Changeset 04b4a71


Ignore:
Timestamp:
Jun 3, 2020, 2:35:13 PM (18 months ago)
Author:
Peter A. Buhr <pabuhr@…>
Branches:
arm-eh, jacob/cs343-translation, master, new-ast, new-ast-unique-expr
Children:
fe9cf9e
Parents:
4e7c0fc0
Message:

update concurrency paper with referee changes and generate a response to the referee's report

Location:
doc/papers
Files:
1 added
3 edited

Legend:

Unmodified
Added
Removed
  • doc/papers/AMA/AMA-stix/ama/WileyNJD-v2.cls

    r4e7c0fc0 r04b4a71  
    24442444     \@afterheading}
    24452445
    2446 \renewcommand\section{\@startsection{section}{1}{\z@}{-25pt \@plus -2pt \@minus -2pt}{12\p@}{\sectionfont}}%
    2447 \renewcommand\subsection{\@startsection{subsection}{2}{\z@}{-22pt \@plus -2pt \@minus -2pt}{5\p@}{\subsectionfont}}%
     2446\renewcommand\section{\@startsection{section}{1}{\z@}{-20pt \@plus -2pt \@minus -2pt}{8\p@}{\sectionfont}}%
     2447\renewcommand\subsection{\@startsection{subsection}{2}{\z@}{-18pt \@plus -2pt \@minus -2pt}{5\p@}{\subsectionfont}}%
    24482448\renewcommand\subsubsection{\@startsection{subsubsection}{3}{\z@}{-20pt \@plus -2pt \@minus -2pt}{2\p@}{\subsubsectionfont}}%
    24492449%
  • doc/papers/concurrency/Paper.tex

    r4e7c0fc0 r04b4a71  
    9999\newcommand{\CRT}{\global\columnposn=\gcolumnposn}
    100100
    101 % Denote newterms in particular font and index them without particular font and in lowercase, e.g., \newterm{abc}.
    102 % The option parameter provides an index term different from the new term, e.g., \newterm[\texttt{abc}]{abc}
     101% Denote newterms in particular font and index them without particular font and in lowercase, \eg \newterm{abc}.
     102% The option parameter provides an index term different from the new term, \eg \newterm[\texttt{abc}]{abc}
    103103% The star version does not lowercase the index information, e.g., \newterm*{IBM}.
    104104\newcommand{\newtermFontInline}{\emph}
     
    235235{\lstset{language=python,moredelim=**[is][\protect\color{red}]{`}{`},#1}\lstset{#1}}
    236236{}
     237\lstnewenvironment{java}[1][]
     238{\lstset{language=java,moredelim=**[is][\protect\color{red}]{`}{`},#1}\lstset{#1}}
     239{}
    237240
    238241% inline code @...@
     
    272275% Library extension for executors, futures, and actors are built on these basic mechanisms.
    273276The runtime provides significant programmer simplification and safety by eliminating spurious wakeup and monitor barging.
    274 The runtime also ensures multiple monitors can be safely acquired \emph{simultaneously} (deadlock free), and this feature is fully integrated with all monitor synchronization mechanisms.
     277The runtime also ensures multiple monitors can be safely acquired in a deadlock-free way, and this feature is fully integrated with all monitor synchronization mechanisms.
    275278All control-flow features integrate with the \CFA polymorphic type-system and exception handling, while respecting the expectations and style of C programmers.
    276279Experimental results show comparable performance of the new features with similar mechanisms in other concurrent programming languages.
     
    293296% "Trait inheritance" works for me. "Interface inheritance" might also be a good choice, and distinguish clearly from implementation inheritance.
    294297% You'll want to be a little bit careful with terms like "structural" and "nominal" inheritance as well. CFA has structural inheritance (I think Go as well) -- it's inferred based on the structure of the code. Java, Rust, and Haskell (not sure about Swift) have nominal inheritance, where there needs to be a specific statement that "this type inherits from this type".
    295 However, functions \emph{cannot} be nested in structures, so there is no lexical binding between a structure and set of functions (member/method) implemented by an implicit \lstinline@this@ (receiver) parameter.},
     298However, functions \emph{cannot} be nested in structures, so there is no lexical binding between a structure and set of functions implemented by an implicit \lstinline@this@ (receiver) parameter.},
    296299backwards-compatible extension of the C programming language.
    297 In many ways, \CFA is to C as Scala~\cite{Scala} is to Java, providing a \emph{research vehicle} for new typing and control-flow capabilities on top of a highly popular programming language\footnote{
    298 The TIOBE index~\cite{TIOBE} for December 2019 ranks the top five \emph{popular} programming languages as Java 17\%, C 16\%, Python 10\%, and \CC 6\%, \Csharp 5\% = 54\%, and over the past 30 years, C has always ranked either first or second in popularity.}
     300In many ways, \CFA is to C as Scala~\cite{Scala} is to Java, providing a vehicle for new typing and control-flow capabilities on top of a highly popular programming language\footnote{
     301The TIOBE index~\cite{TIOBE} for May 2020 ranks the top five \emph{popular} programming languages as C 17\%, Java 16\%, Python 9\%, \CC 6\%, and \Csharp 4\% = 52\%, and over the past 30 years, C has always ranked either first or second in popularity.}
    299302allowing immediate dissemination.
    300303This paper discusses the design philosophy and implementation of advanced language-level control-flow and concurrent/parallel features in \CFA and its runtime, which is written entirely in \CFA.
     
    304307% the concurrency extensions allow high-level management of threads.
    305308
    306 Call/return control-flow with argument/parameter passing appeared in the first programming languages.
     309Call/return control-flow with argument and parameter passing appeared in the first programming languages.
    307310Over the past 50 years, call/return has been augmented with features like static/dynamic call, exceptions (multi-level return) and generators/coroutines (retain state between calls).
    308 While \CFA has mechanisms for dynamic call (algebraic effects) and exceptions\footnote{
     311While \CFA has mechanisms for dynamic call (algebraic effects~\cite{Zhang19}) and exceptions\footnote{
    309312\CFA exception handling will be presented in a separate paper.
    310 The key feature that dovetails with this paper is nonlocal exceptions allowing exceptions to be raised across stacks, with synchronous exceptions raised among coroutines and asynchronous exceptions raised among threads, similar to that in \uC~\cite[\S~5]{uC++}}, this work only discusses retaining state between calls via generators/coroutines.
     313The key feature that dovetails with this paper is nonlocal exceptions allowing exceptions to be raised across stacks, with synchronous exceptions raised among coroutines and asynchronous exceptions raised among threads, similar to that in \uC~\cite[\S~5]{uC++}}, this work only discusses retaining state between calls via generators and coroutines.
    311314\newterm{Coroutining} was introduced by Conway~\cite{Conway63} (1963), discussed by Knuth~\cite[\S~1.4.2]{Knuth73V1}, implemented in Simula67~\cite{Simula67}, formalized by Marlin~\cite{Marlin80}, and is now popular and appears in old and new programming languages: CLU~\cite{CLU}, \Csharp~\cite{Csharp}, Ruby~\cite{Ruby}, Python~\cite{Python}, JavaScript~\cite{JavaScript}, Lua~\cite{Lua}, \CCtwenty~\cite{C++20Coroutine19}.
    312315Coroutining is sequential execution requiring direct handoff among coroutines, \ie only the programmer is controlling execution order.
     
    314317Coroutines are only a stepping stone towards concurrency where the commonality is that coroutines and threads retain state between calls.
    315318
    316 \Celeven/\CCeleven define concurrency~\cite[\S~7.26]{C11}, but it is largely wrappers for a subset of the pthreads library~\cite{Pthreads}.\footnote{Pthreads concurrency is based on simple thread fork/join in a function and mutex/condition locks, which is low-level and error-prone}
    317 Interestingly, almost a decade after the \Celeven standard, neither gcc-9, clang-9 nor msvc-19 (most recent versions) support the \Celeven include @threads.h@, indicating no interest in the C11 concurrency approach (possibly because of the recent effort to add concurrency to \CC).
     319\Celeven/\CCeleven define concurrency~\cite[\S~7.26]{C11}, but it is largely wrappers for a subset of the pthreads library~\cite{Pthreads}.\footnote{Pthreads concurrency is based on simple thread fork and join in a function and mutex or condition locks, which is low-level and error-prone}
     320Interestingly, almost a decade after the \Celeven standard, the most recent versions of gcc, clang, and msvc do not support the \Celeven include @threads.h@, indicating no interest in the C11 concurrency approach (possibly because of the recent effort to add concurrency to \CC).
    318321While the \Celeven standard does not state a threading model, the historical association with pthreads suggests implementations would adopt kernel-level threading (1:1)~\cite{ThreadModel}, as for \CC.
    319322In contrast, there has been a renewed interest during the past decade in user-level (M:N, green) threading in old and new programming languages.
     
    323326As a result, many current languages implementations adopt the 1:1 kernel-threading model, like Java (Scala), Objective-C~\cite{obj-c-book}, \CCeleven~\cite{C11}, C\#~\cite{Csharp} and Rust~\cite{Rust}, with a variety of presentation mechanisms.
    324327From 2000 onwards, several language implementations have championed the M:N user-threading model, like Go~\cite{Go}, Erlang~\cite{Erlang}, Haskell~\cite{Haskell}, D~\cite{D}, and \uC~\cite{uC++,uC++book}, including putting green threads back into Java~\cite{Quasar}, and many user-threading libraries have appeared~\cite{Qthreads,MPC,Marcel}.
    325 The main argument for user-level threading is that it is lighter weight than kernel threading (locking and context switching do not cross the kernel boundary), so there is less restriction on programming styles that encourages large numbers of threads performing medium-sized work to facilitate load balancing by the runtime~\cite{Verch12}.
     328The main argument for user-level threading is that it is lighter weight than kernel threading because locking and context switching do not cross the kernel boundary, so there is less restriction on programming styles that encourages large numbers of threads performing medium-sized work to facilitate load balancing by the runtime~\cite{Verch12}.
    326329As well, user-threading facilitates a simpler concurrency approach using thread objects that leverage sequential patterns versus events with call-backs~\cite{Adya02,vonBehren03}.
    327 Finally, performant user-threading implementations (both time and space) meet or exceed direct kernel-threading implementations, while achieving the programming advantages of high concurrency levels and safety.
     330Finally, performant user-threading implementations, both time and space, meet or exceed direct kernel-threading implementations, while achieving the programming advantages of high concurrency levels and safety.
    328331
    329332A further effort over the past two decades is the development of language memory models to deal with the conflict between language features and compiler/hardware optimizations, \eg some language features are unsafe in the presence of aggressive sequential optimizations~\cite{Buhr95a,Boehm05}.
    330 The consequence is that a language must provide sufficient tools to program around safety issues, as inline and library code is all sequential to the compiler.
    331 One solution is low-level qualifiers and functions (\eg @volatile@ and atomics) allowing \emph{programmers} to explicitly write safe (race-free~\cite{Boehm12}) programs.
    332 A safer solution is high-level language constructs so the \emph{compiler} knows the concurrency boundaries (where mutual exclusion and synchronization are acquired/released) and provide implicit safety at and across these boundaries.
     333The consequence is that a language must provide sufficient tools to program around safety issues, as inline and library code is compiled as sequential without any explicit concurrent directive.
     334One solution is low-level qualifiers and functions, \eg @volatile@ and atomics, allowing \emph{programmers} to explicitly write safe, race-free~\cite{Boehm12} programs.
     335A safer solution is high-level language constructs so the \emph{compiler} knows the concurrency boundaries, \ie where mutual exclusion and synchronization are acquired and released, and provide implicit safety at and across these boundaries.
    333336While the optimization problem is best known with respect to concurrency, it applies to other complex control-flow, like exceptions and coroutines.
    334337As well, language solutions allow matching the language paradigm with the approach, \eg matching the functional paradigm with data-flow programming or the imperative paradigm with thread programming.
    335338
    336 Finally, it is important for a language to provide safety over performance \emph{as the default}, allowing careful reduction of safety (unsafe code) for performance when necessary.
    337 Two concurrency violations of this philosophy are \emph{spurious wakeup} (random wakeup~\cite[\S~9]{Buhr05a}) and \emph{barging}\footnote{
     339Finally, it is important for a language to provide safety over performance \emph{as the default}, allowing careful reduction of safety for performance when necessary.
     340Two concurrency violations of this philosophy are \emph{spurious or random wakeup}~\cite[\S~9]{Buhr05a}) and \emph{barging}\footnote{
    338341Barging is competitive succession instead of direct handoff, \ie after a lock is released both arriving and preexisting waiter threads compete to acquire the lock.
    339342Hence, an arriving thread can temporally \emph{barge} ahead of threads already waiting for an event, which can repeat indefinitely leading to starvation of waiter threads.
    340 } (signals-as-hints~\cite[\S~8]{Buhr05a}), where one is a consequence of the other, \ie once there is spurious wakeup, signals-as-hints follow.
     343} or signals-as-hints~\cite[\S~8]{Buhr05a}, where one is a consequence of the other, \ie once there is spurious wakeup, barging follows.
    341344(Author experience teaching concurrency is that students are confused by these semantics.)
    342345However, spurious wakeup is \emph{not} a foundational concurrency property~\cite[\S~9]{Buhr05a};
     
    356359
    357360\item
    358 monitor synchronization without barging, and the ability to safely acquiring multiple monitors \emph{simultaneously} (deadlock free), while seamlessly integrating these capabilities with all monitor synchronization mechanisms,
     361monitor synchronization without barging, and the ability to safely acquiring multiple monitors in a deadlock-free way, while seamlessly integrating these capabilities with all monitor synchronization mechanisms,
    359362
    360363\item
     
    368371
    369372\item
    370 a dynamic partitioning mechanism to segregate groups of executing user and kernel threads performing specialized work (\eg web-server or compute engine) or requiring different scheduling (\eg NUMA or real-time).
     373a dynamic partitioning mechanism to segregate groups of executing user and kernel threads performing specialized work, \eg web-server or compute engine, or requiring different scheduling, \eg NUMA or real-time.
    371374
    372375% \item
     
    380383Section~\ref{s:StatefulFunction} begins advanced control by introducing sequential functions that retain data and execution state between calls producing constructs @generator@ and @coroutine@.
    381384Section~\ref{s:Concurrency} begins concurrency, or how to create (fork) and destroy (join) a thread producing the @thread@ construct.
    382 Section~\ref{s:MutualExclusionSynchronization} discusses the two mechanisms to restricted nondeterminism when controlling shared access to resources (mutual exclusion) and timing relationships among threads (synchronization).
     385Section~\ref{s:MutualExclusionSynchronization} discusses the two mechanisms to restricted nondeterminism when controlling shared access to resources, called mutual exclusion, and timing relationships among threads, called synchronization.
    383386Section~\ref{s:Monitor} shows how both mutual exclusion and synchronization are safely embedded in the @monitor@ and @thread@ constructs.
    384 Section~\ref{s:CFARuntimeStructure} describes the large-scale mechanism to structure (cluster) threads and virtual processors (kernel threads).
     387Section~\ref{s:CFARuntimeStructure} describes the large-scale mechanism to structure threads and virtual processors (kernel threads).
    385388Section~\ref{s:Performance} uses a series of microbenchmarks to compare \CFA threading with pthreads, Java 11.0.6, Go 1.12.6, Rust 1.37.0, Python 3.7.6, Node.js 12.14.1, and \uC 7.0.0.
    386389
     
    392395To this end, the control-flow features created for \CFA are based on the fundamental properties of any language with function-stack control-flow (see also \uC~\cite[pp.~140-142]{uC++}).
    393396The fundamental properties are execution state, thread, and mutual-exclusion/synchronization (MES).
    394 These independent properties can be used alone, in pairs, or in triplets to compose different language features, forming a compositional hierarchy where the most advanced feature has all the properties (state/thread/MES).
    395 While it is possible for a language to only support the most advanced feature~\cite{Hermes90}, this unnecessarily complicates and makes inefficient solutions to certain classes of problems.
    396 As is shown, each of the (non-rejected) composed features solves a particular set of problems, and hence, has a defensible position in a programming language.
    397 If a compositional feature is missing, a programmer has too few/many fundamental properties resulting in a complex and/or is inefficient solution.
     397These independent properties can be used to compose different language features, forming a compositional hierarchy, where the combination of all three is the most advanced feature, called a thread/task/process.
     398While it is possible for a language to only provide threads for composing programs~\cite{Hermes90}, this unnecessarily complicates and makes inefficient solutions to certain classes of problems.
     399As is shown, each of the non-rejected composed language features solves a particular set of problems, and hence, has a defensible position in a programming language.
     400If a compositional feature is missing, a programmer has too few fundamental properties resulting in a complex and/or is inefficient solution.
    398401
    399402In detail, the fundamental properties are:
    400403\begin{description}[leftmargin=\parindent,topsep=3pt,parsep=0pt]
    401404\item[\newterm{execution state}:]
    402 is the state information needed by a control-flow feature to initialize, manage compute data and execution location(s), and de-initialize.
    403 State is retained in fixed-sized aggregate structures and dynamic-sized stack(s), often allocated in the heap(s) managed by the runtime system.
     405is the state information needed by a control-flow feature to initialize, manage compute data and execution location(s), and de-initialize, \eg calling a function initializes a stack frame including contained objects with constructors, manages local data in blocks and return locations during calls, and de-initializes the frame by running any object destructors and management operations.
     406State is retained in fixed-sized aggregate structures (objects) and dynamic-sized stack(s), often allocated in the heap(s) managed by the runtime system.
    404407The lifetime of the state varies with the control-flow feature, where longer life-time and dynamic size provide greater power but also increase usage complexity and cost.
    405 Control-flow transfers among execution states occurs in multiple ways, such as function call, context switch, asynchronous await, etc.
     408Control-flow transfers among execution states in multiple ways, such as function call, context switch, asynchronous await, etc.
    406409Because the programming language determines what constitutes an execution state, implicitly manages this state, and defines movement mechanisms among states, execution state is an elementary property of the semantics of a programming language.
    407410% An execution-state is related to the notion of a process continuation \cite{Hieb90}.
     
    410413is execution of code that occurs independently of other execution, \ie the execution resulting from a thread is sequential.
    411414Multiple threads provide \emph{concurrent execution};
    412 concurrent execution becomes parallel when run on multiple processing units (hyper-threading, cores, sockets).
    413 There must be language mechanisms to create, block/unblock, and join with a thread.
     415concurrent execution becomes parallel when run on multiple processing units, \eg hyper-threading, cores, or sockets.
     416There must be language mechanisms to create, block and unblock, and join with a thread.
    414417
    415418\item[\newterm{MES}:]
    416419is the concurrency mechanisms to perform an action without interruption and establish timing relationships among multiple threads.
    417 These two properties are independent, \ie mutual exclusion cannot provide synchronization and vice versa without introducing additional threads~\cite[\S~4]{Buhr05a}.
     420We contented these two properties are independent, \ie mutual exclusion cannot provide synchronization and vice versa without introducing additional threads~\cite[\S~4]{Buhr05a}.
    418421Limiting MES, \eg no access to shared data, results in contrived solutions and inefficiency on multi-core von Neumann computers where shared memory is a foundational aspect of its design.
    419422\end{description}
     
    421424
    422425
    423 \subsection{Execution Properties}
    424 
    425 Table~\ref{t:ExecutionPropertyComposition} shows how the three fundamental execution properties: state, thread, and mutual exclusion compose a hierarchy of control-flow features needed in a programming language.
    426 (When doing case analysis, not all combinations are meaningful.)
    427 Note, basic von Neumann execution requires at least one thread and an execution state providing some form of call stack.
     426\subsection{Structuring Execution Properties}
     427
     428Programming languages seldom present the fundamental execution properties directly to programmers.
     429Instead, the properties are packaged into higher-level constructs that encapsulate details and provide safety to these low-level mechanisms.
     430Interestingly, language designers often pick and choose among these execution properties proving a varying subset of constructs.
     431
     432Table~\ref{t:ExecutionPropertyComposition} shows all combinations of the three fundamental execution properties available to language designers.
     433(When doing combination case-analysis, not all combinations are meaningful.)
     434The combinations of state, thread, and mutual exclusion compose a hierarchy of control-flow features all of which have appeared in prior programming languages, where each of these languages have found the feature useful.
     435To understand the table, it is important to review the basic von Neumann execution requirement of at least one thread and execution state providing some form of call stack.
    428436For table entries missing these minimal components, the property is borrowed from the invoker (caller).
    429 
    430 Case 1 is a function that borrows storage for its state (stack frame/activation) and a thread from its invoker and retains this state across \emph{callees}, \ie function local-variables are retained on the stack across calls.
    431 Case 2 is case 1 with access to shared state so callers are restricted during update (mutual exclusion) and scheduling for other threads (synchronization).
    432 Case 3 is a stateful function supporting resume/suspend along with call/return to retain state across \emph{callers}, but has some restrictions because the function's state is stackless.
    433 Note, stackless functions still borrow the caller's stack and thread, where the stack is used to preserve state across its callees.
    434 Case 4 is cases 2 and 3 with protection to shared state for stackless functions.
    435 Cases 5 and 6 are the same as 3 and 4 but only the thread is borrowed as the function state is stackful, so resume/suspend is a context switch from the caller's to the function's stack.
    436 Cases 7 and 8 are rejected because a function that is given a new thread must have its own stack where the thread begins and stack frames are stored for calls, \ie there is no stack to borrow.
    437 Cases 9 and 10 are rejected because a thread with a fixed state (no stack) cannot accept calls, make calls, block, or be preempted, all of which require an unknown amount of additional dynamic state.
    438 Hence, once started, this kind of thread must execute to completion, \ie computation only, which severely restricts runtime management.
    439 Cases 11 and 12 have a stackful thread with and without safe access to shared state.
    440 Execution properties increase the cost of creation and execution along with complexity of usage.
     437Each entry in the table, numbered \textbf{1}--\textbf{12}, is discussed with respect to how the execution properties combine to generate a high-level language construct.
    441438
    442439\begin{table}
     
    452449\hline   
    453450\hline   
    454 No                                      & No            & \textbf{1}\ \ \ function                              & \textbf{2}\ \ \ @monitor@ function    \\
     451No                                      & No            & \textbf{1}\ \ \ @struct@                              & \textbf{2}\ \ \ @mutex@ @struct@              \\
    455452\hline   
    456 Yes (stackless)         & No            & \textbf{3}\ \ \ @generator@                   & \textbf{4}\ \ \ @monitor@ @generator@ \\
     453Yes (stackless)         & No            & \textbf{3}\ \ \ @generator@                   & \textbf{4}\ \ \ @mutex@ @generator@   \\
    457454\hline   
    458 Yes (stackful)          & No            & \textbf{5}\ \ \ @coroutine@                   & \textbf{6}\ \ \ @monitor@ @coroutine@ \\
     455Yes (stackful)          & No            & \textbf{5}\ \ \ @coroutine@                   & \textbf{6}\ \ \ @mutex@ @coroutine@   \\
    459456\hline   
    460457No                                      & Yes           & \textbf{7}\ \ \ {\color{red}rejected} & \textbf{8}\ \ \ {\color{red}rejected} \\
     
    462459Yes (stackless)         & Yes           & \textbf{9}\ \ \ {\color{red}rejected} & \textbf{10}\ \ \ {\color{red}rejected} \\
    463460\hline   
    464 Yes (stackful)          & Yes           & \textbf{11}\ \ \ @thread@                             & \textbf{12}\ \ @monitor@ @thread@             \\
     461Yes (stackful)          & Yes           & \textbf{11}\ \ \ @thread@                             & \textbf{12}\ \ @mutex@ @thread@               \\
    465462\end{tabular}
    466463\end{table}
    467464
    468 Given the execution-properties taxonomy, programmers can now answer three basic questions: is state necessary across calls and how much, is a separate thread necessary, is access to shared state necessary.
    469 The answers define the optimal language feature need for implementing a programming problem.
    470 The next sections discusses how \CFA fills in the table with language features, while other programming languages may only provide a subset of the table.
     465Case 1 is a structure where access functions borrow local state (stack frame/activation) and thread from the invoker and retain this state across \emph{callees}, \ie function local-variables are retained on the borrowed stack during calls.
     466Structures are a foundational mechanism for data organization, and access functions provide interface abstraction and code sharing in all programming languages.
     467Case 2 is case 1 with thread safety to a structure's state where access functions provide serialization (mutual exclusion) and scheduling among calling threads (synchronization).
     468A @mutex@ structure, often called a \newterm{monitor}, provides a high-level interface for race-free access of shared data in concurrent programming-languages.
     469Case 3 is case 1 where the structure can implicitly retain execution state and access functions use this execution state to resume/suspend across \emph{callers}, but resume/suspend does not retain a function's local state.
     470A stackless structure, often called a \newterm{generator} or \emph{iterator}, is \newterm{stackless} because it still borrow the caller's stack and thread, where the stack is used only to preserve state across its callees not callers.
     471Generators provide the first step toward directly solving problems like finite-state machines that retain data and execution state between calls, whereas normal functions restart on each call.
     472Case 4 is cases 2 and 3 with thread safety during execution of the generator's access functions.
     473A @mutex@ generator extends generators into the concurrent domain.
     474Cases 5 and 6 are like cases 3 and 4 where the structure is extended with an implicit separate stack, so only the thread is borrowed by access functions.
     475A stackful generator, often called a \newterm{coroutine}, is \newterm{stackful} because resume/suspend now context switch to/from the caller's and coroutine's stack.
     476A coroutine extends the state retained between calls beyond the generator's structure to arbitrary call depth in the access functions.
     477Cases 7 and 8 are rejected because a new thread must have its own stack, where the thread begins and stack frames are stored for calls, \ie it is unrealistic for a thread to borrow a stack.
     478Cases 9 and 10 are rejected because a thread needs a growable stack to accept calls, make calls, block, or be preempted, all of which compound to require an unknown amount of execution state.
     479If this kind of thread exists, it must execute to completion, \ie computation only, which severely restricts runtime management.
     480Cases 11 and 12 are a stackful thread with and without safe access to shared state.
     481A thread is the language mechanism to start another thread of control in a program with growable execution state for call/return execution.
     482In general, more execution properties increase the cost of creation and execution along with complexity of usage.
     483
     484Given the execution-properties taxonomy, programmers now ask three basic questions: is state necessary across callers and how much, is a separate thread necessary, is thread safety necessary.
     485Table~\ref{t:ExecutionPropertyComposition} then suggests the optimal language feature needed for implementing a programming problem.
     486The following sections describe how \CFA fills in \emph{all} the non-rejected table entries with language features, while other programming languages may only provide a subset of the table.
    471487
    472488
     
    481497\item
    482498Direct interaction among language features must be possible allowing any feature to be selected without restricting comm\-unication.
    483 For example, many concurrent languages do not provide direct communication (calls) among threads, \ie threads only communicate indirectly through monitors, channels, messages, and/or futures.
    484 Indirect communication increases the number of objects, consuming more resources, and require additional synchronization and possibly data transfer.
     499For example, many concurrent languages do not provide direct communication calls among threads, \ie threads only communicate indirectly through monitors, channels, messages, and/or futures.
     500Indirect communication increases the number of objects, consuming more resources, and requires additional synchronization and possibly data transfer.
    485501
    486502\item
     
    493509
    494510\item
    495 MES must be available implicitly in language constructs as well as explicitly for specialized requirements, because requiring programmers to build MES using low-level locks often leads to incorrect programs.
     511MES must be available implicitly in language constructs, \eg Java built-in monitors, as well as explicitly for specialized requirements, \eg @java.util.concurrent@, because requiring programmers to build MES using low-level locks often leads to incorrect programs.
    496512Furthermore, reducing synchronization scope by encapsulating it within language constructs further reduces errors in concurrent programs.
    497513
     
    502518\item
    503519Synchronization must be able to control the service order of requests including prioritizing selection from different kinds of outstanding requests, and postponing a request for an unspecified time while continuing to accept new requests.
    504 Otherwise, certain concurrency problems are difficult, e.g.\ web server, disk scheduling, and the amount of concurrency is inhibited~\cite{Gentleman81}.
     520Otherwise, certain concurrency problems are difficult, \eg web server, disk scheduling, and the amount of concurrency is inhibited~\cite{Gentleman81}.
    505521\end{itemize}
    506522We have satisfied these requirements in \CFA while maintaining backwards compatibility with the huge body of legacy C programs.
     
    511527
    512528Asynchronous await/call is a caller mechanism for structuring programs and/or increasing concurrency, where the caller (client) postpones an action into the future, which is subsequently executed by a callee (server).
    513 The caller detects the action's completion through a \newterm{future}/\newterm{promise}.
     529The caller detects the action's completion through a \newterm{future} or \newterm{promise}.
    514530The benefit is asynchronous caller execution with respect to the callee until future resolution.
    515531For single-threaded languages like JavaScript, an asynchronous call passes a callee action, which is queued in the event-engine, and continues execution with a promise.
     
    517533A promise-completion call-back can be part of the callee action or the caller is rescheduled;
    518534in either case, the call back is executed after the promise is fulfilled.
    519 While asynchronous calls generate new callee (server) events, we content this mechanism is insufficient for advanced control-flow mechanisms like generators or coroutines (which are discussed next).
     535While asynchronous calls generate new callee (server) events, we contend this mechanism is insufficient for advanced control-flow mechanisms like generators or coroutines, which are discussed next.
    520536Specifically, control between caller and callee occurs indirectly through the event-engine precluding direct handoff and cycling among events, and requires complex resolution of a control promise and data.
    521537Note, @async-await@ is just syntactic-sugar over the event engine so it does not solve these deficiencies.
    522538For multi-threaded languages like Java, the asynchronous call queues a callee action with an executor (server), which subsequently executes the work by a thread in the executor thread-pool.
    523 The problem is when concurrent work-units need to interact and/or block as this effects the executor, \eg stops threads.
     539The problem is when concurrent work-units need to interact and/or block as this effects the executor by stopping threads.
    524540While it is possible to extend this approach to support the necessary mechanisms, \eg message passing in Actors, we show monitors and threads provide an equally competitive approach that does not deviate from normal call communication and can be used to build asynchronous call, as is done in Java.
    525541
     
    540556There are two styles of activating a stateful function, \emph{asymmetric} or \emph{symmetric}, identified by resume/suspend (no cycles) and resume/resume (cycles).
    541557These styles \emph{do not} cause incremental stack growth, \eg a million resume/suspend or resume/resume cycles do not remember each cycle just the last resumer for each cycle.
    542 Selecting between stackless/stackful semantics and asymmetric/symmetric style is a tradeoff between programming requirements, performance, and design, where stackless is faster and smaller (modified call/return between closures), stackful is more general but slower and larger (context switching between distinct stacks), and asymmetric is simpler control-flow than symmetric.
    543 Additionally, storage management for the closure/stack (especially in unmanaged languages, \ie no garbage collection) must be factored into design and performance.
     558Selecting between stackless/stackful semantics and asymmetric/symmetric style is a tradeoff between programming requirements, performance, and design, where stackless is faster and smaller using modified call/return between closures, stackful is more general but slower and larger using context switching between distinct stacks, and asymmetric is simpler control-flow than symmetric.
     559Additionally, storage management for the closure/stack must be factored into design and performance, especially in unmanaged languages without garbage collection.
    544560Note, creation cost (closure/stack) is amortized across usage, so activation cost (resume/suspend) is usually the dominant factor.
    545561
     
    690706\hspace{3pt}
    691707\subfloat[C generated code for \CFA version]{\label{f:CFibonacciSim}\usebox\myboxC}
    692 \caption{Fibonacci (output) asymmetric generator}
     708\caption{Fibonacci output asymmetric generator}
    693709\label{f:FibonacciAsymmetricGenerator}
    694710
     
    765781\subfloat[C generated code for \CFA version]{\label{f:CFormatGenImpl}\usebox\myboxB}
    766782\hspace{3pt}
    767 \caption{Formatter (input) asymmetric generator}
     783\caption{Formatter input asymmetric generator}
    768784\label{f:FormatterAsymmetricGenerator}
    769785\end{figure}
    770786
    771 Figure~\ref{f:FibonacciAsymmetricGenerator} shows an unbounded asymmetric generator for an infinite sequence of Fibonacci numbers written (left to right) in C, \CFA, and showing the underlying C implementation for the \CFA version.
     787Figure~\ref{f:FibonacciAsymmetricGenerator} shows an unbounded asymmetric generator for an infinite sequence of Fibonacci numbers written left to right in C, \CFA, and showing the underlying C implementation for the \CFA version.
    772788This generator is an \emph{output generator}, producing a new result on each resumption.
    773789To compute Fibonacci, the previous two values in the sequence are retained to generate the next value, \ie @fn1@ and @fn@, plus the execution location where control restarts when the generator is resumed, \ie top or middle.
    774 An additional requirement is the ability to create an arbitrary number of generators (of any kind), \ie retaining one state in global variables is insufficient;
     790An additional requirement is the ability to create an arbitrary number of generators of any kind, \ie retaining one state in global variables is insufficient;
    775791hence, state is retained in a closure between calls.
    776792Figure~\ref{f:CFibonacci} shows the C approach of manually creating the closure in structure @Fib@, and multiple instances of this closure provide multiple Fibonacci generators.
     
    794810Figure~\ref{f:CFibonacciSim} shows the C implementation of the \CFA asymmetric generator.
    795811Only one execution-state field, @restart@, is needed to subscript the suspension points in the generator.
    796 At the start of the generator main, the @static@ declaration, @states@, is initialized to the N suspend points in the generator (where operator @&&@ dereferences/references a label~\cite{gccValueLabels}).
     812At the start of the generator main, the @static@ declaration, @states@, is initialized to the N suspend points in the generator, where operator @&&@ dereferences or references a label~\cite{gccValueLabels}.
    797813Next, the computed @goto@ selects the last suspend point and branches to it.
    798 The  cost of setting @restart@ and branching via the computed @goto@ adds very little cost to the suspend/resume calls.
     814The  cost of setting @restart@ and branching via the computed @goto@ adds very little cost to the suspend and resume calls.
    799815
    800816An advantage of the \CFA explicit generator type is the ability to allow multiple type-safe interface functions taking and returning arbitrary types.
     
    917933When a message byte arrives, it triggers an interrupt, and the operating system services the interrupt by calling the device driver with the byte read from a hardware register.
    918934The device driver returns a status code of its current state, and when a complete message is obtained, the operating system read the message accumulated in the supplied buffer.
    919 Hence, the device driver is an input/output generator, where the cost of resuming the device-driver generator is the same as call/return, so performance in an operating-system kernel is excellent.
     935Hence, the device driver is an input/output generator, where the cost of resuming the device-driver generator is the same as call and return, so performance in an operating-system kernel is excellent.
    920936The key benefits of using a generator are correctness, safety, and maintenance because the execution states are transcribed directly into the programming language rather than table lookup or stack ripping.
    921937The conclusion is that FSMs are complex and occur in important domains, so direct generator support is important in a system programming language.
     
    982998Constructing the cycle must deal with definition-before-use to close the cycle, \ie, the first generator must know about the last generator, which is not within scope.
    983999(This issue occurs for any cyclic data structure.)
    984 The example creates the generators, @ping@/@pong@, and then assigns the partners that form the cycle.
     1000The example creates the generators, @ping@ and @pong@, and then assigns the partners that form the cycle.
    9851001% (Alternatively, the constructor can assign the partners as they are declared, except the first, and the first-generator partner is set after the last generator declaration to close the cycle.)
    9861002Once the cycle is formed, the program main resumes one of the generators, @ping@, and the generators can then traverse an arbitrary cycle using @resume@ to activate partner generator(s).
     
    10661082While the semantics of call forward is a tail-call optimization, which compilers perform, the generator state is different on each call rather a common state for a tail-recursive function (i.e., the parameter to the function never changes during the forward calls.
    10671083However, this assembler code depends on what entry code is generated, specifically if there are local variables and the level of optimization.
    1068 Hence, internal compiler support is necessary for any forward call (or backwards return), \eg LLVM has various coroutine support~\cite{CoroutineTS}, and \CFA can leverage this support should it eventually fork @clang@.
    1069 For this reason, \CFA does not support general symmetric generators at this time, but, it is possible to hand generate any symmetric generators (as in Figure~\ref{f:CPingPongSim}) for proof of concept and performance testing.
     1084Hence, internal compiler support is necessary for any forward call or backwards return, \eg LLVM has various coroutine support~\cite{CoroutineTS}, and \CFA can leverage this support should it eventually fork @clang@.
     1085For this reason, \CFA does not support general symmetric generators at this time, but, it is possible to hand generate any symmetric generators, as in Figure~\ref{f:CPingPongSim}, for proof of concept and performance testing.
    10701086
    10711087Finally, part of this generator work was inspired by the recent \CCtwenty coroutine proposal~\cite{C++20Coroutine19}, which uses the general term coroutine to mean generator.
     
    11261142\begin{cfa}
    11271143int Crc() {
    1128         `suspend;`
    1129         short int crc = byte << 8;
    1130         `suspend;`
    1131         status = (crc | byte) == sum ? MSG : ECRC;
     1144        `suspend;`  short int crc = byte << 8;
     1145        `suspend;`  status = (crc | byte) == sum ? MSG : ECRC;
    11321146        return crc;
    11331147}
     
    11391153\begin{comment}
    11401154Figure~\ref{f:Coroutine3States} creates a @coroutine@ type, @`coroutine` Fib { int fn; }@, which provides communication, @fn@, for the \newterm{coroutine main}, @main@, which runs on the coroutine stack, and possibly multiple interface functions, \eg @restart@.
    1141 Like the structure in Figure~\ref{f:ExternalState}, the coroutine type allows multiple instances, where instances of this type are passed to the (overloaded) coroutine main.
     1155Like the structure in Figure~\ref{f:ExternalState}, the coroutine type allows multiple instances, where instances of this type are passed to the overloaded coroutine main.
    11421156The coroutine main's stack holds the state for the next generation, @f1@ and @f2@, and the code represents the three states in the Fibonacci formula via the three suspend points, to context switch back to the caller's @resume@.
    11431157The interface function @restart@, takes a Fibonacci instance and context switches to it using @resume@;
     
    13731387
    13741388Figure~\ref{f:ProdCons} shows the ping-pong example in Figure~\ref{f:CFAPingPongGen} extended into a producer/consumer symmetric-coroutine performing bidirectional communication.
    1375 This example is illustrative because both producer/consumer have two interface functions with @resume@s that suspend execution in these interface (helper) functions.
     1389This example is illustrative because both producer and consumer have two interface functions with @resume@s that suspend execution in these interface functions.
    13761390The program main creates the producer coroutine, passes it to the consumer coroutine in its initialization, and closes the cycle at the call to @start@ along with the number of items to be produced.
    13771391The call to @start@ is the first @resume@ of @prod@, which remembers the program main as the starter and creates @prod@'s stack with a frame for @prod@'s coroutine main at the top, and context switches to it.
     
    14011415The question now is where does control continue?
    14021416
    1403 The na\"{i}ve semantics for coroutine-cycle termination is to context switch to the last resumer, like executing a @suspend@/@return@ in a generator.
     1417The na\"{i}ve semantics for coroutine-cycle termination is to context switch to the last resumer, like executing a @suspend@ or @return@ in a generator.
    14041418However, for coroutines, the last resumer is \emph{not} implicitly below the current stack frame, as for generators, because each coroutine's stack is independent.
    14051419Unfortunately, it is impossible to determine statically if a coroutine is in a cycle and unrealistic to check dynamically (graph-cycle problem).
     
    14101424For asymmetric coroutines, it is common for the first resumer (starter) coroutine to be the only resumer;
    14111425for symmetric coroutines, it is common for the cycle creator to persist for the lifetime of the cycle.
    1412 For other scenarios, it is always possible to devise a solution with additional programming effort, such as forcing the cycle forward (backward) to a safe point before starting termination.
     1426For other scenarios, it is always possible to devise a solution with additional programming effort, such as forcing the cycle forward or backward to a safe point before starting termination.
    14131427
    14141428Note, the producer/consumer example does not illustrate the full power of the starter semantics because @cons@ always ends first.
    14151429Assume generator @PingPong@ in Figure~\ref{f:PingPongSymmetricGenerator} is converted to a coroutine.
    14161430Unlike generators, coroutines have a starter structure with multiple levels, where the program main starts @ping@ and @ping@ starts @pong@.
    1417 By adjusting $N$ for either @ping@/@pong@, it is possible to have either finish first.
     1431By adjusting $N$ for either @ping@ or @pong@, it is possible to have either finish first.
    14181432If @pong@ ends first, it resumes its starter @ping@ in its coroutine main, then @ping@ ends and resumes its starter the program main on return;
    14191433if @ping@ ends first, it resumes its starter the program main on return.
     
    14251439\subsection{Generator / Coroutine Implementation}
    14261440
    1427 A significant implementation challenge for generators/coroutines (and threads in Section~\ref{s:threads}) is adding extra fields to the custom types and related functions, \eg inserting code after/before the coroutine constructor/destructor and @main@ to create/initialize/de-initialize/destroy any extra fields, \eg stack.
     1441A significant implementation challenge for generators and coroutines (and threads in Section~\ref{s:threads}) is adding extra fields to the custom types and related functions, \eg inserting code after/before the coroutine constructor/destructor and @main@ to create/initialize/de-initialize/destroy any extra fields, \eg the coroutine stack.
    14281442There are several solutions to these problem, which follow from the object-oriented flavour of adopting custom types.
    14291443
     
    14331447\end{cfa}
    14341448% The problem is that the programming language and its tool chain, \eg debugger, @valgrind@, need to understand @baseCoroutine@ because it infers special property, so type @baseCoroutine@ becomes a de facto keyword and all types inheriting from it are implicitly custom types.
    1435 The problem is that some special properties are not handled by existing language semantics, \eg the execution of constructors/destructors is in the wrong order to implicitly start threads because the thread must start \emph{after} all constructors as it relies on a completely initialized object, but the inherited constructor runs \emph{before} the derived.
     1449The problem is that some special properties are not handled by existing language semantics, \eg the execution of constructors and destructors is in the wrong order to implicitly start threads because the thread must start \emph{after} all constructors as it relies on a completely initialized object, but the inherited constructor runs \emph{before} the derived.
    14361450Alternatives, such as explicitly starting threads as in Java, are repetitive and forgetting to call start is a common source of errors.
    14371451An alternative is composition:
     
    14611475forall( `dtype` T | is_coroutine(T) ) void $suspend$( T & ), resume( T & );
    14621476\end{cfa}
    1463 Note, copying generators/coroutines/threads is undefined because muliple objects cannot execute on a shared stack and stack copying does not work in unmanaged languages (no garbage collection), like C, because the stack may contain pointers to objects within it that require updating for the copy.
    1464 The \CFA @dtype@ property provides no \emph{implicit} copying operations and the @is_coroutine@ trait provides no \emph{explicit} copying operations, so all coroutines must be passed by reference (pointer).
    1465 The function definitions ensure there is a statically typed @main@ function that is the starting point (first stack frame) of a coroutine, and a mechanism to get (read) the coroutine descriptor from its handle.
    1466 The @main@ function has no return value or additional parameters because the coroutine type allows an arbitrary number of interface functions with corresponding arbitrary typed input/output values versus fixed ones.
     1477Note, copying generators, coroutines, and threads is undefined because muliple objects cannot execute on a shared stack and stack copying does not work in unmanaged languages (no garbage collection), like C, because the stack may contain pointers to objects within it that require updating for the copy.
     1478The \CFA @dtype@ property provides no \emph{implicit} copying operations and the @is_coroutine@ trait provides no \emph{explicit} copying operations, so all coroutines must be passed by reference or pointer.
     1479The function definitions ensure there is a statically typed @main@ function that is the starting point (first stack frame) of a coroutine, and a mechanism to read the coroutine descriptor from its handle.
     1480The @main@ function has no return value or additional parameters because the coroutine type allows an arbitrary number of interface functions with corresponding arbitrary typed input and output values versus fixed ones.
    14671481The advantage of this approach is that users can easily create different types of coroutines, \eg changing the memory layout of a coroutine is trivial when implementing the @get_coroutine@ function, and possibly redefining \textsf{suspend} and @resume@.
    14681482
     
    15061520
    15071521Figure~\ref{f:CoroutineMemoryLayout} shows different memory-layout options for a coroutine (where a thread is similar).
    1508 The coroutine handle is the @coroutine@ instance containing programmer specified type global/communication variables across interface functions.
     1522The coroutine handle is the @coroutine@ instance containing programmer specified type global and communication variables across interface functions.
    15091523The coroutine descriptor contains all implicit declarations needed by the runtime, \eg @suspend@/@resume@, and can be part of the coroutine handle or separate.
    15101524The coroutine stack can appear in a number of locations and be fixed or variable sized.
     
    15131527Once allocated, a VLS is fixed sized.}
    15141528on the allocating stack, provided the allocating stack is large enough.
    1515 For a VLS stack allocation/deallocation is an inexpensive adjustment of the stack pointer, modulo any stack constructor costs (\eg initial frame setup).
    1516 For stack allocation in the heap, allocation/deallocation is an expensive allocation, where the heap can be a shared resource, modulo any stack constructor costs.
    1517 It is also possible to use a split (segmented) stack calling convention, available with gcc and clang, allowing a variable-sized stack via a set of connected blocks in the heap.
    1518 Currently, \CFA supports stack/heap allocated descriptors but only fixed-sized heap allocated stacks.
     1529For a VLS stack allocation and deallocation is an inexpensive adjustment of the stack pointer, modulo any stack constructor costs to initial frame setup.
     1530For stack allocation in the heap, allocation and deallocation is an expensive allocation, where the heap can be a shared resource, modulo any stack constructor costs.
     1531It is also possible to use a split or segmented stack calling convention, available with gcc and clang, allowing a variable-sized stack via a set of connected blocks in the heap.
     1532Currently, \CFA supports stack and heap allocated descriptors but only fixed-sized heap allocated stacks.
    15191533In \CFA debug-mode, the fixed-sized stack is terminated with a write-only page, which catches most stack overflows.
    15201534Experience teaching concurrency with \uC~\cite{CS343} shows fixed-sized stacks are rarely an issue for students.
     
    15391553The transition to concurrency, even for a single thread with multiple stacks, occurs when coroutines context switch to a \newterm{scheduling coroutine}, introducing non-determinism from the coroutine perspective~\cite[\S~3]{Buhr05a}.
    15401554Therefore, a minimal concurrency system requires coroutines \emph{in conjunction with a nondeterministic scheduler}.
    1541 The resulting execution system now follows a cooperative threading-model~\cite{Adya02,libdill} because context-switching points to the scheduler (blocking) are known, but the next unblocking point is unknown due to the scheduler.
     1555The resulting execution system now follows a cooperative threading-model~\cite{Adya02,libdill} because context-switching points to the scheduler are known, but the next unblocking point is unknown due to the scheduler.
    15421556Adding \newterm{preemption} introduces \newterm{non-cooperative} or \newterm{preemptive} scheduling, where context switching points to the scheduler are unknown as they can occur randomly between any two instructions often based on a timer interrupt.
    15431557Uncertainty gives the illusion of parallelism on a single processor and provides a mechanism to access and increase performance on multiple processors.
    1544 The reason is that the scheduler/runtime have complete knowledge about resources and how to best utilized them.
     1558The reason is that the scheduler and runtime have complete knowledge about resources and how to best utilized them.
    15451559However, the introduction of unrestricted nondeterminism results in the need for \newterm{mutual exclusion} and \newterm{synchronization}~\cite[\S~4]{Buhr05a}, which restrict nondeterminism for correctness;
    15461560otherwise, it is impossible to write meaningful concurrent programs.
     
    15891603\vspace{1pt}
    15901604\par\noindent
    1591 \CFA has a simpler approach using a custom @thread@ type and leveraging declaration semantics (allocation/deallocation), where threads implicitly @fork@ after construction and @join@ before destruction.
     1605\CFA has a simpler approach using a custom @thread@ type and leveraging declaration semantics, allocation and deallocation, where threads implicitly @fork@ after construction and @join@ before destruction.
    15921606\begin{cfa}
    15931607thread MyThread {};
     
    15981612} $\C{// deallocate stack-based threads, implicit joins before destruction}$
    15991613\end{cfa}
    1600 This semantic ensures a thread is started and stopped exactly once, eliminating some programming error, and scales to multiple threads for basic (termination) synchronization.
    1601 For block allocation to arbitrary depth, including recursion, threads are created/destroyed in a lattice structure (tree with top and bottom).
     1614This semantic ensures a thread is started and stopped exactly once, eliminating some programming error, and scales to multiple threads for basic termination synchronization.
     1615For block allocation to arbitrary depth, including recursion, threads are created and destroyed in a lattice structure (tree with top and bottom).
    16021616Arbitrary topologies are possible using dynamic allocation, allowing threads to outlive their declaration scope, identical to normal dynamic allocation.
    16031617\begin{cfa}
     
    16701684\end{tabular}
    16711685\end{cquote}
    1672 Like coroutines, the @dtype@ property prevents \emph{implicit} copy operations and the @is_thread@ trait provides no \emph{explicit} copy operations, so threads must be passed by reference (pointer).
    1673 Similarly, the function definitions ensure there is a statically typed @main@ function that is the thread starting point (first stack frame), a mechanism to get (read) the thread descriptor from its handle, and a special destructor to prevent deallocation while the thread is executing.
     1686Like coroutines, the @dtype@ property prevents \emph{implicit} copy operations and the @is_thread@ trait provides no \emph{explicit} copy operations, so threads must be passed by reference or pointer.
     1687Similarly, the function definitions ensure there is a statically typed @main@ function that is the thread starting point (first stack frame), a mechanism to read the thread descriptor from its handle, and a special destructor to prevent deallocation while the thread is executing.
    16741688(The qualifier @mutex@ for the destructor parameter is discussed in Section~\ref{s:Monitor}.)
    16751689The difference between the coroutine and thread is that a coroutine borrows a thread from its caller, so the first thread resuming a coroutine creates the coroutine's stack and starts running the coroutine main on the stack;
    16761690whereas, a thread is scheduling for execution in @main@ immediately after its constructor is run.
    1677 No return value or additional parameters are necessary for this function because the @thread@ type allows an arbitrary number of interface functions with corresponding arbitrary typed input/output values.
     1691No return value or additional parameters are necessary for this function because the @thread@ type allows an arbitrary number of interface functions with corresponding arbitrary typed input and output values.
    16781692
    16791693
     
    16831697Unrestricted nondeterminism is meaningless as there is no way to know when a result is completed and safe to access.
    16841698To produce meaningful execution requires clawing back some determinism using mutual exclusion and synchronization, where mutual exclusion provides access control for threads using shared data, and synchronization is a timing relationship among threads~\cite[\S~4]{Buhr05a}.
    1685 The shared data protected by mutual exlusion is called a \newterm{critical section}~\cite{Dijkstra65}, and the protection can be simple (only 1 thread) or complex (only N kinds of threads, \eg group~\cite{Joung00} or readers/writer~\cite{Courtois71}).
    1686 Without synchronization control in a critical section, an arriving thread can barge ahead of preexisting waiter threads resulting in short/long-term starvation, staleness/freshness problems, and/or incorrect transfer of data.
     1699The shared data protected by mutual exlusion is called a \newterm{critical section}~\cite{Dijkstra65}, and the protection can be simple, only 1 thread, or complex, only N kinds of threads, \eg group~\cite{Joung00} or readers/writer~\cite{Courtois71} problems.
     1700Without synchronization control in a critical section, an arriving thread can barge ahead of preexisting waiter threads resulting in short/long-term starvation, staleness and freshness problems, and incorrect transfer of data.
    16871701Preventing or detecting barging is a challenge with low-level locks, but made easier through higher-level constructs.
    16881702This challenge is often split into two different approaches: barging \emph{avoidance} and \emph{prevention}.
     
    16961710Some concurrent systems eliminate mutable shared-state by switching to non-shared communication like message passing~\cite{Thoth,Harmony,V-Kernel,MPI} (Erlang, MPI), channels~\cite{CSP} (CSP,Go), actors~\cite{Akka} (Akka, Scala), or functional techniques (Haskell).
    16971711However, these approaches introduce a new communication mechanism for concurrency different from the standard communication using function call/return.
    1698 Hence, a programmer must learn and manipulate two sets of design/programming patterns.
     1712Hence, a programmer must learn and manipulate two sets of design and programming patterns.
    16991713While this distinction can be hidden away in library code, effective use of the library still has to take both paradigms into account.
    1700 In contrast, approaches based on shared-state models more closely resemble the standard call/return programming model, resulting in a single programming paradigm.
     1714In contrast, approaches based on shared-state models more closely resemble the standard call and return programming model, resulting in a single programming paradigm.
    17011715Finally, a newer approach for restricting non-determinism is transactional memory~\cite{Herlihy93}.
    17021716While this approach is pursued in hardware~\cite{Nakaike15} and system languages, like \CC~\cite{Cpp-Transactions}, the performance and feature set is still too restrictive~\cite{Cascaval08,Boehm09} to be the main concurrency paradigm for system languages.
     
    17111725For these reasons, \CFA selected monitors as the core high-level concurrency construct, upon which higher-level approaches can be easily constructed.
    17121726
    1713 Specifically, a \textbf{monitor} is a set of functions that ensure mutual exclusion when accessing shared state.
    1714 More precisely, a monitor is a programming technique that implicitly binds mutual exclusion to static function scope by call/return, as opposed to locks, where mutual-exclusion is defined by acquire/release calls, independent of lexical context (analogous to block and heap storage allocation).
    1715 Restricting acquire/release points eases programming, comprehension, and maintenance, at a slight cost in flexibility and efficiency.
    1716 \CFA uses a custom @monitor@ type and leverages declaration semantics (deallocation) to protect active or waiting threads in a monitor.
    1717 
    1718 The following is a \CFA monitor implementation of an atomic counter.
    1719 \begin{cfa}
    1720 `monitor` Aint { int cnt; }; $\C[4.25in]{// atomic integer counter}$
    1721 int ++?( Aint & `mutex` this ) with( this ) { return ++cnt; } $\C{// increment}$
    1722 int ?=?( Aint & `mutex` lhs, int rhs ) with( lhs ) { cnt = rhs; } $\C{// conversions with int, mutex optional}\CRT$
    1723 int ?=?( int & lhs, Aint & `mutex` rhs ) with( rhs ) { lhs = cnt; }
    1724 \end{cfa}
    1725 The operators use the parameter-only declaration type-qualifier @mutex@ to mark which parameters require locking during function execution to protect from race conditions.
    1726 The assignment operators provide bidirectional conversion between an atomic and normal integer without accessing field @cnt@.
    1727 (These operations only need @mutex@, if reading/writing the implementation type is not atomic.)
    1728 The atomic counter is used without any explicit mutual-exclusion and provides thread-safe semantics.
    1729 \begin{cfa}
     1727Figure~\ref{f:AtomicCounter} compares a \CFA and Java monitor implementing an atomic counter.\footnote{
     1728Like other concurrent programming languages, \CFA and Java have performant specializations for the basic types using atomic instructions.}
     1729A \newterm{monitor} is a set of functions that ensure mutual exclusion when accessing shared state.
     1730(Note, in \CFA, @monitor@ is short-hand for @mutex struct@.)
     1731More precisely, a monitor is a programming technique that implicitly binds mutual exclusion to static function scope by call and return, as opposed to locks, where mutual exclusion is defined by acquire/release calls, independent of lexical context (analogous to block and heap storage allocation).
     1732Restricting acquire and release points eases programming, comprehension, and maintenance, at a slight cost in flexibility and efficiency.
     1733As for other special types, \CFA has a custom @monitor@ type.
     1734
     1735\begin{figure}
     1736\centering
     1737
     1738\begin{lrbox}{\myboxA}
     1739\begin{cfa}[aboveskip=0pt,belowskip=0pt]
     1740`monitor` Aint { // atomic integer counter
     1741        int cnt;
     1742};
     1743int ++?( Aint & `mutex` this ) with(this) { return ++cnt; }
     1744int ?=?( Aint & `mutex` lhs, int rhs ) with(lhs) { cnt = rhs; }
     1745int ?=?(int & lhs, Aint & rhs) with(rhs) { lhs = cnt; }
     1746
    17301747int i = 0, j = 0, k = 5;
    1731 Aint x = { 0 }, y = { 0 }, z = { 5 }; $\C{// no mutex required}$
    1732 ++x; ++y; ++z; $\C{// safe increment by multiple threads}$
    1733 x = 2; y = i; z = k; $\C{// conversions}$
    1734 i = x; j = y; k = z;
    1735 \end{cfa}
    1736 Note, like other concurrent programming languages, \CFA has specializations for the basic types using atomic instructions for performance and a general trait similar to the \CC template @std::atomic@.
    1737 
    1738 \CFA monitors have \newterm{multi-acquire} semantics so the thread in the monitor may acquire it multiple times without deadlock, allowing recursion and calling other interface functions.
    1739 \newpage
    1740 \begin{cfa}
    1741 monitor M { ... } m;
    1742 void foo( M & mutex m ) { ... } $\C{// acquire mutual exclusion}$
    1743 void bar( M & mutex m ) { $\C{// acquire mutual exclusion}$
    1744         ... `bar( m );` ... `foo( m );` ... $\C{// reacquire mutual exclusion}$
    1745 }
    1746 \end{cfa}
    1747 \CFA monitors also ensure the monitor lock is released regardless of how an acquiring function ends (normal or exceptional), and returning a shared variable is safe via copying before the lock is released.
     1748Aint x = { 0 }, y = { 0 }, z = { 5 }; // no mutex
     1749++x; ++y; ++z;     // mutex
     1750x = 2; y = i; z = k;  // mutex
     1751i = x; j = y; k = z;  // no mutex
     1752\end{cfa}
     1753\end{lrbox}
     1754
     1755\begin{lrbox}{\myboxB}
     1756\begin{java}[aboveskip=0pt,belowskip=0pt]
     1757class Aint {
     1758    private int cnt;
     1759    public Aint( int init ) { cnt = init; }
     1760    `synchronized` public int inc() { return ++cnt; }
     1761    `synchronized` public void set( int rhs ) {cnt=rhs;}
     1762    public int get() { return cnt; }
     1763}
     1764int i = 0, j = 0, k = 5;
     1765Aint x=new Aint(0), y=new Aint(0), z=new Aint(5);
     1766x.inc(); y.inc(); z.inc();
     1767x.set( 2 ); y.set( i ); z.set( k );
     1768i = x.get(); j = y.get(); k = z.get();
     1769\end{java}
     1770\end{lrbox}
     1771
     1772\subfloat[\CFA]{\label{f:AtomicCounterCFA}\usebox\myboxA}
     1773\hspace{3pt}
     1774\vrule
     1775\hspace{3pt}
     1776\subfloat[Java]{\label{f:AtomicCounterJava}\usebox\myboxB}
     1777\caption{Atomic counter}
     1778\label{f:AtomicCounter}
     1779\end{figure}
     1780
     1781Like Java, \CFA monitors have \newterm{multi-acquire} semantics so the thread in the monitor may acquire it multiple times without deadlock, allowing recursion and calling other interface functions.
     1782% \begin{cfa}
     1783% monitor M { ... } m;
     1784% void foo( M & mutex m ) { ... } $\C{// acquire mutual exclusion}$
     1785% void bar( M & mutex m ) { $\C{// acquire mutual exclusion}$
     1786%       ... `bar( m );` ... `foo( m );` ... $\C{// reacquire mutual exclusion}$
     1787% }
     1788% \end{cfa}
     1789\CFA monitors also ensure the monitor lock is released regardless of how an acquiring function ends, normal or exceptional, and returning a shared variable is safe via copying before the lock is released.
    17481790Similar safety is offered by \emph{explicit} opt-in disciplines like \CC RAII versus the monitor \emph{implicit} language-enforced safety guarantee ensuring no programmer usage errors.
    1749 Furthermore, RAII mechanisms cannot handle complex synchronization within a monitor, where the monitor lock may not be released on function exit because it is passed to an unblocking thread;
     1791However, RAII mechanisms cannot handle complex synchronization within a monitor, where the monitor lock may not be released on function exit because it is passed to an unblocking thread;
    17501792RAII is purely a mutual-exclusion mechanism (see Section~\ref{s:Scheduling}).
     1793
     1794Both Java and \CFA use a keyword @mutex@/\lstinline[language=java]|synchronized| to designate functions that implicitly acquire/release the monitor lock on call/return providing mutual exclusion to the stared data.
     1795Non-designated functions provide no mutual exclusion for read-only access or as an interface to a multi-step protocol requiring several steps of acquiring and releasing the monitor.
     1796Monitor objects can be passed through multiple helper functions without acquiring mutual exclusion, until a designated function associated with the object is called.
     1797\CFA designated functions are marked by an explicitly parameter-only pointer/reference qualifier @mutex@ (discussed further in Section\ref{s:MutexAcquisition}).
     1798Whereas, Java designated members are marked with \lstinline[language=java]|synchronized| that applies to the implicit reference parameter @this@.
     1799In the example, the increment and setter operations need mutual exclusion while the read-only getter operation can be non-mutex if reading the implementation is atomic.
    17511800
    17521801
     
    17711820\end{tabular}
    17721821\end{cquote}
    1773 The @dtype@ property prevents \emph{implicit} copy operations and the @is_monitor@ trait provides no \emph{explicit} copy operations, so monitors must be passed by reference (pointer).
    1774 Similarly, the function definitions ensures there is a mechanism to get (read) the monitor descriptor from its handle, and a special destructor to prevent deallocation if a thread using the shared data.
     1822The @dtype@ property prevents \emph{implicit} copy operations and the @is_monitor@ trait provides no \emph{explicit} copy operations, so monitors must be passed by reference or pointer.
     1823Similarly, the function definitions ensures there is a mechanism to read the monitor descriptor from its handle, and a special destructor to prevent deallocation if a thread is using the shared data.
    17751824The custom monitor type also inserts any locks needed to implement the mutual exclusion semantics.
     1825\CFA relies heavily on traits as an abstraction mechanism, so the @mutex@ qualifier prevents coincidentally matching of a monitor trait with a type that is not a monitor, similar to coincidental inheritance where a shape and playing card can both be drawable.
    17761826
    17771827
     
    17791829\label{s:MutexAcquisition}
    17801830
    1781 While the monitor lock provides mutual exclusion for shared data, there are implementation options for when and where the locking/unlocking occurs.
    1782 (Much of this discussion also applies to basic locks.)
    1783 For example, a monitor may be passed through multiple helper functions before it is necessary to acquire the monitor's mutual exclusion.
    1784 
    1785 \CFA requires programmers to identify the kind of parameter with the @mutex@ keyword and uses no keyword to mean \lstinline[morekeywords=nomutex]@nomutex@, because @mutex@ parameters are rare and no keyword is the \emph{normal} parameter semantics.
    1786 Hence, @mutex@ parameters are documentation, at the function and its prototype, to both programmer and compiler, without other redundant keywords.
    1787 Furthermore, \CFA relies heavily on traits as an abstraction mechanism, so the @mutex@ qualifier prevents coincidentally matching of a monitor trait with a type that is not a monitor, similar to coincidental inheritance where a shape and playing card can both be drawable.
    1788 
    1789 The next semantic decision is establishing which parameter \emph{types} may be qualified with @mutex@.
    1790 The following has monitor parameter types that are composed of multiple objects.
    1791 \begin{cfa}
    1792 monitor M { ... }
     1831For object-oriented programming languages, the mutex property applies to one object, the implicit pointer/reference to the monitor type.
     1832Because \CFA uses a pointer qualifier, other possibilities exist, \eg:
     1833\begin{cfa}
     1834monitor M { ... };
    17931835int f1( M & mutex m ); $\C{// single parameter object}$
    17941836int f2( M * mutex m ); $\C{// single or multiple parameter object}$
     
    17961838int f4( stack( M * ) & mutex m ); $\C{// multiple parameters object}$
    17971839\end{cfa}
    1798 Function @f1@ has a single parameter object, while @f2@'s indirection could be a single or multi-element array, where static array size is often unknown in C.
    1799 Function @f3@ has a multiple object matrix, and @f4@ a multiple object data structure.
    1800 While shown shortly, multiple object acquisition is possible, but the number of objects must be statically known.
    1801 Therefore, \CFA only acquires one monitor per parameter with exactly one level of indirection, and exclude pointer types to unknown sized arrays.
    1802 
    1803 For object-oriented monitors, \eg Java, calling a mutex member \emph{implicitly} acquires mutual exclusion of the receiver object, @`rec`.foo(...)@.
    1804 \CFA has no receiver, and hence, the explicit @mutex@ qualifier is used to specify which objects acquire mutual exclusion.
    1805 A positive consequence of this design decision is the ability to support multi-monitor functions,\footnote{
     1840Function @f1@ has a single object parameter, while functions @f2@ to @f4@ can be a single or multi-element parameter with statically unknown size.
     1841Because of the statically unknown size, \CFA only supports a single reference @mutex@ parameter, @f1@.
     1842
     1843The \CFA @mutex@ qualifier does allow the ability to support multi-monitor functions,\footnote{
    18061844While object-oriented monitors can be extended with a mutex qualifier for multiple-monitor members, no prior example of this feature could be found.}
    1807 called \newterm{bulk acquire}.
     1845where the number of acquisitions is statically known, called \newterm{bulk acquire}.
    18081846\CFA guarantees bulk acquisition order is consistent across calls to @mutex@ functions using the same monitors as arguments, so acquiring multiple monitors in a bulk acquire is safe from deadlock.
    18091847Figure~\ref{f:BankTransfer} shows a trivial solution to the bank transfer problem~\cite{BankTransfer}, where two resources must be locked simultaneously, using \CFA monitors with implicit locking and \CC with explicit locking.
     
    19331971% There are many aspects of scheduling in a concurrency system, all related to resource utilization by waiting threads, \ie which thread gets the resource next.
    19341972% Different forms of scheduling include access to processors by threads (see Section~\ref{s:RuntimeStructureCluster}), another is access to a shared resource by a lock or monitor.
    1935 This section discusses scheduling for waiting threads eligible for monitor entry, \ie which user thread gets the shared resource next. (See Section~\ref{s:RuntimeStructureCluster} for scheduling kernel threads on virtual processors.)
     1973This section discusses scheduling for waiting threads eligible for monitor entry~\cite{Buhr95b}, \ie which user thread gets the shared resource next. (See Section~\ref{s:RuntimeStructureCluster} for scheduling kernel threads on virtual processors.)
    19361974While monitor mutual-exclusion provides safe access to its shared data, the data may indicate a thread cannot proceed, \eg a bounded buffer may be full/\-empty so produce/consumer threads must block.
    19371975Leaving the monitor and retrying (busy waiting) is impractical for high-level programming.
     
    19391977Monitors eliminate busy waiting by providing synchronization within the monitor critical-section to schedule threads needing access to the shared data, where threads block versus spin.
    19401978Synchronization is generally achieved with internal~\cite{Hoare74} or external~\cite[\S~2.9.2]{uC++} scheduling.
    1941 \newterm{Internal} (largely) schedules threads located \emph{inside} the monitor and is accomplished using condition variables with signal and wait.
    1942 \newterm{External} (largely) schedules threads located \emph{outside} the monitor and is accomplished with the @waitfor@ statement.
    1943 Note, internal scheduling has a small amount of external scheduling and vice versus, so the naming denotes where the majority of the block threads reside (inside or outside) for scheduling.
    1944 For complex scheduling, the approaches can be combined, so there can be an equal number of threads waiting inside and outside.
    1945 
    1946 \CFA monitors do not allow calling threads to barge ahead of signalled threads (via barging prevention), which simplifies synchronization among threads in the monitor and increases correctness.
     1979\newterm{Internal} largely schedules threads located \emph{inside} the monitor and is accomplished using condition variables with signal and wait.
     1980\newterm{External} largely schedules threads located \emph{outside} the monitor and is accomplished with the @waitfor@ statement.
     1981Note, internal scheduling has a small amount of external scheduling and vice versa, so the naming denotes where the majority of the block threads reside (inside or outside) for scheduling.
     1982For complex scheduling, the approaches can be combined, so there are threads waiting inside and outside.
     1983
     1984\CFA monitors do not allow calling threads to barge ahead of signalled threads via barging prevention, which simplifies synchronization among threads in the monitor and increases correctness.
    19471985A direct consequence of this semantics is that unblocked waiting threads are not required to recheck the waiting condition, \ie waits are not in a starvation-prone busy-loop as required by the signals-as-hints style with barging.
    19481986Preventing barging comes directly from Hoare's semantics in the seminal paper on monitors~\cite[p.~550]{Hoare74}.
     
    19531991Furthermore, \CFA concurrency has no spurious wakeup~\cite[\S~9]{Buhr05a}, which eliminates an implicit self barging.
    19541992
    1955 Monitor mutual-exclusion means signalling cannot have the signaller and signalled thread in the monitor simultaneously, so only the signaller or signallee can proceed.
    1956 Figure~\ref{f:MonitorScheduling} shows internal/external scheduling for the bounded-buffer examples in Figure~\ref{f:GenericBoundedBuffer}.
    1957 For internal scheduling in Figure~\ref{f:BBInt}, the @signal@ moves the signallee (front thread of the specified condition queue) to urgent and the signaller continues (solid line).
     1993Monitor mutual-exclusion means signalling cannot have the signaller and signalled thread in the monitor simultaneously, so only the signaller or signallee can proceed and the other waits on an implicit urgent list~\cite[p.~551]{Hoare74}.
     1994Figure~\ref{f:MonitorScheduling} shows internal and external scheduling for the bounded-buffer examples in Figure~\ref{f:GenericBoundedBuffer}.
     1995For internal scheduling in Figure~\ref{f:BBInt}, the @signal@ moves the signallee, front thread of the specified condition queue, to the urgent list (see Figure~\ref{f:MonitorScheduling}) and the signaller continues (solid line).
    19581996Multiple signals move multiple signallees to urgent until the condition queue is empty.
    1959 When the signaller exits or waits, a thread is implicitly unblocked from urgent (if available) before unblocking a calling thread to prevent barging.
     1997When the signaller exits or waits, a thread is implicitly unblocked from urgent, if available, before unblocking a calling thread to prevent barging.
    19601998(Java conceptually moves the signalled thread to the calling queue, and hence, allows barging.)
    1961 Signal is used when the signaller is providing the cooperation needed by the signallee (\eg creating an empty slot in a buffer for a producer) and the signaller immediately exits the monitor to run concurrently (consume the buffer element) and passes control of the monitor to the signalled thread, which can immediately take advantage of the state change.
     1999Signal is used when the signaller is providing the cooperation needed by the signallee, \eg creating an empty slot in a buffer for a producer, and the signaller immediately exits the monitor to run concurrently consuming the buffer element, and passes control of the monitor to the signalled thread, which can immediately take advantage of the state change.
    19622000Specifically, the @wait@ function atomically blocks the calling thread and implicitly releases the monitor lock(s) for all monitors in the function's parameter list.
    19632001Signalling is unconditional because signalling an empty condition queue does nothing.
    19642002It is common to declare condition queues as monitor fields to prevent shared access, hence no locking is required for access as the queues are protected by the monitor lock.
    1965 In \CFA, a condition queue can be created/stored independently.
     2003In \CFA, a condition queue can be created and stored independently.
    19662004
    19672005\begin{figure}
     
    20492087\end{figure}
    20502088
    2051 The @signal_block@ provides the opposite unblocking order, where the signaller is moved to urgent and the signallee continues and a thread is implicitly unblocked from urgent when the signallee exits or waits (dashed line).
    2052 Signal block is used when the signallee is providing the cooperation needed by the signaller (\eg if the buffer is removed and a producer hands off an item to a consumer, as in Figure~\ref{f:DatingSignalBlock}) so the signaller must wait until the signallee unblocks, provides the cooperation, exits the monitor to run concurrently, and passes control of the monitor to the signaller, which can immediately take advantage of the state change.
     2089The @signal_block@ provides the opposite unblocking order, where the signaller is moved to urgent and the signallee continues and a thread is implicitly unblocked from urgent when the signallee exits or waits (dashed line)~\cite[p.~551]{Hoare74}.
     2090Signal block is used when the signallee is providing the cooperation needed by the signaller, \eg if the buffer is removed and a producer hands off an item to a consumer as in Figure~\ref{f:DatingSignalBlock}, so the signaller must wait until the signallee unblocks, provides the cooperation, exits the monitor to run concurrently, and passes control of the monitor to the signaller, which can immediately take advantage of the state change.
    20532091Using @signal@ or @signal_block@ can be a dynamic decision based on whether the thread providing the cooperation arrives before or after the thread needing the cooperation.
    20542092
    2055 External scheduling in Figure~\ref{f:BBExt} simplifies internal scheduling by eliminating condition queues and @signal@/@wait@ (cases where it cannot are discussed shortly), and has existed in the programming language Ada for almost 40 years with variants in other languages~\cite{SR,ConcurrentC++,uC++}.
     2093For external scheduling in Figure~\ref{f:BBExt}, the internal scheduling is replaced, eliminating condition queues and @signal@/@wait@ (cases where it cannot are discussed shortly), and has existed in the programming language Ada for almost 40 years with variants in other languages~\cite{SR,ConcurrentC++,uC++}.
    20562094While prior languages use external scheduling solely for thread interaction, \CFA generalizes it to both monitors and threads.
    20572095External scheduling allows waiting for events from other threads while restricting unrelated events, that would otherwise have to wait on condition queues in the monitor.
     
    20622100Now when a producer/consumer detects a full/empty buffer, the necessary cooperation for continuation is specified by indicating the next function call that can occur.
    20632101For example, a producer detecting a full buffer must have cooperation from a consumer to remove an item so function @remove@ is accepted, which prevents producers from entering the monitor, and after a consumer calls @remove@, the producer waiting on urgent is \emph{implicitly} unblocked because it can now continue its insert operation.
    2064 Hence, this mechanism is done in terms of control flow, next call, versus in terms of data, channels, as in Go/Rust @select@.
     2102Hence, this mechanism is done in terms of control flow, next call, versus in terms of data, channels, as in Go and Rust @select@.
    20652103While both mechanisms have strengths and weaknesses, \CFA uses the control-flow mechanism to be consistent with other language features.
    20662104
    2067 Figure~\ref{f:ReadersWriterLock} shows internal/external scheduling for a readers/writer lock with no barging and threads are serviced in FIFO order to eliminate staleness/freshness among the reader/writer threads.
     2105Figure~\ref{f:ReadersWriterLock} shows internal and external scheduling for a readers/writer lock with no barging and threads are serviced in FIFO order to eliminate staleness and freshness among the reader/writer threads.
    20682106For internal scheduling in Figure~\ref{f:RWInt}, the readers and writers wait on the same condition queue in FIFO order, making it impossible to tell if a waiting thread is a reader or writer.
    20692107To clawback the kind of thread, a \CFA condition can store user data in the node for a blocking thread at the @wait@, \ie whether the thread is a @READER@ or @WRITER@.
     
    22292267For signal scheduling, the @exchange@ condition is necessary to block the thread finding the match, while the matcher unblocks to take the opposite number, post its phone number, and unblock the partner.
    22302268For signal-block scheduling, the implicit urgent-queue replaces the explicit @exchange@-condition and @signal_block@ puts the finding thread on the urgent stack and unblocks the matcher.
    2231 
    2232 The dating service is an important example of a monitor that cannot be written using external scheduling.
    2233 First, because scheduling requires knowledge of calling parameters to make matching decisions, and parameters of calling threads are unavailable within the monitor.
    2234 For example, a girl thread within the monitor cannot examine the @ccode@ of boy threads waiting on the calling queue to determine if there is a matching partner.
    2235 Second, because a scheduling decision may be delayed when there is no immediate match, which requires a condition queue for waiting, and condition queues imply internal scheduling.
    2236 For example, if a girl thread could determine there is no calling boy with the same @ccode@, it must wait until a matching boy arrives.
    2237 Finally, barging corrupts the dating service during an exchange because a barger may also match and change the phone numbers, invalidating the previous exchange phone number.
     2269Note, barging corrupts the dating service during an exchange because a barger may also match and change the phone numbers, invalidating the previous exchange phone number.
    22382270This situation shows rechecking the waiting condition and waiting again (signals-as-hints) fails, requiring significant restructured to account for barging.
    22392271
    2240 Both internal and external scheduling extend to multiple monitors in a natural way.
     2272Given external and internal scheduling, what guidelines can a programmer use to select between them?
     2273In general, external scheduling is easier to understand and code because only the next logical action (mutex function(s)) is stated, and the monitor implicitly handles all the details.
     2274Therefore, there are no condition variables, and hence, no wait and signal, which reduces coding complexity and synchronization errors.
     2275If external scheduling is simpler than internal, why not use it all the time?
     2276Unfortunately, external scheduling cannot be used if: scheduling depends on parameter value(s) or scheduling must block across an unknown series of calls on a condition variable, \ie internal scheduling.
     2277For example, the dating service cannot be written using external scheduling.
     2278First, scheduling requires knowledge of calling parameters to make matching decisions and parameters of calling threads are unavailable within the monitor.
     2279Specifically, a thread within the monitor cannot examine the @ccode@ of threads waiting on the calling queue to determine if there is a matching partner.
     2280(Similarly, if the bounded buffer or readers/writer are restructured with a single interface function with a parameter denoting producer/consumer or reader/write, they cannot be solved with external scheduling.)
     2281Second, a scheduling decision may be delayed across an unknown number of calls when there is no immediate match so the thread in the monitor must block on a condition.
     2282Specifically, if a thread determines there is no opposite calling thread with the same @ccode@, it must wait an unknown period until a matching thread arrives.
     2283For complex synchronization, both external and internal scheduling can be used to take advantage of best of properties of each.
     2284
     2285Finally, both internal and external scheduling extend to multiple monitors in a natural way.
    22412286\begin{cquote}
    22422287\begin{tabular}{@{}l@{\hspace{2\parindentlnth}}l@{}}
     
    22742319Similarly, for @waitfor( rtn )@, the default semantics is to atomically block the acceptor and release all acquired mutex parameters, \ie @waitfor( rtn : m1, m2 )@.
    22752320To override the implicit multi-monitor wait, specific mutex parameter(s) can be specified, \eg @waitfor( rtn : m1 )@.
    2276 @waitfor@ does statically verify the monitor types passed are the same as the acquired mutex-parameters of the given function or function pointer, hence the function (pointer) prototype must be accessible.
     2321@waitfor@ does statically verify the monitor types passed are the same as the acquired mutex-parameters of the given function or function pointer, hence the prototype must be accessible.
    22772322% When an overloaded function appears in an @waitfor@ statement, calls to any function with that name are accepted.
    2278 % The rationale is that members with the same name should perform a similar function, and therefore, all should be eligible to accept a call.
     2323% The rationale is that functions with the same name should perform a similar actions, and therefore, all should be eligible to accept a call.
    22792324Overloaded functions can be disambiguated using a cast
    22802325\begin{cfa}
     
    22852330
    22862331The ability to release a subset of acquired monitors can result in a \newterm{nested monitor}~\cite{Lister77} deadlock (see Section~\ref{s:MutexAcquisition}).
    2287 \newpage
    22882332\begin{cfa}
    22892333void foo( M & mutex m1, M & mutex m2 ) {
     
    23002344
    23012345Figure~\ref{f:ExtendedWaitfor} shows the extended form of the @waitfor@ statement to conditionally accept one of a group of mutex functions, with an optional statement to be performed \emph{after} the mutex function finishes.
    2302 For a @waitfor@ clause to be executed, its @when@ must be true and an outstanding call to its corresponding member(s) must exist.
     2346For a @waitfor@ clause to be executed, its @when@ must be true and an outstanding call to its corresponding function(s) must exist.
    23032347The \emph{conditional-expression} of a @when@ may call a function, but the function must not block or context switch.
    2304 If there are multiple acceptable mutex calls, selection occurs top-to-bottom (prioritized) among the @waitfor@ clauses, whereas some programming languages with similar mechanisms accept nondeterministically for this case, \eg Go \lstinline[morekeywords=select]@select@.
    2305 If some accept guards are true and there are no outstanding calls to these members, the acceptor is blocked until a call to one of these members is made.
     2348If there are multiple acceptable mutex calls, selection is prioritized top-to-bottom among the @waitfor@ clauses, whereas some programming languages with similar mechanisms accept nondeterministically for this case, \eg Go \lstinline[morekeywords=select]@select@.
     2349If some accept guards are true and there are no outstanding calls to these functions, the acceptor is blocked until a call to one of these functions is made.
    23062350If there is a @timeout@ clause, it provides an upper bound on waiting.
    23072351If all the accept guards are false, the statement does nothing, unless there is a terminating @else@ clause with a true guard, which is executed instead.
    23082352Hence, the terminating @else@ clause allows a conditional attempt to accept a call without blocking.
    23092353If both @timeout@ and @else@ clause are present, the @else@ must be conditional, or the @timeout@ is never triggered.
    2310 There is also a traditional future wait queue (not shown) (\eg Microsoft @WaitForMultipleObjects@), to wait for a specified number of future elements in the queue.
     2354% There is also a traditional future wait queue (not shown) (\eg Microsoft @WaitForMultipleObjects@), to wait for a specified number of future elements in the queue.
    23112355Finally, there is a shorthand for specifying multiple functions using the same set of monitors: @waitfor( f, g, h : m1, m2, m3 )@.
    23122356
     
    23152359\begin{cfa}
    23162360`when` ( $\emph{conditional-expression}$ )      $\C{// optional guard}$
    2317         waitfor( $\emph{mutex-member-name}$ ) $\emph{statement}$ $\C{// action after call}$
     2361        waitfor( $\emph{mutex-function-name}$ ) $\emph{statement}$ $\C{// action after call}$
    23182362`or` `when` ( $\emph{conditional-expression}$ ) $\C{// any number of functions}$
    2319         waitfor( $\emph{mutex-member-name}$ ) $\emph{statement}$
     2363        waitfor( $\emph{mutex-function-name}$ ) $\emph{statement}$
    23202364`or`    ...
    23212365`when` ( $\emph{conditional-expression}$ ) $\C{// optional guard}$
     
    23352379The left example only accepts @mem1@ if @C1@ is true or only @mem2@ if @C2@ is true.
    23362380The right example accepts either @mem1@ or @mem2@ if @C1@ and @C2@ are true.
     2381Hence, the @waitfor@ has parallel semantics, accepting any true @when@ clause.
    23372382
    23382383An interesting use of @waitfor@ is accepting the @mutex@ destructor to know when an object is deallocated, \eg assume the bounded buffer is restructured from a monitor to a thread with the following @main@.
     
    24322477If W1 waited first, the signaller must retain @m1@ amd @m2@ until completion of the outer mutex statement and then pass both to W1.
    24332478% Furthermore, there is an execution sequence where the signaller always finds waiter W2, and hence, waiter W1 starves.
    2434 To support this efficient semantics (and prevent barging), the implementation maintains a list of monitors acquired for each blocked thread.
    2435 When a signaller exits or waits in a monitor function/statement, the front waiter on urgent is unblocked if all its monitors are released.
     2479To support this efficient semantics and prevent barging, the implementation maintains a list of monitors acquired for each blocked thread.
     2480When a signaller exits or waits in a mutex function or statement, the front waiter on urgent is unblocked if all its monitors are released.
    24362481Implementing a fast subset check for the necessary released monitors is important and discussed in the following sections.
    24372482% The benefit is encapsulating complexity into only two actions: passing monitors to the next owner when they should be released and conditionally waking threads if all conditions are met.
     
    24422487
    24432488In a statically-typed object-oriented programming language, a class has an exhaustive list of members, even when members are added via static inheritance (see Figure~\ref{f:uCinheritance}).
    2444 Knowing all members at compilation (even separate compilation) allows uniquely numbered them so the accept-statement implementation can use a fast/compact bit mask with $O(1)$ compare.
     2489Knowing all members at compilation, even separate compilation, allows uniquely numbered them so the accept-statement implementation can use a fast and compact bit mask with $O(1)$ compare.
    24452490
    24462491\begin{figure}
     
    24932538Hence, function pointers are used to identify the functions listed in the @waitfor@ statement, stored in a variable-sized array.
    24942539Then, the same implementation approach used for the urgent stack (see Section~\ref{s:Scheduling}) is used for the calling queue.
    2495 Each caller has a list of monitors acquired, and the @waitfor@ statement performs a (short) linear search matching functions in the @waitfor@ list with called functions, and then verifying the associated mutex locks can be transfers.
     2540Each caller has a list of monitors acquired, and the @waitfor@ statement performs a short linear search matching functions in the @waitfor@ list with called functions, and then verifying the associated mutex locks can be transfers.
    24962541
    24972542
     
    25712616
    25722617struct Msg { int i, j; };
    2573 monitor thread GoRtn { int i;  float f;  Msg m; };
     2618mutex thread GoRtn { int i;  float f;  Msg m; };
    25742619void mem1( GoRtn & mutex gortn, int i ) { gortn.i = i; }
    25752620void mem2( GoRtn & mutex gortn, float f ) { gortn.f = f; }
     
    25772622void ^?{}( GoRtn & mutex ) {}
    25782623
    2579 void main( GoRtn & gortn ) with( gortn ) { // thread starts
     2624void main( GoRtn & mutex gortn ) with(gortn) { // thread starts
    25802625
    25812626        for () {
     
    26442689
    26452690\begin{cfa}
    2646 monitor thread DatingService {
     2691mutex thread DatingService {
    26472692        condition Girls[CompCodes], Boys[CompCodes];
    26482693        int girlPhoneNo, boyPhoneNo, ccode;
     
    27082753% \label{f:pingpong}
    27092754% \end{figure}
    2710 Note, the ping/pong threads are globally declared, @pi@/@po@, and hence, start (and possibly complete) before the program main starts.
     2755Note, the ping/pong threads are globally declared, @pi@/@po@, and hence, start and possibly complete before the program main starts.
    27112756\end{comment}
    27122757
    27132758
    2714 \subsection{\texorpdfstring{\protect\lstinline@monitor@ Generators / Coroutines / Threads}{monitor Generators / Coroutines / Threads}}
    2715 
    2716 \CFA generators, coroutines, and threads can also be monitors (Table~\ref{t:ExecutionPropertyComposition} cases 4, 6, 12) allowing safe \emph{direct communication} with threads, \ie the custom types can have mutex functions that are called by other threads.
     2759\subsection{\texorpdfstring{\protect\lstinline@mutex@ Generators / Coroutines / Threads}{monitor Generators / Coroutines / Threads}}
     2760
     2761\CFA generators, coroutines, and threads can also be @mutex@ (Table~\ref{t:ExecutionPropertyComposition} cases 4, 6, 12) allowing safe \emph{direct communication} with threads, \ie the custom types can have mutex functions that are called by other threads.
    27172762All monitor features are available within these mutex functions.
    2718 For example, if the formatter generator (or coroutine equivalent) in Figure~\ref{f:CFAFormatGen} is extended with the monitor property and this interface function is used to communicate with the formatter:
     2763For example, if the formatter generator or coroutine equivalent in Figure~\ref{f:CFAFormatGen} is extended with the monitor property and this interface function is used to communicate with the formatter:
    27192764\begin{cfa}
    27202765void fmt( Fmt & mutex fmt, char ch ) { fmt.ch = ch; resume( fmt ) }
     
    27242769Figure~\ref{f:DirectCommunicationComparison} shows a comparison of direct call-communication in \CFA versus indirect channel-communication in Go.
    27252770(Ada has a similar mechanism to \CFA direct communication.)
    2726 The program thread in \CFA @main@ uses the call/return paradigm to directly communicate with the @GoRtn main@, whereas Go switches to the channel paradigm to indirectly communicate with the goroutine.
     2771% The thread main function is by default @mutex@, so the @mutex@ qualifier for the thread parameter is optional.
     2772% The reason is that the thread logically starts instantaneously in the thread main acquiring its mutual exclusion, so it starts before any calls to prepare for synchronizing these calls.
     2773The \CFA program @main@ uses the call/return paradigm to directly communicate with the @GoRtn main@, whereas Go switches to the unbuffered channel paradigm to indirectly communicate with the goroutine.
    27272774Communication by multiple threads is safe for the @gortn@ thread via mutex calls in \CFA or channel assignment in Go.
     2775The different between call and channel send occurs for buffered channels making the send asynchronous.
     2776In \CFA, asynchronous call and multiple buffers is provided using an administrator and worker threads~\cite{Gentleman81} and/or futures (not discussed).
    27282777
    27292778Figure~\ref{f:DirectCommunicationDatingService} shows the dating-service problem in Figure~\ref{f:DatingServiceMonitor} extended from indirect monitor communication to direct thread communication.
    2730 When converting a monitor to a thread (server), the coding pattern is to move as much code as possible from the accepted members into the thread main so it does an much work as possible.
     2779When converting a monitor to a thread (server), the coding pattern is to move as much code as possible from the accepted functions into the thread main so it does an much work as possible.
    27312780Notice, the dating server is postponing requests for an unspecified time while continuing to accept new requests.
    2732 For complex servers (web-servers), there can be hundreds of lines of code in the thread main and safe interaction with clients can be complex.
     2781For complex servers, \eg web-servers, there can be hundreds of lines of code in the thread main and safe interaction with clients can be complex.
    27332782
    27342783
     
    27682817
    27692818In contrast to direct threading is indirect \newterm{thread pools}, \eg Java @executor@, where small jobs (work units) are inserted into a work pool for execution.
    2770 If the jobs are dependent, \ie interact, there is an implicit/explicit dependency graph that ties them together.
     2819If the jobs are dependent, \ie interact, there is an implicit dependency graph that ties them together.
    27712820While removing direct concurrency, and hence the amount of context switching, thread pools significantly limit the interaction that can occur among jobs.
    27722821Indeed, jobs should not block because that also blocks the underlying thread, which effectively means the CPU utilization, and therefore throughput, suffers.
     
    28572906\label{s:RuntimeStructureProcessor}
    28582907
    2859 A virtual processor is implemented by a kernel thread (\eg UNIX process), which are scheduled for execution on a hardware processor by the underlying operating system.
     2908A virtual processor is implemented by a kernel thread, \eg UNIX process, which are scheduled for execution on a hardware processor by the underlying operating system.
    28602909Programs may use more virtual processors than hardware processors.
    28612910On a multiprocessor, kernel threads are distributed across the hardware processors resulting in virtual processors executing in parallel.
     
    28722921\label{s:Implementation}
    28732922
    2874 A primary implementation challenge is avoiding contention from dynamically allocating memory because of bulk acquire, \eg the internal-scheduling design is (almost) free of allocations.
     2923A primary implementation challenge is avoiding contention from dynamically allocating memory because of bulk acquire, \eg the internal-scheduling design is almost free of allocations.
    28752924All blocking operations are made by parking threads onto queues, therefore all queues are designed with intrusive nodes, where each node has preallocated link fields for chaining.
    28762925Furthermore, several bulk-acquire operations need a variable amount of memory.
     
    29182967
    29192968There are two versions of the \CFA runtime kernel: debug and non-debug.
    2920 The debugging version has many runtime checks and internal assertions, \eg stack (non-writable) guard page, and checks for stack overflow whenever context switches occur among coroutines and threads, which catches most stack overflows.
     2969The debugging version has many runtime checks and internal assertions, \eg stack non-writable guard page, and checks for stack overflow whenever context switches occur among coroutines and threads, which catches most stack overflows.
    29212970After a program is debugged, the non-debugging version can be used to significantly decrease space and increase performance.
    29222971
     
    29262975
    29272976To test the performance of the \CFA runtime, a series of microbenchmarks are used to compare \CFA with pthreads, Java 11.0.6, Go 1.12.6, Rust 1.37.0, Python 3.7.6, Node.js 12.14.1, and \uC 7.0.0.
    2928 For comparison, the package must be multi-processor (M:N), which excludes libdill/libmil~\cite{libdill} (M:1)), and use a shared-memory programming model, \eg not message passing.
     2977For comparison, the package must be multi-processor (M:N), which excludes libdil and /libmil~\cite{libdill} (M:1)), and use a shared-memory programming model, \eg not message passing.
    29292978The benchmark computer is an AMD Opteron\texttrademark\ 6380 NUMA 64-core, 8 socket, 2.5 GHz processor, running Ubuntu 16.04.6 LTS, and pthreads/\CFA/\uC are compiled with gcc 9.2.1.
    29302979
     
    29402989% tar --exclude-ignore=exclude -cvhf benchmark.tar benchmark
    29412990
     2991\paragraph{Creation}
     2992
     2993Creation is measured by creating and deleting a specific kind of control-flow object.
     2994Figure~\ref{f:creation} shows the code for \CFA with results in Table~\ref{t:creation}.
     2995Note, the call stacks of \CFA coroutines are lazily created on the first resume, therefore the cost of creation with and without a stack are presented.
     2996
     2997\begin{multicols}{2}
     2998\lstset{language=CFA,moredelim=**[is][\color{red}]{@}{@},deletedelim=**[is][]{`}{`}}
     2999\begin{cfa}
     3000@coroutine@ MyCoroutine {};
     3001void ?{}( MyCoroutine & this ) {
     3002#ifdef EAGER
     3003        resume( this );
     3004#endif
     3005}
     3006void main( MyCoroutine & ) {}
     3007int main() {
     3008        BENCH( for ( N ) { @MyCoroutine c;@ } )
     3009        sout | result;
     3010}
     3011\end{cfa}
     3012\captionof{figure}{\CFA creation benchmark}
     3013\label{f:creation}
     3014
     3015\columnbreak
     3016
     3017\vspace*{-16pt}
     3018\captionof{table}{Creation comparison (nanoseconds)}
     3019\label{t:creation}
     3020
     3021\begin{tabular}[t]{@{}r*{3}{D{.}{.}{5.2}}@{}}
     3022\multicolumn{1}{@{}c}{} & \multicolumn{1}{c}{Median} & \multicolumn{1}{c}{Average} & \multicolumn{1}{c@{}}{Std Dev} \\
     3023\CFA generator                  & 0.6           & 0.6           & 0.0           \\
     3024\CFA coroutine lazy             & 13.4          & 13.1          & 0.5           \\
     3025\CFA coroutine eager    & 144.7         & 143.9         & 1.5           \\
     3026\CFA thread                             & 466.4         & 468.0         & 11.3          \\
     3027\uC coroutine                   & 155.6         & 155.7         & 1.7           \\
     3028\uC thread                              & 523.4         & 523.9         & 7.7           \\
     3029Python generator                & 123.2         & 124.3         & 4.1           \\
     3030Node.js generator               & 32.3          & 32.2          & 0.3           \\
     3031Goroutine thread                & 751.0         & 750.5         & 3.1           \\
     3032Rust tokio thread               & 1860.0        & 1881.1        & 37.6          \\
     3033Rust thread                             & 53801.0       & 53896.8       & 274.9         \\
     3034Java thread                             & 120274.0      & 120722.9      & 2356.7        \\
     3035Pthreads thread                 & 31465.5       & 31419.5       & 140.4
     3036\end{tabular}
     3037\end{multicols}
     3038
    29423039\paragraph{Context Switching}
    29433040
    29443041In procedural programming, the cost of a function call is important as modularization (refactoring) increases.
    29453042(In many cases, a compiler inlines function calls to increase the size and number of basic blocks for optimizing.)
    2946 Similarly, when modularization extends to coroutines/threads, the time for a context switch becomes a relevant factor.
     3043Similarly, when modularization extends to coroutines and threads, the time for a context switch becomes a relevant factor.
    29473044The coroutine test is from resumer to suspender and from suspender to resumer, which is two context switches.
    29483045%For async-await systems, the test is scheduling and fulfilling @N@ empty promises, where all promises are allocated before versus interleaved with fulfillment to avoid garbage collection.
     
    29503047The thread test is using yield to enter and return from the runtime kernel, which is two context switches.
    29513048The difference in performance between coroutine and thread context-switch is the cost of scheduling for threads, whereas coroutines are self-scheduling.
    2952 Figure~\ref{f:ctx-switch} shows the \CFA code for a coroutine/thread with results in Table~\ref{t:ctx-switch}.
     3049Figure~\ref{f:ctx-switch} shows the \CFA code for a coroutine and thread with results in Table~\ref{t:ctx-switch}.
    29533050
    29543051% From: Gregor Richards <gregor.richards@uwaterloo.ca>
     
    29963093\uC thread                      & 97.3          & 97.4          & 1.0   \\
    29973094Python generator        & 40.9          & 41.3          & 1.5   \\
     3095Node.js await           & 1852.2        & 1854.7        & 16.4  \\
    29983096Node.js generator       & 32.6          & 32.2          & 1.0   \\
    2999 Node.js await           & 1852.2        & 1854.7        & 16.4  \\
    30003097Goroutine thread        & 143.0         & 143.3         & 1.1   \\
     3098Rust async await        & 32.0          & 32.0          & 0.0   \\
     3099Rust tokio thread       & 143.0         & 143.0         & 1.7   \\
    30013100Rust thread                     & 332.0         & 331.4         & 2.4   \\
    30023101Java thread                     & 405.0         & 415.0         & 17.6  \\
     
    30053104\end{multicols}
    30063105
     3106\vspace*{-10pt}
    30073107\paragraph{Internal Scheduling}
    30083108
     
    30363136}
    30373137\end{cfa}
     3138\vspace*{-8pt}
    30383139\captionof{figure}{\CFA Internal-scheduling benchmark}
    30393140\label{f:schedint}
     
    31043205\paragraph{Mutual-Exclusion}
    31053206
    3106 Uncontented mutual exclusion, which frequently occurs, is measured by entering/leaving a critical section.
    3107 For monitors, entering and leaving a monitor function is measured, otherwise the language-appropriate mutex-lock is measured.
     3207Uncontented mutual exclusion, which frequently occurs, is measured by entering and leaving a critical section.
     3208For monitors, entering and leaving a mutex function is measured, otherwise the language-appropriate mutex-lock is measured.
    31083209For comparison, a spinning (versus blocking) test-and-test-set lock is presented.
    31093210Figure~\ref{f:mutex} shows the code for \CFA with results in Table~\ref{t:mutex}.
     
    31423243\end{multicols}
    31433244
    3144 \paragraph{Creation}
    3145 
    3146 Creation is measured by creating/deleting a specific kind of control-flow object.
    3147 Figure~\ref{f:creation} shows the code for \CFA with results in Table~\ref{t:creation}.
    3148 Note, the call stacks of \CFA coroutines are lazily created on the first resume, therefore the cost of creation with and without a stack are presented.
    3149 
    3150 \begin{multicols}{2}
    3151 \lstset{language=CFA,moredelim=**[is][\color{red}]{@}{@},deletedelim=**[is][]{`}{`}}
    3152 \begin{cfa}
    3153 @coroutine@ MyCoroutine {};
    3154 void ?{}( MyCoroutine & this ) {
    3155 #ifdef EAGER
    3156         resume( this );
    3157 #endif
    3158 }
    3159 void main( MyCoroutine & ) {}
    3160 int main() {
    3161         BENCH( for ( N ) { @MyCoroutine c;@ } )
    3162         sout | result;
    3163 }
    3164 \end{cfa}
    3165 \captionof{figure}{\CFA creation benchmark}
    3166 \label{f:creation}
    3167 
    3168 \columnbreak
    3169 
    3170 \vspace*{-16pt}
    3171 \captionof{table}{Creation comparison (nanoseconds)}
    3172 \label{t:creation}
    3173 
    3174 \begin{tabular}[t]{@{}r*{3}{D{.}{.}{5.2}}@{}}
    3175 \multicolumn{1}{@{}c}{} & \multicolumn{1}{c}{Median} & \multicolumn{1}{c}{Average} & \multicolumn{1}{c@{}}{Std Dev} \\
    3176 \CFA generator                  & 0.6           & 0.6           & 0.0           \\
    3177 \CFA coroutine lazy             & 13.4          & 13.1          & 0.5           \\
    3178 \CFA coroutine eager    & 144.7         & 143.9         & 1.5           \\
    3179 \CFA thread                             & 466.4         & 468.0         & 11.3          \\
    3180 \uC coroutine                   & 155.6         & 155.7         & 1.7           \\
    3181 \uC thread                              & 523.4         & 523.9         & 7.7           \\
    3182 Python generator                & 123.2         & 124.3         & 4.1           \\
    3183 Node.js generator               & 32.3          & 32.2          & 0.3           \\
    3184 Goroutine thread                & 751.0         & 750.5         & 3.1           \\
    3185 Rust thread                             & 53801.0       & 53896.8       & 274.9         \\
    3186 Java thread                             & 120274.0      & 120722.9      & 2356.7        \\
    3187 Pthreads thread                 & 31465.5       & 31419.5       & 140.4
    3188 \end{tabular}
    3189 \end{multicols}
    3190 
    31913245
    31923246\subsection{Discussion}
    31933247
    3194 Languages using 1:1 threading based on pthreads can at best meet or exceed (due to language overhead) the pthread results.
     3248Languages using 1:1 threading based on pthreads can at best meet or exceed, due to language overhead, the pthread results.
    31953249Note, pthreads has a fast zero-contention mutex lock checked in user space.
    31963250Languages with M:N threading have better performance than 1:1 because there is no operating-system interactions.
     
    32003254
    32013255
    3202 \section{Conclusion}
     3256\section{Conclusions and Future Work}
    32033257
    32043258Advanced control-flow will always be difficult, especially when there is temporal ordering and nondeterminism.
     
    32073261Combining these properties creates a number of high-level, efficient, and maintainable control-flow types: generator, coroutine, thread, each of which can be a monitor.
    32083262Eliminated from \CFA are barging and spurious wakeup, which are nonintuitive and lead to errors, and having to work with a bewildering set of low-level locks and acquisition techniques.
    3209 \CFA high-level race-free monitors and threads provide the core mechanisms for mutual exclusion and synchronization, without having to resort to magic qualifiers like @volatile@/@atomic@.
     3263\CFA high-level race-free monitors and threads, when used with mutex access function, provide the core mechanisms for mutual exclusion and synchronization, without having to resort to magic qualifiers like @volatile@ or @atomic@.
    32103264Extending these mechanisms to handle high-level deadlock-free bulk acquire across both mutual exclusion and synchronization is a unique contribution.
    32113265The \CFA runtime provides concurrency based on a preemptive M:N user-level threading-system, executing in clusters, which encapsulate scheduling of work on multiple kernel threads providing parallelism.
    32123266The M:N model is judged to be efficient and provide greater flexibility than a 1:1 threading model.
    32133267These concepts and the \CFA runtime-system are written in the \CFA language, extensively leveraging the \CFA type-system, which demonstrates the expressiveness of the \CFA language.
    3214 Performance comparisons with other concurrent systems/languages show the \CFA approach is competitive across all basic operations, which translates directly into good performance in well-written applications with advanced control-flow.
     3268Performance comparisons with other concurrent systems and languages show the \CFA approach is competitive across all basic operations, which translates directly into good performance in well-written applications with advanced control-flow.
    32153269C programmers should feel comfortable using these mechanisms for developing complex control-flow in applications, with the ability to obtain maximum available performance by selecting mechanisms at the appropriate level of need using only calling communication.
    32163270
    3217 
    3218 \section{Future Work}
    3219 
    32203271While control flow in \CFA has a strong start, development is still underway to complete a number of missing features.
    32213272
     3273\vspace{-5pt}
    32223274\paragraph{Flexible Scheduling}
    32233275\label{futur:sched}
    32243276
    32253277An important part of concurrency is scheduling.
    3226 Different scheduling algorithms can affect performance (both in terms of average and variation).
     3278Different scheduling algorithms can affect performance, both in terms of average and variation.
    32273279However, no single scheduler is optimal for all workloads and therefore there is value in being able to change the scheduler for given programs.
    32283280One solution is to offer various tuning options, allowing the scheduler to be adjusted to the requirements of the workload.
     
    32303282Currently, the \CFA pluggable scheduler is too simple to handle complex scheduling, \eg quality of service and real-time, where the scheduler must interact with mutex objects to deal with issues like priority inversion~\cite{Buhr00b}.
    32313283
     3284\vspace{-5pt}
    32323285\paragraph{Non-Blocking I/O}
    32333286\label{futur:nbio}
    3234 
    32353287Many modern workloads are not bound by computation but IO operations, common cases being web servers and XaaS~\cite{XaaS} (anything as a service).
    32363288These types of workloads require significant engineering to amortizing costs of blocking IO-operations.
     
    32413293A non-blocking I/O library is currently under development for \CFA.
    32423294
     3295\vspace{-5pt}
    32433296\paragraph{Other Concurrency Tools}
    32443297\label{futur:tools}
     
    32493302As well, new \CFA extensions should make it possible to create a uniform interface for virtually all mutual exclusion, including monitors and low-level locks.
    32503303
     3304\vspace{-5pt}
    32513305\paragraph{Implicit Threading}
    32523306\label{futur:implcit}
    32533307
    3254 Basic concurrent (embarrassingly parallel) applications can benefit greatly from implicit concurrency, where sequential programs are converted to concurrent, possibly with some help from pragmas to guide the conversion.
     3308Basic \emph{embarrassingly parallel} applications can benefit greatly from implicit concurrency, where sequential programs are converted to concurrent, possibly with some help from pragmas to guide the conversion.
    32553309This type of concurrency can be achieved both at the language level and at the library level.
    32563310The canonical example of implicit concurrency is concurrent nested @for@ loops, which are amenable to divide and conquer algorithms~\cite{uC++book}.
  • doc/papers/concurrency/mail2

    r4e7c0fc0 r04b4a71  
    512512Software: Practice and Experience Editorial Office
    513513
     514
     515
     516Date: Sat, 18 Apr 2020 10:42:13 +0000
     517From: Richard Jones <onbehalfof@manuscriptcentral.com>
     518Reply-To: R.E.Jones@kent.ac.uk
     519To: tdelisle@uwaterloo.ca, pabuhr@uwaterloo.ca
     520Subject: Software: Practice and Experience - Decision on Manuscript ID
     521 SPE-19-0219.R1
     522
     52318-Apr-2020
     524
     525Dear Dr Buhr,
     526
     527Many thanks for submitting SPE-19-0219.R1 entitled "Advanced Control-flow and Concurrency in Cforall" to Software: Practice and Experience. The paper has now been reviewed and the comments of the referees are included at the bottom of this letter.
     528
     529I believe that we are making progress here towards a paper that can be published in Software: Practice and Experience.  However the referees still have significant concerns about the paper. The journal's focus is on practice and experience, and one of the the reviewers' concerns remains that your submission should focus the narrative more on the perspective of the programmer than the language designer. I agree that this would strengthen your submission, and I ask you to address this as well as the referees' other comments.
     530
     531A revised version of your manuscript that takes into account the comments of the referee(s) will be reconsidered for publication.
     532
     533Please note that submitting a revision of your manuscript does not guarantee eventual acceptance, and that your revision may be subject to re-review by the referees before a decision is rendered.
     534
     535You have 90 days from the date of this email to submit your revision. If you are unable to complete the revision within this time, please contact me to request a short extension.
     536
     537You can upload your revised manuscript and submit it through your Author Center. Log into https://mc.manuscriptcentral.com/spe  and enter your Author Center, where you will find your manuscript title listed under "Manuscripts with Decisions".
     538
     539When submitting your revised manuscript, you will be able to respond to the comments made by the referee(s) in the space provided.  You can use this space to document any changes you make to the original manuscript.
     540
     541If you would like help with English language editing, or other article preparation support, Wiley Editing Services offers expert help with English Language Editing, as well as translation, manuscript formatting, and figure formatting at www.wileyauthors.com/eeo/preparation. You can also check out our resources for Preparing Your Article for general guidance about writing and preparing your manuscript at www.wileyauthors.com/eeo/prepresources.
     542
     543Once again, thank you for submitting your manuscript to Software: Practice and Experience and I look forward to receiving your revision.
     544
     545Sincerely,
     546Richard
     547
     548Prof. Richard Jones
     549Software: Practice and Experience
     550R.E.Jones@kent.ac.uk
     551
     552
     553Referee(s)' Comments to Author:
     554
     555Reviewing: 1
     556
     557Comments to the Author
     558(A relatively short second review)
     559
     560I thank the authors for their revisions and comprehensive response to
     561reviewers' comments --- many of my comments have been successfully
     562addressed by the revisions.  Here I'll structure my comments around
     563the main salient points in that response which I consider would
     564benefit from further explanation.
     565
     566>  Table 1 is moved to the start and explained in detail.
     567
     568I consider this change makes a significant improvement to the paper,
     569laying out the landscape of language features at the start, and thus
     570addresses my main concerns about the paper.
     571
     572I still have a couple of issues --- perhaps the largest is that it's
     573still not clear at this point in the paper what some of these options
     574are, or crucially how they would be used. I don't know if it's
     575possbile to give high-level examples or use cases to be clear about
     576these up front - or if that would duplicate too much information from
     577later in the paper - either way expanding out the discussion - even if
     578just two a couple of sentences for each row - would help me more.  The
     579point is not just to define these categories but to ensure the
     580readers' understanding of these definitons agrees with that used in
     581the paper.
     582
     583in a little more detail:
     584
     585 * 1st para section 2 begs the question: why not support each
     586   dimension independently, and let the programmer or library designer
     587   combiine features?
     588
     589 * "execution state" seems a relatively low-level description here.
     590  I don't think of e.g. the lambda calculus that way. Perhaps it's as
     591  good a term as any.
     592
     593 * Why must there "be language mechanisms to create, block/unblock,
     594   and join with a thread"?  There aren't in Smalltalk (although there
     595   are in the runtime).  Especially given in Cforall those mechanisms
     596   are *implicit* on thread creation and destruction?
     597
     598 * "Case 1 is a function that borrows storage for its state (stack
     599   frame/activation) and a thread from its invoker"
     600
     601   this much makes perfect sense to me, but I don't understand how a
     602   non-stateful, non-theaded function can then retain
     603
     604   "this state across callees, ie, function local-variables are
     605   retained on the stack across calls."
     606
     607   how can it retain function-local values *across calls* when it
     608   doesn't have any functional-local state?
     609
     610   I'm not sure if I see two separate cases here - rougly equivalent
     611   to C functions without static storage, and then C functions *with*
     612   static storage. I assumed that was the distinction between cases 1
     613   & 3; but perhpas the actual distinction is that 3 has a
     614   suspend/resume point, and so the "state" in figure 1 is this
     615   component of execution state (viz figs 1 & 2), not the state
     616   representing the cross-call variables?
     617
     618>    but such evaluation isn't appropriate for garbage-collected or JITTed
     619   languages like Java or Go.
     620
     621For JITTed languages in particular, reporting peak performance needs
     622to "warm up" the JIT with a number of iterators before beginning
     623measurement. Actually for JIT's its even worse: see Edd Barrett et al
     624OOPSLA 2017.
     625   
     626
     627
     628minor issues:
     629
     630 * footnote A - I've looked at various other papers & the website to
     631   try to understand how "object-oriented" Cforall is - I'm still not
     632   sure.  This footnote says Cforall has "virtuals" - presumably
     633   virtual functions, i.e. dynamic dispatch - and inheritance: that
     634   really is OO as far as I (and most OO people) are concerned.  For
     635   example Haskell doesn't have inheritance, so it's not OO; while
     636   CLOS (the Common Lisp *Object* System) or things like Cecil and
     637   Dylan are considered OO even though they have "multiple function
     638   parameters as receivers", lack "lexical binding between a structure
     639   and set of functions", and don't have explicit receiver invocation
     640   syntax.  Python has receiver syntax, but unlike Java or Smalltalk
     641   or C++, method declarations still need to have an explicit "self"
     642   receiver parameter.  Seems to me that Go, for example, is
     643   more-or-less OO with interfaces, methods, and dynamic dispatch (yes
     644   also and an explicit receiver syntax but that's not
     645   determiniative); while Rust lacks dynamic dispatch built-in.  C is
     646   not OO as a language, but as you say given it supports function
     647   pointers with structures, it does support an OO programm style.
     648 
     649   This is why I again recommend just not buying into this fight: not
     650   making any claims about whether Cforall is OO or is not - because
     651   as I see it, the rest of the paper doesn't depend on whether
     652   Cforall is OO or not.  That said: this is just a recommendation,
     653   and I won't quibble over this any further.
     654
     655 * is a "monitor function" the same as a "mutex function"?
     656   if so the paper should pick one term; if not, make the distinction clear.
     657
     658
     659 * "As stated on line 1 because state declarations from the generator
     660    type can be moved out of the coroutine type into the coroutine main"
     661
     662    OK sure, but again: *why* would a programmer want to do that?
     663    (Other than, I guess, to show the difference between coroutines &
     664    generators?)  Perhaps another way to put this is that the first
     665    para of 3.2 gives the disadvantages of coroutines vs-a-vs
     666    generators, briefly describes the extended semantics, but never
     667    actualy says why a programmer may want those extended semantics,
     668    or how they would benefit.  I don't mean to belabour the point,
     669    but (generalist?) readers like me would generally benefit from
     670    those kinds of discussions about each feature throughout the
     671    paper: why might a programmer want to use them?
     672   
     673
     674> p17 if the multiple-monitor entry procedure really is novel, write a paper
     675> about that, and only about that.
     676
     677> We do not believe this is a practical suggestion.
     678
     679 * I'm honestly not trying to be snide here: I'm not an expert on
     680   monitor or concurrent implementations. Brinch Hansen's original
     681   monitors were single acquire; this draft does not cite any other
     682   previous work that I could see. I'm not suggesting that the brief
     683   mention of this mechanism necessarily be removed from this paper,
     684   but if this is novel (and a clear advance over a classical OO
     685   monitor a-la Java which only acquires the distinguished reciever)
     686   then that would be worth another paper in itself.
     687 
     688> * conclusion should conclude the paper, not the related.
     689> We do not understand this comment.if ithis
     690
     691My typo: the paper's conclusion should come at the end, after the
     692future work section.
     693
     694
     695
     696
     697To encourage accountability, I'm signing my reviews in 2020.
     698For the record, I am James Noble, kjx@ecs.vuw.ac.nz.
     699
     700
     701Reviewing: 2
     702
     703Comments to the Author
     704I thank the authors for their detailed response. To respond to a couple of points raised  in response to my review (number 2):
     705
     706- on the Boehm paper and whether code is "all sequential to the compiler": I now understand the authors' position better and suspect we are in violent agreement, except for whether it's appropriate to use the rather breezy phrase "all sequential to the compiler". It would be straightforward to clarify that code not using the atomics features is optimized *as if* it were sequential, i.e. on the assumption of a lack of data races.
     707
     708- on the distinction between "mutual exclusion" and "synchronization": the added citation does help, in that it makes a coherent case for the definition the authors prefer. However, the text could usefully clarify that this is a matter of definition not of fact, given especially that in my assessment the authors' preferred definition is not the most common one. (Although the mention of Hoare's apparent use of this definition is one data point, countervailing ones are found in many contemporaneous or later papers, e.g. Habermann's 1972 "Synchronization of Communicating Processes" (CACM 15(3)), Reed & Kanodia's 1979 "Synchronization with eventcounts and sequencers" (CACM (22(2)) and so on.)
     709
     710I am glad to see that the authors have taken on board most of the straightforward improvements I suggested.
     711
     712However, a recurring problem of unclear writing still remains through many parts of the paper, including much of sections 2, 3 and 6. To highlight a couple of problem patches (by no means exhaustive):
     713
     714- section 2 (an expanded version of what was previously section 5.9) lacks examples and is generally obscure and allusory ("the most advanced feature" -- name it! "in triplets" -- there is only one triplet!; what are "execution locations"? "initialize" and "de-initialize" what? "borrowed from the invoker" is a concept in need of explaining or at least a fully explained example -- in what sense does a plain function borrow" its stack frame? "computation only" as opposed to what? in 2.2, in what way is a "request" fundamental to "synchronization"? and the "implicitly" versus "explicitly" point needs stating as elsewhere, with a concrete example e.g. Java built-in mutexes versus java.util.concurrent).
     715
     716- section 6: 6.2 omits the most important facts in preference for otherwise inscrutable detail: "identify the kind of parameter" (first say *that there are* kinds of parameter, and what "kinds" means!); "mutex parameters are documentation" is misleading (they are also semantically significant!) and fails to say *what* they mean; the most important thing is surely that 'mutex' is a language feature for performing lock/unlock operations at function entry/exit. So say it! The meanings of examples f3 and f4 remain unclear. Meanwhile in 6.3, "urgent" is not introduced (we are supposed to infer its meaning from Figure 12, but that Figure is incomprehensible to me), and we are told of "external scheduling"'s long history in Ada but not clearly what it actually means; 6.4's description of "waitfor" tells us it is different from an if-else chain but tries to use two *different* inputs to tell us that the behavior is different; tell us an instance where *the same* values of C1 and C2 give different behavior (I even wrote out a truth table and still don't see the semantic difference)
     717
     718The authors frequently use bracketed phrases, and sometimes slashes "/", in ways that are confusing and/or detrimental to readability. Page 13 line 2's "forward (backward)" is one particularly egregious example. In general I would recommend the the authors try to limit their use of parentheses and slashes as a means of forcing a clearer wording to emerge. Also, the use of "eg." is often cursory and does not explain the examples given, which are frequently a one- or two-word phrase of unclear referent.
     719
     720Considering the revision more broadly, none of the more extensive or creative rewrites I suggested in my previous review have been attempted, nor any equivalent efforts to improve its readability. The hoisting of the former section 5.9 is a good idea, but the newly added material accompanying it (around Table 1) suffers fresh deficiencies in clarity. Overall the paper is longer than before, even though (as my previous review stated), I believe a shorter paper is required in order to serve the likely purpose of publication. (Indeed, the authors' letter implies that a key goal of publication is to build community and gain external users.)
     721
     722Given this trajectory, I no longer see a path to an acceptable revision of the present submission. Instead I suggest the authors consider splitting the paper in two: one half about coroutines and stack management, the other about mutexes, monitors and the runtime. (A briefer presentation of the runtime may be helpful in the first paper also, and a brief recap of the generator and coroutine support is obviously needed in the second too.) Both of these new papers would need to be written with a strong emphasis on clarity, paying great care to issues of structure, wording, choices of example, and restraint (saying what's important, not everything that could be said). I am confident the authors could benefit from getting early feedback from others at their institution. For the performance experiments, of course these do not split evenly -- most (but not all) belong in the second of these two hypothetical papers. But the first of them would still have plenty of meat to it; for me, a clear and thorough study of the design space around coroutines is the most interesting and tantalizing prospect.
     723
     724I do not buy the authors' defense of the limited practical experience or "non-micro" benchmarking presented. Yes, gaining external users is hard and I am sympathetic on that point. But building something at least *somewhat* substantial with your own system should be within reach, and without it the "practice and experience" aspects of the work have not been explored. Clearly C\/ is the product of a lot of work over an extended period, so it is a surprise that no such experience is readily available for inclusion.
     725
     726Some smaller points:
     727
     728It does not seem right to state that a stack is essential to Von Neumann architectures -- since the earliest Von Neumann machines (and indeed early Fortran) did not use one.
     729
     730To elaborate on something another reviewer commented on: it is a surprise to find a "Future work" section *after* the "Conclusion" section. A "Conclusions and future work" section often works well.
     731
     732
     733Reviewing: 3
     734
     735Comments to the Author
     736This is the second round of reviewing.
     737
     738As in the first review, I found that the paper (and Cforall) contains
     739a lot of really interesting ideas, but it remains really difficult to
     740have a good sense of which idea I should use and when. This applies in
     741different ways to different features from the language:
     742
     743* coroutines/generators/threads: here there is
     744  some discussion, but it can be improved.
     745* interal/external scheduling: I didn't find any direct comparison
     746  between these features, except by way of example.
     747
     748I requested similar things in my previous review and I see that
     749content was added in response to those requests. Unfortunately, I'm
     750not sure that I can say it improved the paper's overall read. I think
     751in some sense the additions were "too much" -- I would have preferred
     752something more like a table or a few paragraphs highlighting the key
     753reasons one would pick one construct or the other.
     754
     755In general, I do wonder if the paper is just trying to do too much.
     756The discussion of clusters and pre-emption in particular feels quite
     757rushed.
     758
     759## Summary
     760
     761I make a number of suggestions below but the two most important
     762I think are:
     763
     764* Recommend to shorten the comparison on coroutine/generator/threads
     765  in Section 2 to a paragraph with a few examples, or possibly a table
     766  explaining the trade-offs between the constructs
     767* Recommend to clarify the relationship between internal/external
     768  scheduling -- is one more general but more error-prone or low-level?
     769
     770## Coroutines/generators/threads
     771
     772There is obviously a lot of overlap between these features, and in
     773particular between coroutines and generators. As noted in the previous
     774review, many languages have chosen to offer *only* generators, and to
     775build coroutines by stacks of generators invoking one another.
     776
     777I believe the newly introduced Section 2 of the paper is trying to
     778motivate why each of these constructs exist, but I did not find it
     779effective. It was dense and difficult to understand. I think the
     780problem is that Section 2 seems to be trying to derive "from first
     781principles" why each construct exists, but I think that a more "top
     782down" approach would be easier to understand.
     783
     784In fact, the end of Section 2.1 (on page 5) contains a particular
     785paragraph that embodies this "top down" approach. It starts,
     786"programmers can now answer three basic questions", and thus gives
     787some practical advice for which construct you should use and when. I
     788think giving some examples of specific applications that this
     789paragraph, combined with some examples of cases where each construct
     790was needed, would be a better approach.
     791
     792I don't think this compariosn needs to be very long. It seems clear
     793enough that one would
     794
     795* prefer generators for simple computations that yield up many values,
     796* prefer coroutines for more complex processes that have significant
     797  internal structure,
     798* prefer threads for cases where parallel execution is desired or
     799  needed.
     800
     801I did appreciate the comparison in Section 2.3 between async-await in
     802JS/Java and generators/coroutines. I agree with its premise that those
     803mechanisms are a poor replacement for generators (and, indeed, JS has
     804a distinct generator mechanism, for example, in part for this reason).
     805I believe I may have asked for this in a previous review, but having
     806read it, I wonder if it is really necessary, since those mechanisms
     807are so different in purpose.
     808
     809## Internal vs external scheduling
     810
     811I find the motivation for supporting both internal and external
     812scheduling to be fairly implicit. After several reads through the
     813section, I came to the conclusion that internal scheduling is more
     814expressive than external scheduling, but sometimes less convenient or
     815clear. Is this correct? If not, it'd be useful to clarify where
     816external scheduling is more expressive.
     817
     818The same is true, I think, of the `signal_block` function, which I
     819have not encountered before; it seems like its behavior can be modeled
     820with multiple condition variables, but that's clearly more complex.
     821
     822One question I had about `signal_block`: what happens if one signals
     823but no other thread is waiting? Does it block until some other thread
     824waits? Or is that user error?
     825
     826I would find it very interesting to try and capture some of the
     827properties that make internal vs external scheduling the better
     828choice.
     829
     830For example, it seems to me that external scheduling works well if
     831there are only a few "key" operations, but that internal scheduling
     832might be better otherwise, simply because it would be useful to have
     833the ability to name a signal that can be referenced by many
     834methods. Consider the bounded buffer from Figure 13: if it had
     835multiple methods for removing elements, and not just `remove`, then
     836the `waitfor(remove)` call in `insert` might not be sufficient.
     837
     838## Comparison of external scheduling to messaging
     839
     840I did enjoy the section comparing external scheduling to Go's
     841messaging mechanism, which I believe is a new addition.
     842
     843I believe that one difference between the Go program and the Cforall
     844equivalent is that the Goroutine has an associated queue, so that
     845multiple messages could be enqueued, whereas the Cforall equivalent is
     846effectively a "bounded buffer" of length 1. Is that correct? I think
     847this should be stated explicitly. (Presumably, one could modify the
     848Cforall program to include an explicit vector of queued messages if
     849desired, but you would also be reimplementing the channel
     850abstraction.)
     851
     852Also, in Figure 20, I believe that there is a missing `mutex` keyword.
     853The fiugre states:
     854
     855```
     856void main(GoRtn & gortn) with(gortn) {
     857```
     858
     859but I think it should probably be as follows:
     860
     861```
     862void main(GoRtn & mutex gortn) with(gortn) {
     863```
     864
     865Unless there is some implicit `mutex` associated with being a main
     866function for a `monitor thread`.
     867
     868## Atomic operations and race freedom
     869
     870I was glad to see that the paper acknowledged that Cforall still had
     871low-level atomic operations, even if their use is discouraged in favor
     872of higher-level alternatives.
     873
     874However, I still feel that the conclusion overstates the value of the
     875contribution here when it says that "Cforall high-level race-free
     876monitors and threads provide the core mechanisms for mutual exclusion
     877and synchronization, without the need for volatile and atomics". I
     878feel confident that Java programmers, for example, would be advised to
     879stick with synchronized methods whenever possible, and it seems to me
     880that they offer similar advantages -- but they sometimes wind up using
     881volatiles for performance reasons.
     882
     883I was also confused by the term "race-free" in that sentence. In
     884particular, I don't think that Cforall has any mechanisms for
     885preventing *data races*, and it clearly doesn't prevent "race
     886conditions" (which would bar all sorts of useful programs). I suppose
     887that "race free" here might be referring to the improvements such as
     888removing barging behavior.
     889
     890## Performance comparisons
     891
     892In my previous review, I requested comparisons against Rust and
     893node.js, and I see that the new version of the paper includes both,
     894which is a good addition.
     895
     896One note on the Rust results: I believe that the results are comparing
     897against the threads found in Rust's standard library, which are
     898essentially a shallow wrapper around pthreads, and hence the
     899performance is quite close to pthread performance (as one would
     900expect). It would perhaps be more interesting to see a comparison
     901built using [tokio] or [async-std], two of the more prominent
     902user-space threading libraries that build on Rust's async-await
     903feature (which operates quite differently than Javascript's
     904async-await, in that it doesn't cause every aync function call to
     905schedule a distinct task).
     906
     907[tokio]: https://tokio.rs/
     908[async-std]: https://async.rs/
     909
     910That said, I am satisfied with the performance results as they are in
     911the current revision.
     912
     913## Minor notes and typos
     914
     915Several figures used the `with` keyword. I deduced that `with(foo)`
     916permits one to write `bar` instead of `foo.bar`. It seems worth
     917introducing. Apologies if this is stated in the paper, if so I missed
     918it.
     919
     920On page 20, section 6.3, "external scheduling and vice versus" should be
     921"external scheduling and vice versa".
     922
     923On page 5, section 2.3, the paper states "we content" but it should be
     924"we contend".
     925
     926Reviewing: Editor
     927
     928A few small comments in addition to those of the referees.
     929
     930Page 1. I don't believe that it s fair to imply that Scala is  "research vehicle" as it is used by major players, Twitter being the most prominent example.
     931
     932Page 15. Must Cforall threads start after construction (e.g. see your example on page 15, line 21)? I can think of examples where it is not desirable that threads start immediately after construction, e.g. a game with N players, each of whom is expensive to create, but all of whom should be started at the same time.
     933
     934Page 18, line 17: is using
     935
Note: See TracChangeset for help on using the changeset viewer.