source: doc/proposals/concurrency/thesis.tex @ eb182b0

ADTaaron-thesisarm-ehast-experimentalcleanup-dtorsdeferred_resndemanglerenumforall-pointer-decayjacob/cs343-translationjenkins-sandboxnew-astnew-ast-unique-exprnew-envno_listpersistent-indexerpthread-emulationqualifiedEnumresolv-newwith_gc
Last change on this file since eb182b0 was dbfb35d, checked in by Thierry Delisle <tdelisle@…>, 7 years ago

Renamed concurrency to thesis before splitting into chapters

  • Property mode set to 100644
File size: 82.0 KB
Line 
1% requires tex packages: texlive-base texlive-latex-base tex-common texlive-humanities texlive-latex-extra texlive-fonts-recommended
2
3% inline code ©...© (copyright symbol) emacs: C-q M-)
4% red highlighting ®...® (registered trademark symbol) emacs: C-q M-.
5% blue highlighting ß...ß (sharp s symbol) emacs: C-q M-_
6% green highlighting ¢...¢ (cent symbol) emacs: C-q M-"
7% LaTex escape §...§ (section symbol) emacs: C-q M-'
8% keyword escape ¶...¶ (pilcrow symbol) emacs: C-q M-^
9% math escape $...$ (dollar symbol)
10
11\documentclass[letterpaper,12pt,titlepage,oneside,final]{book}
12
13%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
14
15% Latex packages used in the document.
16\usepackage[T1]{fontenc}                                        % allow Latin1 (extended ASCII) characters
17\usepackage{textcomp}
18\usepackage[latin1]{inputenc}
19\usepackage{fullpage,times,comment}
20\usepackage{epic,eepic}
21\usepackage{upquote}                                            % switch curled `'" to straight
22\usepackage{calc}
23\usepackage{xspace}
24\usepackage{graphicx}
25\usepackage{tabularx}
26\usepackage{multicol}
27\usepackage[acronym]{glossaries}
28\usepackage{varioref}   
29\usepackage{listings}                                           % format program code
30\usepackage[flushmargin]{footmisc}                              % support label/reference in footnote
31\usepackage{latexsym}                                           % \Box glyph
32\usepackage{mathptmx}                                           % better math font with "times"
33\usepackage[usenames]{color}
34\usepackage[pagewise]{lineno}
35\usepackage{fancyhdr}
36\renewcommand{\linenumberfont}{\scriptsize\sffamily}
37\input{style}                                                   % bespoke macros used in the document
38\usepackage[dvips,plainpages=false,pdfpagelabels,pdfpagemode=UseNone,colorlinks=true,pagebackref=true,linkcolor=blue,citecolor=blue,urlcolor=blue,pagebackref=true,breaklinks=true]{hyperref}
39\usepackage{breakurl}
40
41\usepackage{tikz}
42\def\checkmark{\tikz\fill[scale=0.4](0,.35) -- (.25,0) -- (1,.7) -- (.25,.15) -- cycle;}
43
44\renewcommand{\UrlFont}{\small\sf}
45
46\setlength{\topmargin}{-0.45in}                         % move running title into header
47\setlength{\headsep}{0.25in}
48
49%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
50
51% Names used in the document.
52
53\newcommand{\Version}{1.0.0}
54\newcommand{\CS}{C\raisebox{-0.9ex}{\large$^\sharp$}\xspace}
55
56\newcommand{\Textbf}[2][red]{{\color{#1}{\textbf{#2}}}}
57\newcommand{\Emph}[2][red]{{\color{#1}\textbf{\emph{#2}}}}
58\newcommand{\R}[1]{\Textbf{#1}}
59\newcommand{\B}[1]{{\Textbf[blue]{#1}}}
60\newcommand{\G}[1]{{\Textbf[OliveGreen]{#1}}}
61\newcommand{\uC}{$\mu$\CC}
62\newcommand{\cit}{\textsuperscript{[Citation Needed]}\xspace}
63\newcommand{\code}[1]{\lstinline[language=CFA]{#1}}
64\newcommand{\pscode}[1]{\lstinline[language=pseudo]{#1}}
65\newcommand{\TODO}{{\Textbf{TODO}}}
66
67\input{glossary}
68
69\newsavebox{\LstBox}
70
71%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
72
73\setcounter{secnumdepth}{3}                             % number subsubsections
74\setcounter{tocdepth}{3}                                % subsubsections in table of contents
75% \linenumbers                                            % comment out to turn off line numbering
76\makeindex
77\pagestyle{fancy}
78\fancyhf{}
79\cfoot{\thepage}
80\rfoot{v\input{version}}
81
82%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
83
84\begin{document}
85% \linenumbers
86
87\title{Concurrency in \CFA}
88\author{Thierry Delisle \\
89School of Computer Science, University of Waterloo, \\ Waterloo, Ontario, Canada
90}
91
92\maketitle
93
94% ### #     # ####### ######  #######
95%  #  ##    #    #    #     # #     #
96%  #  # #   #    #    #     # #     #
97%  #  #  #  #    #    ######  #     #
98%  #  #   # #    #    #   #   #     #
99%  #  #    ##    #    #    #  #     #
100% ### #     #    #    #     # #######
101
102\chapter{Introduction}
103This proposal provides a minimal core concurrency API that is both simple, efficient and can be reused to build higher-level features. The simplest possible concurrency core is a thread and a lock but this low-level approach is hard to master. An easier approach for users is to support higher-level constructs as the basis of the concurrency in \CFA. Indeed, for highly productive parallel programming, high-level approaches are much more popular~\cite{HPP:Study}. Examples are task based, message passing and implicit threading.
104
105There are actually two problems that need to be solved in the design of the concurrency for a programming language: which concurrency tools are available to the users and which parallelism tools are available. While these two concepts are often seen together, they are in fact distinct concepts that require different sorts of tools~\cite{Buhr05a}. Concurrency tools need to handle mutual exclusion and synchronization, while parallelism tools are more about performance, cost and resource utilization.
106
107%  #####  ####### #     #  #####  #     # ######  ######  ####### #     #  #####  #     #
108% #     # #     # ##    # #     # #     # #     # #     # #       ##    # #     #  #   #
109% #       #     # # #   # #       #     # #     # #     # #       # #   # #         # #
110% #       #     # #  #  # #       #     # ######  ######  #####   #  #  # #          #
111% #       #     # #   # # #       #     # #   #   #   #   #       #   # # #          #
112% #     # #     # #    ## #     # #     # #    #  #    #  #       #    ## #     #    #
113%  #####  ####### #     #  #####   #####  #     # #     # ####### #     #  #####     #
114
115\chapter{Concurrency}
116Several tool can be used to solve concurrency challenges. Since these challenges always appear with the use of mutable shared-state, some languages and libraries simply disallow mutable shared-state (Erlang~\cite{Erlang}, Haskell~\cite{Haskell}, Akka (Scala)~\cite{Akka}). In these paradigms, interaction among concurrent objects relies on message passing~\cite{Thoth,Harmony,V-Kernel} or other paradigms that closely relate to networking concepts (channels\cit for example). However, in languages that use routine calls as their core abstraction mechanism, these approaches force a clear distinction between concurrent and non-concurrent paradigms (i.e., message passing versus routine call). Which in turn means that, in order to be effective, programmers need to learn two sets of designs patterns. This distinction can be hidden away in library code, but effective use of the librairy still has to take both paradigms into account. Approaches based on shared memory are more closely related to non-concurrent paradigms since they often rely on basic constructs like routine calls and objects. At a lower level these can be implemented as locks and atomic operations. Many such mechanisms have been proposed, including semaphores~\cite{Dijkstra68b} and path expressions~\cite{Campbell74}. However, for productivity reasons it is desireable to have a higher-level construct be the core concurrency paradigm~\cite{HPP:Study}. An approach that is worth mentionning because it is gaining in popularity is transactionnal memory~\cite{Dice10}[Check citation]. While this approach is even pursued by system languages like \CC\cit, the performance and feature set is currently too restrictive to add such a paradigm to a language like C or \CC\cit, which is why it was rejected as the core paradigm for concurrency in \CFA. One of the most natural, elegant, and efficient mechanisms for synchronization and communication, especially for shared memory systems, is the \emph{monitor}. Monitors were first proposed by Brinch Hansen~\cite{Hansen73} and later described and extended by C.A.R.~Hoare~\cite{Hoare74}. Many programming languages---e.g., Concurrent Pascal~\cite{ConcurrentPascal}, Mesa~\cite{Mesa}, Modula~\cite{Modula-2}, Turing~\cite{Turing:old}, Modula-3~\cite{Modula-3}, NeWS~\cite{NeWS}, Emerald~\cite{Emerald}, \uC~\cite{Buhr92a} and Java~\cite{Java}---provide monitors as explicit language constructs. In addition, operating-system kernels and device drivers have a monitor-like structure, although they often use lower-level primitives such as semaphores or locks to simulate monitors. For these reasons, this project proposes monitors as the core concurrency construct.
117
118% #     # ####### #     # ### ####### ####### ######   #####
119% ##   ## #     # ##    #  #     #    #     # #     # #     #
120% # # # # #     # # #   #  #     #    #     # #     # #
121% #  #  # #     # #  #  #  #     #    #     # ######   #####
122% #     # #     # #   # #  #     #    #     # #   #         #
123% #     # #     # #    ##  #     #    #     # #    #  #     #
124% #     # ####### #     # ###    #    ####### #     #  #####
125
126\section{Monitors}
127A monitor is a set of routines that ensure mutual exclusion when accessing shared state. This concept is generally associated with Object-Oriented Languages like Java~\cite{Java} or \uC~\cite{uC++book} but does not strictly require OOP semantics. The only requirements is the ability to declare a handle to a shared object and a set of routines that act on it :
128\begin{cfacode}
129        typedef /*some monitor type*/ monitor;
130        int f(monitor & m);
131
132        int main() {
133                monitor m;
134                f(m);
135        }
136\end{cfacode}
137
138%  #####     #    #       #
139% #     #   # #   #       #
140% #        #   #  #       #
141% #       #     # #       #
142% #       ####### #       #
143% #     # #     # #       #
144%  #####  #     # ####### #######
145
146\subsection{Call semantics} \label{call}
147The above monitor example displays some of the intrinsic characteristics. Indeed, it is necessary to use pass-by-reference over pass-by-value for monitor routines. This semantics is important because at their core, monitors are implicit mutual-exclusion objects (locks), and these objects cannot be copied. Therefore, monitors are implicitly non-copyable.
148
149Another aspect to consider is when a monitor acquires its mutual exclusion. For example, a monitor may need to be passed through multiple helper routines that do not acquire the monitor mutual-exclusion on entry. Pass through can be both generic helper routines (\code{swap}, \code{sort}, etc.) or specific helper routines like the following to implement an atomic counter :
150
151\begin{cfacode}
152        monitor counter_t { /*...see section $\ref{data}$...*/ };
153
154        void ?{}(counter_t & nomutex this); //constructor
155        size_t ++?(counter_t & mutex this); //increment
156
157        //need for mutex is platform dependent here
158        void ?{}(size_t * this, counter_t & mutex cnt); //conversion
159\end{cfacode}
160
161Here, the constructor(\code{?\{\}}) uses the \code{nomutex} keyword to signify that it does not acquire the monitor mutual exclusion when constructing. This semantics is because an object not yet constructed should never be shared and therefore does not require mutual exclusion. The prefix increment operator uses \code{mutex} to protect the incrementing process from race conditions. Finally, there is a conversion operator from \code{counter_t} to \code{size_t}. This conversion may or may not require the \code{mutex} key word depending on whether or not reading an \code{size_t} is an atomic operation or not.
162
163Having both \code{mutex} and \code{nomutex} keywords could be argued to be redundant based on the meaning of a routine having neither of these keywords. For example, given a routine without quualifiers \code{void foo(counter_t & this)} then one could argue that it should default to the safest option \code{mutex}. On the other hand, the option of having routine \code{void foo(counter_t & this)} mean \code{nomutex} is unsafe by default and may easily cause subtle errors. It can be argued that \code{nomutex} is the more "normal" behaviour, the \code{nomutex} keyword effectively stating explicitly that "this routine has nothing special". Another alternative is to make having exactly one of these keywords mandatory, which would provide the same semantics but without the ambiguity of supporting routine \code{void foo(counter_t & this)}. Mandatory keywords would also have the added benefice of being self-documented but at the cost of extra typing. However, since \CFA relies heavily on traits as an abstraction mechanism, the distinction between a type that is a monitor and a type that looks like a monitor can become blurred. For this reason, \CFA only has the \code{mutex} keyword.
164
165
166The next semantic decision is to establish when \code{mutex} may be used as a type qualifier. Consider the following declarations:
167\begin{cfacode}
168        int f1(monitor & mutex m);
169        int f2(const monitor & mutex m);
170        int f3(monitor ** mutex m);
171        int f4(monitor *[] mutex m);
172        int f5(graph(monitor*) & mutex m);
173\end{cfacode}
174The problem is to indentify which object(s) should be acquired. Furthermore, each object needs to be acquired only once. In the case of simple routines like \code{f1} and \code{f2} it is easy to identify an exhaustive list of objects to acquire on entry. Adding indirections (\code{f3}) still allows the compiler and programmer to indentify which object is acquired. However, adding in arrays (\code{f4}) makes it much harder. Array lengths are not necessarily known in C and even then making sure we only acquire objects once becomes also none trivial. This can be extended to absurd limits like \code{f5}, which uses a graph of monitors. To keep everyone as sane as possible~\cite{Chicken}, this projects imposes the requirement that a routine may only acquire one monitor per parameter and it must be the type of the parameter with one level of indirection (ignoring potential qualifiers). Also note that while routine \code{f3} can be supported, meaning that monitor \code{**m} is be acquired, passing an array to this routine would be type safe and yet result in undefined behavior because only the first element of the array is acquired. However, this ambiguity is part of the C type system with respects to arrays. For this reason, \code{mutex} is disallowed in the context where arrays may be passed.
175
176Finally, for convenience, monitors support multiple acquireing, that is acquireing a monitor while already holding it does not cause a deadlock. It simply increments an internal counter which is then used to release the monitor after the number of acquires and releases match up.
177
178% ######     #    #######    #
179% #     #   # #      #      # #
180% #     #  #   #     #     #   #
181% #     # #     #    #    #     #
182% #     # #######    #    #######
183% #     # #     #    #    #     #
184% ######  #     #    #    #     #
185
186\subsection{Data semantics} \label{data}
187Once the call semantics are established, the next step is to establish data semantics. Indeed, until now a monitor is used simply as a generic handle but in most cases monitors contian shared data. This data should be intrinsic to the monitor declaration to prevent any accidental use of data without its appropriate protection. For example, here is a complete version of the counter showed in section \ref{call}:
188\begin{cfacode}
189        monitor counter_t {
190                int value;
191        };
192
193        void ?{}(counter_t & this) {
194                this.cnt = 0;
195        }
196
197        int ++?(counter_t & mutex this) {
198                return ++this.value;
199        }
200
201        //need for mutex is platform dependent here
202        void ?{}(int * this, counter_t & mutex cnt) {
203                *this = (int)cnt;
204        }
205\end{cfacode}
206
207This simple counter is used as follows:
208\begin{center}
209\begin{tabular}{c @{\hskip 0.35in} c @{\hskip 0.35in} c}
210\begin{cfacode}
211        //shared counter
212        counter_t cnt;
213
214        //multiple threads access counter
215        thread 1 : cnt++;
216        thread 2 : cnt++;
217        thread 3 : cnt++;
218          ...
219        thread N : cnt++;
220\end{cfacode}
221\end{tabular}
222\end{center}
223
224Notice how the counter is used without any explicit synchronisation and yet supports thread-safe semantics for both reading and writting. Unlike object-oriented monitors, where calling a mutex member \emph{implicitly} acquires mutual-exclusion, \CFA uses an explicit mechanism to acquire mutual-exclusion. A consequence of this approach is that it extends to multi-monitor calls.
225\begin{cfacode}
226        int f(MonitorA & mutex a, MonitorB & mutex b);
227
228        MonitorA a;
229        MonitorB b;
230        f(a,b);
231\end{cfacode}
232This code acquires both locks before entering the critical section, called \emph{\gls{group-acquire}}. In practice, writing multi-locking routines that do not lead to deadlocks is tricky. Having language support for such a feature is therefore a significant asset for \CFA. In the case presented above, \CFA guarantees that the order of aquisition is consistent across calls to routines using the same monitors as arguments. However, since \CFA monitors use multi-acquisition locks, users can effectively force the acquiring order. For example, notice which routines use \code{mutex}/\code{nomutex} and how this affects aquiring order :
233\begin{cfacode}
234        void foo(A & mutex a, B & mutex b) { //acquire a & b
235                //...
236        }
237
238        void bar(A & mutex a, B & /*nomutex*/ b) { //acquire a
239                //...
240                foo(a, b); //acquire b
241                //...
242        }
243
244        void baz(A & /*nomutex*/ a, B & mutex b) { //acquire b
245                //...
246                foo(a, b); //acquire a
247                //...
248        }
249\end{cfacode}
250
251The multi-acquisition monitor lock allows a monitor lock to be acquired by both \code{bar} or \code{baz} and acquired again in \code{foo}. In the calls to \code{bar} and \code{baz} the monitors are acquired in opposite order. such use leads to nested monitor call problems~\cite{Lister77}, which is a more specific variation of the lock acquiring order problem. In the example above, the user uses implicit ordering in the case of function \code{foo} but explicit ordering in the case of \code{bar} and \code{baz}. This subtle mistake means that calling these routines concurrently may lead to deadlock and is therefore undefined behavior. As shown on several occasion\cit, solving this problem requires :
252\begin{enumerate}
253        \item Dynamically tracking of the monitor-call order.
254        \item Implement rollback semantics.
255\end{enumerate}
256
257While the first requirement is already a significant constraint on the system, implementing a general rollback semantics in a C-like language is prohibitively complex \cit. In \CFA, users simply need to be carefull when acquiring multiple monitors at the same time.
258
259% ######  ####### #######    #    ### #        #####
260% #     # #          #      # #    #  #       #     #
261% #     # #          #     #   #   #  #       #
262% #     # #####      #    #     #  #  #        #####
263% #     # #          #    #######  #  #             #
264% #     # #          #    #     #  #  #       #     #
265% ######  #######    #    #     # ### #######  #####
266%
267%             ######  ####### #       #     # #     # ####### ######  #     #
268%             #     # #     # #        #   #  ##   ## #     # #     # #     #
269%             #     # #     # #         # #   # # # # #     # #     # #     #
270%  #####    ######  #     # #          #    #  #  # #     # ######  #######
271%             #       #     # #          #    #     # #     # #   #   #     #
272%             #       #     # #          #    #     # #     # #    #  #     #
273%             #       ####### #######    #    #     # ####### #     # #     #
274
275\subsection{Implementation Details: Interaction with polymorphism}
276At first glance, interaction between monitors and \CFA's concept of polymorphism seems complex to support. However, it is shown that entry-point locking can solve most of the issues.
277
278Before looking into complex control flow, it is important to present the difference between the two acquiring options : \gls{callsite-locking} and \gls{entry-point-locking}, i.e. acquiring the monitors before making a mutex call or as the first instruction of the mutex call. For example:
279
280\begin{center}
281\begin{tabular}{|c|c|c|}
282Code & \gls{callsite-locking} & \gls{entry-point-locking} \\
283\CFA & pseudo-code & pseudo-code \\
284\hline
285\begin{cfacode}[tabsize=3]
286void foo(monitor& mutex a){
287
288
289
290        //Do Work
291        //...
292
293}
294
295void main() {
296        monitor a;
297
298
299
300        foo(a);
301
302}
303\end{cfacode} & \begin{pseudo}[tabsize=3]
304foo(& a) {
305
306
307
308        //Do Work
309        //...
310
311}
312
313main() {
314        monitor a;
315        //calling routine
316        //handles concurrency
317        acquire(a);
318        foo(a);
319        release(a);
320}
321\end{pseudo} & \begin{pseudo}[tabsize=3]
322foo(& a) {
323        //called routine
324        //handles concurrency
325        acquire(a);
326        //Do Work
327        //...
328        release(a);
329}
330
331main() {
332        monitor a;
333
334
335
336        foo(a);
337
338}
339\end{pseudo}
340\end{tabular}
341\end{center}
342
343First of all, interaction between \code{otype} polymorphism and monitors is impossible since monitors do not support copying. Therefore, the main question is how to support \code{dtype} polymorphism. Since a monitor's main purpose is to ensure mutual exclusion when accessing shared data, this implies that mutual exclusion is only required for routines that do in fact access shared data. However, since \code{dtype} polymorphism always handles incomplete types (by definition), no \code{dtype} polymorphic routine can access shared data since the data requires knowledge about the type. Therefore, the only concern when combining \code{dtype} polymorphism and monitors is to protect access to routines. \Gls{callsite-locking} would require a significant amount of work, since any \code{dtype} routine may have to obtain some lock before calling a routine, depending on whether or not the type passed is a monitor. However, with \gls{entry-point-locking} calling a monitor routine becomes exactly the same as calling it from anywhere else. Note that the \code{mutex} keyword relies on the resolver, which mean that in cases where generic monitor routines is actually desired, writing mutex routine is possible with the proper trait.
344
345
346% ### #     # #######         #####   #####  #     # ####### ######
347%  #  ##    #    #           #     # #     # #     # #       #     #
348%  #  # #   #    #           #       #       #     # #       #     #
349%  #  #  #  #    #            #####  #       ####### #####   #     #
350%  #  #   # #    #    ###          # #       #     # #       #     #
351%  #  #    ##    #    ###    #     # #     # #     # #       #     #
352% ### #     #    #    ###     #####   #####  #     # ####### ######
353
354\section{Internal scheduling} \label{insched}
355In addition to mutual exclusion, the monitors at the core of \CFA's concurrency can also be used to achieve synchronisation. With monitors, this is generally achieved with internal or external scheduling as in\cit. Since internal scheduling of single monitors is mostly a solved problem, this proposal concentraits on extending internal scheduling to multiple monitors at once. Indeed, like the \gls{group-acquire} semantics, internal scheduling extends to multiple monitors at once in a way that is natural to the user but requires additional complexity on the implementation side.
356
357First, Here is a simple example of such a technique :
358
359\begin{cfacode}
360        monitor A {
361                condition e;
362        }
363
364        void foo(A & mutex a) {
365                // ...
366                // We need someone else to do something now
367                wait(a.e);
368                // ...
369        }
370
371        void bar(A & mutex a) {
372                // Do the thing foo is waiting on
373                // ...
374                // Signal foo it's done
375                signal(a.e);
376        }
377\end{cfacode}
378
379Note that in \CFA, \code{condition} have no particular need to be stored inside a monitor, beyond any software engineering reasons. Here routine \code{foo} waits for the \code{signal} from \code{bar} before making further progress, effectively ensuring a basic ordering. An important aspect to take into account here is that \CFA does not allow barging, which means that once function \code{bar} releases the monitor, foo is guaranteed to resume immediately after (unless some other function waited on the same condition). This guarantees offers the benefit of not having to loop arount waits in order to guarantee that a condition is still met. The main reason \CFA offers this guarantee is that users can easily introduce barging if it becomes a necessity but adding a barging prevention or barging avoidance is more involved without language support.
380
381Supporting barging prevention as well as extending internal scheduling to multiple monitors is the main source of complexity in the design of \CFA concurrency.
382
383\subsection{Internal Scheduling - multi monitor}
384It easier to understand the problem of multi monitor scheduling using a series of pseudo code though experiment. Note that in the following snippets of pseudo-code waiting and signalling is done without the use of a condition variable. While \CFA requires condition variables to use signalling, the variable itself only really holds the data needed for the implementation of internal schedulling. Some languages like JAVA\cit simply define an implicit condition variable for every monitor while other languages like \uC use explicit condition variables. Since the following pseudo-codes are simple and focused experiments, all condition variables are implicit.
385
386\begin{multicols}{2}
387\begin{pseudo}
388acquire A
389        wait A
390release A
391\end{pseudo}
392
393\columnbreak
394
395\begin{pseudo}
396acquire A
397        signal A
398release A
399\end{pseudo}
400\end{multicols}
401
402The previous example shows the simple case of having two threads (one for each column) and a single monitor A. One thread acquires before waiting and the other acquires before signalling. There are a few important things to note here. First, both \code{wait} and \code{signal} must be called with the proper monitor(s) already acquired. This can be hidden on the user side but is a logical requirement for barging prevention. Secondly, as stated above, while it is argued that not all problems regarding single monitors are solved, this paper only regards challenges of \gls{group-acquire} and considers other problems related to monitors as solved.
403
404An important note about this example is that signalling a monitor is a delayed operation. The ownership of the monitor is transferred only when the monitor would have otherwise been released, not at the point of the \code{signal} statement.
405
406A direct extension of the previous example is the \gls{group-acquire} version :
407
408\begin{multicols}{2}
409\begin{pseudo}
410acquire A & B
411        wait A & B
412release A & B
413\end{pseudo}
414
415\columnbreak
416
417\begin{pseudo}
418acquire A & B
419        signal A & B
420release A & B
421\end{pseudo}
422\end{multicols}
423
424This version uses \gls{group-acquire} (denoted using the \& symbol), but the presence of multiple monitors does not add a particularly new meaning. Synchronization will happen between the two threads in exactly the same way and order. The only difference is that mutual exclusion will cover more monitors. On the implementation side, handling multiple monitors at once does add a degree of complexity but it is not significant compared to the next few examples.
425
426For the sake of completeness, here is another example of the single-monitor case, this time with nesting.
427
428\begin{multicols}{2}
429\begin{pseudo}
430acquire A
431        acquire B
432                wait B
433        release B
434release A
435\end{pseudo}
436
437\columnbreak
438
439\begin{pseudo}
440
441acquire B
442        signal B
443release B
444
445\end{pseudo}
446\end{multicols}
447
448While these cases can cause some deadlock issues, we consider that these issues are only a symptom of the fact that locks, and by extension monitors, are not perfectly composable. However, for monitors as for locks, it is possible to write program that using nesting without encountering any problems if they are nested carefully.
449
450The next example is where \gls{group-acquire} adds a significant layer of complexity to the internal signalling semantics.
451
452\begin{multicols}{2}
453\begin{pseudo}
454acquire A
455        // Code Section 1
456        acquire A & B
457                // Code Section 2
458                wait A & B
459                // Code Section 3
460        release A & B
461        // Code Section 4
462release A
463\end{pseudo}
464
465\columnbreak
466
467\begin{pseudo}
468acquire A
469        // Code Section 5
470        acquire A & B
471                // Code Section 6
472                signal A & B
473                // Code Section 7
474        release A & B
475        // Code Section 8
476release A
477\end{pseudo}
478\end{multicols}
479
480It is particularly important to pay attention to code sections 8 and 3 which are where the existing semantics of internal scheduling are undefined. The root of the problem is that \gls{group-acquire} is used in a context where one of the monitors is already acquired. As mentionned in previous sections, monitors support multiple acquiring which means the that nesting \gls{group-acquire} can be done safely. However, in the context of internal scheduling it is important to define the behaviour of the previous pseudo-code. When the signaller thread reaches the location where it should "release A \& B", it actually only needs to release the monitor B. Since the other thread is waiting on monitor B, the signaller thread cannot simply release the monitor into the wild. This would mean that the waiting thread would have to reacquire the monitor and would therefore open the door to barging threads. Since the signalling thread still needs the monitor A, simply transferring ownership to the waiting thread is not an option because it would pottentially violate mutual exclusion. We are therefore left with three options :
481
482\subsubsection{Delaying signals}
483The first more obvious solution to solve the problem of multi-monitor scheduling is to keep ownership of all locks until the last lock is ready to be transferred. It can be argued that that moment is the correct time to transfer ownership when the last lock is no longer needed is what fits most closely to the behaviour of single monitor scheduling. However, this solution can become much more complicated depending on the content of the code section 8. Indeed, nothing prevents a user from signalling monitor A on a different condition variable. In that case, if monitor B is transferred with monitor A, then it means the system needs to handle threads having ownership on more monitors than expected and how to tie monitors together. On the other hand if the signalling thread only transfers monitor A then somehow both monitors A and B have to be transferred to the waiting thread from two different threads. While this solution may work, it was not fully explored because there is no apparent upper bound on the complexity of ownership transfer.
484
485\subsubsection{Dependency graphs}
486In the previous pseudo-code, there is a solution which would statisfy both barging prevention and mutual exclusion. If ownership of both monitors is transferred to the waiter when the signaller releases A and then the waiter transfers back ownership of A when it releases it then the problem is solved. This is the second solution. The problem it encounters is that it effectively boils down to resolving a dependency graph of ownership requirements. Here even the simplest of code snippets requires two transfers and it seems to increase in a manner closer to polynomial. For example the following code which is just a direct extension to three monitors requires at least three ownership transfer and has multiple solutions.
487
488\begin{multicols}{2}
489\begin{pseudo}
490acquire A
491        acquire B
492                acquire C
493                        wait A & B & C
494                release C
495        release B
496release A
497\end{pseudo}
498
499\columnbreak
500
501\begin{pseudo}
502acquire A
503        acquire B
504                acquire C
505                        signal A & B & C
506                release C
507        release B
508release A
509\end{pseudo}
510\end{multicols}
511
512\subsubsection{Partial signalling}
513Finally, the solution that was chosen for \CFA is to use partial signalling. Consider the following case :
514
515\begin{multicols}{2}
516\begin{pseudo}[numbers=left]
517acquire A
518        acquire A & B
519                wait A & B
520        release A & B
521release A
522\end{pseudo}
523
524\columnbreak
525
526\begin{pseudo}[numbers=left, firstnumber=6]
527acquire A
528        acquire A & B
529                signal A & B
530        release A & B
531        // ... More code
532release A
533\end{pseudo}
534\end{multicols}
535
536The partial signalling solution transfers ownership of monitor B at lines 10 but does not wake the waiting thread since it is still using monitor A. Only when it reaches line 11 does it actually wakeup the waiting thread. This solution has the benefit that complexity is encapsulated in to only two actions, passing monitors to the next owner when they should be release and conditionnaly waking threads if all conditions are met. Contrary to the other solutions, this solution quickly hits an upper bound on complexity of implementation.
537
538% Hard extension :
539
540% Incorrect options for the signal :
541
542% \begin{description}
543%  \item[-] Release B and baton pass after Code Section 8 : Passing b without having it
544%  \item[-] Keep B during Code Section 8 : Can lead to deadlocks since we secretly keep a lock longer than specified by the user
545%  \item[-] Instead of release B transfer A and B to waiter then try to reacquire A before running Code Section 8 : This allows barging
546% \end{description}
547
548% Since we don't want barging we need to pass A \& B and somehow block and get A back.
549
550% \begin{center}
551% \begin{tabular}{ c @{\hskip 0.65in} c }
552% \begin{lstlisting}[language=Pseudo]
553% acquire A
554%       acquire B
555%               acquire C
556%                       wait A & B & C
557%               1: release C
558%       2: release B
559% 3: release A
560% \end{lstlisting}&\begin{lstlisting}[language=Pseudo]
561% acquire A
562%       acquire B
563%               acquire C
564%                       signal A & B & C
565%               4: release C
566%       5: release B
567% 6: release A
568% \end{lstlisting}
569% \end{tabular}
570% \end{center}
571
572% To prevent barging :
573
574% \begin{description}
575%  \item[-] When the signaller hits 4 : pass A, B, C to waiter
576%  \item[-] When the waiter hits 2 : pass A, B to signaller
577%  \item[-] When the signaller hits 5 : pass A to waiter
578% \end{description}
579
580
581% \begin{center}
582% \begin{tabular}{ c @{\hskip 0.65in} c }
583% \begin{lstlisting}[language=Pseudo]
584% acquire A
585%       acquire C
586%               acquire B
587%                       wait A & B & C
588%               1: release B
589%       2: release C
590% 3: release A
591% \end{lstlisting}&\begin{lstlisting}[language=Pseudo]
592% acquire B
593%       acquire A
594%               acquire C
595%                       signal A & B & C
596%               4: release C
597%       5: release A
598% 6: release B
599% \end{lstlisting}
600% \end{tabular}
601% \end{center}
602
603% To prevent barging : When the signaller hits 4 : pass A, B, C to waiter. When the waiter hits 1 it must release B,
604
605% \begin{description}
606%  \item[-]
607%  \item[-] When the waiter hits 1 : pass A, B to signaller
608%  \item[-] When the signaller hits 5 : pass A, B to waiter
609%  \item[-] When the waiter hits 2 : pass A to signaller
610% \end{description}
611
612% Monitors also need to schedule waiting threads internally as a mean of synchronization. Internal scheduling is one of the simple examples of such a feature. It allows users to declare condition variables and have threads wait and signaled from them. Here is a simple example of such a technique :
613
614% \begin{lstlisting}
615%       mutex struct A {
616%               condition e;
617%       }
618
619%       void foo(A & mutex a) {
620%               //...
621%               wait(a.e);
622%               //...
623%       }
624
625%       void bar(A & mutex a) {
626%               signal(a.e);
627%       }
628% \end{lstlisting}
629
630% Note that in \CFA, \code{condition} have no particular need to be stored inside a monitor, beyond any software engineering reasons. Here routine \code{foo} waits for the \code{signal} from \code{bar} before making further progress, effectively ensuring a basic ordering.
631
632% As for simple mutual exclusion, these semantics must also be extended to include \gls{group-acquire} :
633% \begin{center}
634% \begin{tabular}{ c @{\hskip 0.65in} c }
635% Thread 1 & Thread 2 \\
636% \begin{lstlisting}
637% void foo(A & mutex a,
638%            A & mutex b) {
639%       //...
640%       wait(a.e);
641%       //...
642% }
643
644% foo(a, b);
645% \end{lstlisting} &\begin{lstlisting}
646% void bar(A & mutex a,
647%            A & mutex b) {
648%       signal(a.e);
649% }
650
651
652
653% bar(a, b);
654% \end{lstlisting}
655% \end{tabular}
656% \end{center}
657
658% To define the semantics of internal scheduling, it is important to look at nesting and \gls{group-acquire}. Indeed, beyond concerns about lock ordering, without scheduling the two following pseudo codes are mostly equivalent. In fact, if we assume monitors are ordered alphabetically, these two pseudo codes would probably lead to exactly the same implementation :
659
660% \begin{table}[h!]
661% \centering
662% \begin{tabular}{c c}
663% \begin{lstlisting}[language=pseudo]
664% monitor A, B, C
665
666% acquire A
667%       acquire B & C
668
669%                       //Do stuff
670
671%       release B & C
672% release A
673% \end{lstlisting} &\begin{lstlisting}[language=pseudo]
674% monitor A, B, C
675
676% acquire A
677%       acquire B
678%               acquire C
679%                       //Do stuff
680%               release C
681%       release B
682% release A
683% \end{lstlisting}
684% \end{tabular}
685% \end{table}
686
687% Once internal scheduling is introduce however, semantics of \gls{group-acquire} become relevant. For example, let us look into the semantics of the following pseudo-code :
688
689% \begin{lstlisting}[language=Pseudo]
690% 1: monitor A, B, C
691% 2: condition c1
692% 3:
693% 4: acquire A
694% 5:            acquire A & B & C
695% 6:                            signal c1
696% 7:            release A & B & C
697% 8: release A
698% \end{lstlisting}
699
700% Without \gls{group-acquire} signal simply baton passes the monitor lock on the next release. In the case above, we therefore need to indentify the next release. If line 8 is picked at the release point, then the signal will attempt to pass A \& B \& C, without having ownership of B \& C. Since this violates mutual exclusion, we conclude that line 7 is the only valid location where signalling can occur. The traditionnal meaning of signalling is to transfer ownership of the monitor(s) and immediately schedule the longest waiting task. However, in the discussed case, the signalling thread expects to maintain ownership of monitor A. This can be expressed in two differents ways : 1) the thread transfers ownership of all locks and reacquires A when it gets schedulled again or 2) it transfers ownership of all three monitors and then expects the ownership of A to be transferred back.
701
702% However, the question is does these behavior motivate supporting acquireing non-disjoint set of monitors. Indeed, if the previous example was modified to only acquire B \& C at line 5 (an release the accordingly) then in respects to scheduling, we could add the simplifying constraint that all monitors in a bulk will behave the same way, simplifying the problem back to a single monitor problem which has already been solved. For this constraint to be acceptble however, we need to demonstrate that in does not prevent any meaningful possibilities. And, indeed, we can look at the two previous interpretation of the above pseudo-code and conclude that supporting the acquiring of non-disjoint set of monitors does not add any expressiveness to the language.
703
704% Option 1 reacquires the lock after the signal statement, this can be rewritten as follows without the need for non-disjoint sets :
705% \begin{lstlisting}[language=Pseudo]
706% monitor A, B, C
707% condition c1
708
709% acquire A & B & C
710%       signal c1
711% release A & B & C
712% acquire A
713
714% release A
715% \end{lstlisting}
716
717% This pseudo code has almost exaclty the same semantics as the code acquiring intersecting sets of monitors.
718
719% Option 2 uses two-way lock ownership transferring instead of reacquiring monitor A. Two-way monitor ownership transfer is normally done using signalBlock semantics, which immedietely transfers ownership of a monitor before getting the ownership back when the other thread no longer needs the monitor. While the example pseudo-code for Option 2 seems toe transfer ownership of A, B and C and only getting A back, this is not a requirement. Getting back all 3 monitors and releasing B and C differs only in performance. For this reason, the second option could arguably be rewritten as :
720
721% \begin{lstlisting}[language=Pseudo]
722% monitor A, B, C
723% condition c1
724
725% acquire A
726%       acquire B & C
727%               signalBlock c1
728%       release B & C
729% release A
730% \end{lstlisting}
731
732% Obviously, the difference between these two snippets of pseudo code is that the first one transfers ownership of A, B and C while the second one only transfers ownership of B and C. However, this limitation can be removed by allowing user to release extra monitors when using internal scheduling, referred to as extended internal scheduling (pattent pending) from this point on. Extended internal scheduling means the two following pseudo-codes are functionnaly equivalent :
733% \begin{table}[h!]
734% \centering
735% \begin{tabular}{c @{\hskip 0.65in} c}
736% \begin{lstlisting}[language=pseudo]
737% monitor A, B, C
738% condition c1
739
740% acquire A
741%       acquire B & C
742%               signalBlock c1 with A
743%       release B & C
744% release A
745% \end{lstlisting} &\begin{lstlisting}[language=pseudo]
746% monitor A, B, C
747% condition c1
748
749% acquire A
750%       acquire A & B & C
751%               signal c1
752%       release A & B & C
753% release A
754% \end{lstlisting}
755% \end{tabular}
756% \end{table}
757
758% It must be stated that the extended internal scheduling only makes sense when using wait and signalBlock, since they need to prevent barging, which cannot be done in the context of signal since the ownership transfer is strictly one-directionnal.
759
760% One critic that could arise is that extended internal schedulling is not composable since signalBlock must be explicitly aware of which context it is in. However, this argument is not relevant since acquire A, B and C in a context where a subset of them is already acquired cannot be achieved without spurriously releasing some locks or having an oracle aware of all monitors. Therefore, composability of internal scheduling is no more an issue than composability of monitors in general.
761
762% The main benefit of using extended internal scheduling is that it offers the same expressiveness as intersecting monitor set acquiring but greatly simplifies the selection of a leader (or representative) for a group of monitor. Indeed, when using intersecting sets, it is not obvious which set intersects with other sets which means finding a leader representing only the smallest scope is a hard problem. Where as when using disjoint sets, any monitor that would be intersecting must be specified in the extended set, the leader can be chosen as any monitor in the primary set.
763
764% We need to make sure the semantics for internally scheduling N monitors are a natural extension of the single monitor semantics. For this reason, we introduce the concept of \gls{mon-ctx}. In terms of context internal scheduling means "releasing a \gls{mon-ctx} and waiting for an other thread to acquire the same \gls{mon-ctx} and baton-pass it back to the initial thread". This definitions requires looking into what a \gls{mon-ctx} is and what the semantics of waiting and baton-passing are.
765
766% \subsubsection{Internal scheduling: Context} \label{insched-context}
767% Monitor scheduling operations are defined in terms of the context they are in. In languages that only supports operations on a single monitor at once, the context is completly defined by which most recently acquired monitors. Indeed, acquiring several monitors will form a stack of monitors which will be released in FILO order. In \CFA, a \gls{mon-ctx} cannot be simply defined by the last monitor that was acquired since \gls{group-acquire} means multiple monitors can be "the last monitor acquired". The \gls{mon-ctx} is therefore defined as the last set of monitors to have been acquired. This means taht when any new monitor is acquired, the group it belongs to is the new \gls{mon-ctx}. Correspondingly, if any monitor is released, the \gls{mon-ctx} reverts back to the context that was used prior to the monitor being acquired. In the most common case, \gls{group-acquire} means every monitor of a group will be acquired in released at the same time. However, since every monitor has its own recursion level, \gls{group-acquire} does not prevent users from reacquiring certain monitors while acquireing new monitors in the same operation. For example :
768
769% \begin{lstlisting}
770% //Forward declarations
771% monitor a, b, c
772% void foo( monitor & mutex a,
773%             monitor & mutex b);
774% void bar( monitor & mutex a,
775%             monitor & mutex b);
776% void baz( monitor & mutex a,
777%             monitor & mutex b,
778%             monitor & mutex c);
779
780% //Routines defined inline to illustrate context changed compared to the stack
781
782% //main thread
783% foo(a, b) {
784%       //thread calls foo
785%       //acquiring context a & b
786
787%       baz(a, b) {
788%               //thread calls baz
789%               //no context change
790
791%               bar(a, b, c) {
792%                       //thread calls bar
793%                       //acquiring context a & b & c
794
795%                       //Do stuff
796
797%                       return;             
798%                       //call to bar returns
799%               }
800%               //context back to a & b
801
802%               return;
803%               //call to baz returns
804%       }
805%       //no context change
806
807%       return;
808%       //call to foo returns
809% }
810% //context back to initial state
811
812% \end{lstlisting}
813
814% As illustrated by the previous example, context changes can be caused by only one of the monitors comming into context or going out of context.
815
816% \subsubsection{Internal scheduling: Waiting} \label{insched-wait}
817
818% \subsubsection{Internal scheduling: Baton Passing} \label{insched-signal}
819% Baton passing in internal scheduling is done in terms of \code{signal} and \code{signalBlock}\footnote{Arguably, \code{signal_now} is a more evocative name and \code{signal} could be changed appropriately. }. While \code{signalBlock} is the more straight forward way of baton passing, transferring ownership immediately, it must rely on \code{signal} which is why t is discussed first.
820% \code{signal} has for effect to transfer the current context to another thread when the context would otherwise be released. This means that instead of releasing the concerned monitors, the first thread on the condition ready-queue is scheduled to run. The monitors are not released and when the signalled thread runs, it assumes it regained ownership of all the monitors it had in its context.
821
822% \subsubsection{Internal scheduling: Implementation} \label{insched-impl}
823% Too implement internal scheduling, three things are need : a data structure for waiting tasks, a data structure for signalled task and a leaving procedure to run the signalled task. In the case of both data structures, it is desireable to have to use intrusive data structures in order to prevent the need for any dynamic allocation. However, in both cases being able to queue several items in the same position in a queue is non trivial, even more so in the presence of concurrency. However, within a given \gls{mon-ctx}, all monitors have exactly the same behavior in regards to scheduling. Therefore, the problem of queuing multiple monitors at once can be ignored by choosing one monitor to represent every monitor in a context. While this could prove difficult in other situations, \gls{group-acquire} requires that the monitors be sorted according to some stable predicate. Since monitors are sorted in all contexts, the representative can simply be the first in the list. Choosing a representative means a simple intrusive queue inside the condition is sufficient to implement the data structure for both waiting and signalled monitors.
824
825% Since \CFA monitors don't have a complete image of the \gls{mon-ctx}, choosing the representative and maintaning the current context information cannot easily be done by any single monitors. However, as discussed in section [Missing section here], monitor mutual exclusion is implemented using an raii object which is already in charge of sorting monitors. This object has a complete picture of the \gls{mon-ctx} which means it is well suited to choose the reprensentative and detect context changes.
826
827% \newpage
828% \begin{lstlisting}
829% void ctor( monitor ** _monitors, int _count ) {
830%       bool ctx_changed = false;
831%       for( mon in _monitors ) {
832%               ctx_changed = acquire( mon ) || ctx_changed;
833%       }
834
835%       if( ctx_changed ) {
836%               set_representative();
837%               set_context();
838%       }
839% }
840
841% void dtor( monitor ** _monitors, int _count ) {
842%       if( context_will_exit( _monitors, count ) ) {
843%               baton_pass();
844%               return;
845%       }
846
847%       for( mon in _monitors ) {
848%               release( mon );
849%       }
850% }
851
852% \end{lstlisting}
853
854
855
856% A direct extension of the single monitor semantics is to release all locks when waiting and transferring ownership of all locks when signalling. However, for the purpose of synchronization it may be usefull to only release some of the locks but keep others. It is possible to support internal scheduling and \gls{group-acquire} without any extra syntax by relying on order of acquisition. Here is an example of the different contexts in which internal scheduling can be used. (Note that here the use of helper routines is irrelevant, only routines acquire mutual exclusion have an impact on internal scheduling):
857
858% \begin{table}[h!]
859% \centering
860% \begin{tabular}{|c|c|c|}
861% Context 1 & Context 2 & Context 3 \\
862% \hline
863% \begin{lstlisting}
864% condition e;
865
866% //acquire a & b
867% void foo(monitor & mutex a,
868%            monitor & mutex b) {
869
870%       wait(e); //release a & b
871% }
872
873
874
875
876
877
878% foo(a,b);
879% \end{lstlisting} &\begin{lstlisting}
880% condition e;
881
882% //acquire a
883% void bar(monitor & mutex a,
884%            monitor & nomutex b) {
885%       foo(a,b);
886% }
887
888% //acquire a & b
889% void foo(monitor & mutex a,
890%            monitor & mutex b) {
891%       wait(e);  //release a & b
892% }
893
894% bar(a, b);
895% \end{lstlisting} &\begin{lstlisting}
896% condition e;
897
898% //acquire a
899% void bar(monitor & mutex a,
900%            monitor & nomutex b) {
901%       baz(a,b);
902% }
903
904% //acquire b
905% void baz(monitor & nomutex a,
906%            monitor & mutex b) {
907%       wait(e);  //release b
908% }
909
910% bar(a, b);
911% \end{lstlisting}
912% \end{tabular}
913% \end{table}
914
915% Context 1 is the simplest way of acquiring more than one monitor (\gls{group-acquire}), using a routine with multiple parameters having the \code{mutex} keyword. Context 2 also uses \gls{group-acquire} as well in routine \code{foo}. However, the routine is called by routine \code{bar}, which only acquires monitor \code{a}. Since monitors can be acquired multiple times this does not cause a deadlock by itself but it does force the acquiring order to \code{a} then \code{b}. Context 3 also forces the acquiring order to be \code{a} then \code{b} but does not use \gls{group-acquire}. The previous example tries to illustrate the semantics that must be established to support releasing monitors in a \code{wait} statement. In all cases, the behavior of the wait statment is to release all the locks that were acquired my the inner-most monitor call. That is \code{a & b} in context 1 and 2 and \code{b} only in context 3. Here are a few other examples of this behavior.
916
917
918% \begin{center}
919% \begin{tabular}{|c|c|c|}
920% \begin{lstlisting}
921% condition e;
922
923% //acquire b
924% void foo(monitor & nomutex a,
925%            monitor & mutex b) {
926%       bar(a,b);
927% }
928
929% //acquire a
930% void bar(monitor & mutex a,
931%            monitor & nomutex b) {
932
933%       wait(e); //release a
934%                 //keep b
935% }
936
937% foo(a, b);
938% \end{lstlisting} &\begin{lstlisting}
939% condition e;
940
941% //acquire a & b
942% void foo(monitor & mutex a,
943%            monitor & mutex b) {
944%       bar(a,b);
945% }
946
947% //acquire b
948% void bar(monitor & mutex a,
949%            monitor & nomutex b) {
950
951%       wait(e); //release b
952%                 //keep a
953% }
954
955% foo(a, b);
956% \end{lstlisting} &\begin{lstlisting}
957% condition e;
958
959% //acquire a & b
960% void foo(monitor & mutex a,
961%            monitor & mutex b) {
962%       bar(a,b);
963% }
964
965% //acquire none
966% void bar(monitor & nomutex a,
967%            monitor & nomutex b) {
968
969%       wait(e); //release a & b
970%                 //keep none
971% }
972
973% foo(a, b);
974% \end{lstlisting}
975% \end{tabular}
976% \end{center}
977% Note the right-most example is actually a trick pulled on the reader. Monitor state information is stored in thread local storage rather then in the routine context, which means that helper routines and other \code{nomutex} routines are invisible to the runtime system in regards to concurrency. This means that in the right-most example, the routine parameters are completly unnecessary. However, calling this routine from outside a valid monitor context is undefined.
978
979% These semantics imply that in order to release of subset of the monitors currently held, users must write (and name) a routine that only acquires the desired subset and simply calls wait. While users can use this method, \CFA offers the \code{wait_release}\footnote{Not sure if an overload of \code{wait} would work...} which will release only the specified monitors. In the center previous examples, the code in the center uses the \code{bar} routine to only release monitor \code{b}. Using the \code{wait_release} helper, this can be rewritten without having the name two routines :
980% \begin{center}
981% \begin{tabular}{ c c c }
982% \begin{lstlisting}
983%       condition e;
984
985%       //acquire a & b
986%       void foo(monitor & mutex a,
987%                  monitor & mutex b) {
988%               bar(a,b);
989%       }
990
991%       //acquire b
992%       void bar(monitor & mutex a,
993%                  monitor & nomutex b) {
994
995%               wait(e); //release b
996%                         //keep a
997%       }
998
999%       foo(a, b);
1000% \end{lstlisting} &\begin{lstlisting}
1001%       =>
1002% \end{lstlisting} &\begin{lstlisting}
1003%       condition e;
1004
1005%       //acquire a & b
1006%       void foo(monitor & mutex a,
1007%                  monitor & mutex b) {
1008%               wait_release(e,b); //release b
1009%                                        //keep a
1010%       }
1011
1012%       foo(a, b);
1013% \end{lstlisting}
1014% \end{tabular}
1015% \end{center}
1016
1017% Regardless of the context in which the \code{wait} statement is used, \code{signal} must be called holding the same set of monitors. In all cases, signal only needs a single parameter, the condition variable that needs to be signalled. But \code{signal} needs to be called from the same monitor(s) that call to \code{wait}. Otherwise, mutual exclusion cannot be properly transferred back to the waiting monitor.
1018
1019% Finally, an additional semantic which can be very usefull is the \code{signal_block} routine. This routine behaves like signal for all of the semantics discussed above, but with the subtelty that mutual exclusion is transferred to the waiting task immediately rather than wating for the end of the critical section.
1020% \\
1021
1022% ####### #     # #######         #####   #####  #     # ####### ######
1023% #        #   #     #           #     # #     # #     # #       #     #
1024% #         # #      #           #       #       #     # #       #     #
1025% #####      #       #            #####  #       ####### #####   #     #
1026% #         # #      #    ###          # #       #     # #       #     #
1027% #        #   #     #    ###    #     # #     # #     # #       #     #
1028% ####### #     #    #    ###     #####   #####  #     # ####### ######
1029\section{External scheduling} \label{extsched}
1030An alternative to internal scheduling is to use external scheduling instead. This method is more constrained and explicit which may help users tone down the undeterministic nature of concurrency. Indeed, as the following examples demonstrates, external scheduling allows users to wait for events from other threads without the concern of unrelated events occuring. External scheduling can generally be done either in terms of control flow (ex: \uC) or in terms of data (ex: Go). Of course, both of these paradigms have their own strenghts and weaknesses but for this project control flow semantics where chosen to stay consistent with the rest of the languages semantics. Two challenges specific to \CFA arise when trying to add external scheduling with loose object definitions and multi-monitor routines. The following example shows a simple use \code{accept} versus \code{wait}/\code{signal} and its advantages.
1031
1032\begin{center}
1033\begin{tabular}{|c|c|}
1034Internal Scheduling & External Scheduling \\
1035\hline
1036\begin{lstlisting}
1037        _Monitor blarg {
1038                condition c;
1039        public:
1040                void f() { signal(c)}
1041                void g() { wait(c); }
1042        private:
1043        }
1044\end{lstlisting}&\begin{lstlisting}
1045        _Monitor blarg {
1046
1047        public:
1048                void f() { /*...*/ }
1049                void g() { _Accept(f); }
1050        private:
1051        }
1052\end{lstlisting}
1053\end{tabular}
1054\end{center}
1055
1056In the case of internal scheduling, the call to \code{wait} only guarantees that \code{g} is the last routine to access the monitor. This intails that the routine \code{f} may have acquired mutual exclusion several times while routine \code{h} was waiting. On the other hand, external scheduling guarantees that while routine \code{h} was waiting, no routine other than \code{g} could acquire the monitor.
1057\\
1058
1059% #       ####### #######  #####  #######    ####### ######        #  #####
1060% #       #     # #     # #     # #          #     # #     #       # #     #
1061% #       #     # #     # #       #          #     # #     #       # #
1062% #       #     # #     #  #####  #####      #     # ######        #  #####
1063% #       #     # #     #       # #          #     # #     # #     #       #
1064% #       #     # #     # #     # #          #     # #     # #     # #     #
1065% ####### ####### #######  #####  #######    ####### ######   #####   #####
1066
1067\subsection{Loose object definitions}
1068In \uC, monitor declarations include an exhaustive list of monitor operations. Since \CFA is not object oriented it becomes both more difficult to implement but also less clear for the user :
1069
1070\begin{lstlisting}
1071        mutex struct A {};
1072
1073        void f(A & mutex a);
1074        void g(A & mutex a) { accept(f); }
1075\end{lstlisting}
1076
1077However, external scheduling is an example where implementation constraints become visible from the interface. Indeed, ince there is no hard limit to the number of threads trying to acquire a monitor concurrently, performance is a significant concern. Here is the pseudo code for the entering phase of a monitor :
1078
1079\begin{center}
1080\begin{tabular}{l}
1081\begin{lstlisting}[language=Pseudo]
1082        if monitor is free :
1083                enter
1084        elif monitor accepts me :
1085                enter
1086        else :
1087                block
1088\end{lstlisting}
1089\end{tabular}
1090\end{center}
1091
1092For the \pscode{monitor is free} condition it is easy to implement a check that can evaluate the condition in a few instruction. However, a fast check for \pscode{monitor accepts me} is much harder to implement depending on the constraints put on the monitors. Indeed, monitors are often expressed as an entry queue and some acceptor queue as in the following figure :
1093
1094\begin{center}
1095{\resizebox{0.4\textwidth}{!}{\input{monitor}}}
1096\end{center}
1097
1098There are other alternatives to these pictures but in the case of this picture implementing a fast accept check is relatively easy. Indeed simply updating a bitmask when the acceptor queue changes is enough to have a check that executes in a single instruction, even with a fairly large number of acceptor. However, this relies on the fact that all the acceptable routines are declared with the monitor type. For OO languages this doesn't compromise much since monitors already have an exhaustive list of member routines. However, for \CFA this isn't the case, routines can be added to a type anywhere after its declaration. Its important to note that the bitmask approach does not actually require an exhaustive list of routines, but it requires a dense unique ordering of routines with an upper-bound and that ordering must be consistent across translation units.
1099The alternative would be to have a picture more like this one:
1100
1101\begin{center}
1102{\resizebox{0.4\textwidth}{!}{\input{ext_monitor}}}
1103\end{center}
1104
1105Not storing the queues inside the monitor means that the storage can vary between routines, allowing for more flexibility and extensions. Storing an array of function-pointers would solve the issue of uniquely identifying acceptable routines. However, the single instruction bitmask compare has been replaced by dereferencing a pointer followed by a linear search. Furthermore, supporting nested external scheduling may now require additionnal searches on calls to accept to check if a routine is already queued in.
1106
1107At this point we must make a decision between flexibility and performance. Many design decisions in \CFA achieve both flexibility and performance, for example polymorphic routines add significant flexibility but inlining them means the optimizer can easily remove any runtime cost. Here however, the cost of flexibility cannot be trivially removed.
1108
1109In either cases here are a few alternatives for the different syntaxes this syntax : \\
1110\begin{center}
1111{\renewcommand{\arraystretch}{1.5}
1112\begin{tabular}[t]{l @{\hskip 0.35in} l}
1113\hline
1114\multicolumn{2}{ c }{\code{accept} on type}\\
1115\hline
1116Alternative 1 & Alternative 2 \\
1117\begin{lstlisting}
1118mutex struct A
1119accept( void f(A & mutex a) )
1120{};
1121\end{lstlisting} &\begin{lstlisting}
1122mutex struct A {}
1123accept( void f(A & mutex a) );
1124
1125\end{lstlisting} \\
1126Alternative 3 & Alternative 4 \\
1127\begin{lstlisting}
1128mutex struct A {
1129        accept( void f(A & mutex a) )
1130};
1131
1132\end{lstlisting} &\begin{lstlisting}
1133mutex struct A {
1134        accept :
1135                void f(A & mutex a) );
1136};
1137\end{lstlisting}\\
1138\hline
1139\multicolumn{2}{ c }{\code{accept} on routine}\\
1140\hline
1141\begin{lstlisting}
1142mutex struct A {};
1143
1144void f(A & mutex a)
1145
1146accept( void f(A & mutex a) )
1147void g(A & mutex a) {
1148        /*...*/
1149}
1150\end{lstlisting}&\\
1151\end{tabular}
1152}
1153\end{center}
1154
1155An other aspect to consider is what happens if multiple overloads of the same routine are used. For the time being it is assumed that multiple overloads of the same routine should be scheduled regardless of the overload used. However, this could easily be extended in the future.
1156
1157% #     # #     # #       ####### ###    #     # ####### #     #
1158% ##   ## #     # #          #     #     ##   ## #     # ##    #
1159% # # # # #     # #          #     #     # # # # #     # # #   #
1160% #  #  # #     # #          #     #     #  #  # #     # #  #  #
1161% #     # #     # #          #     #     #     # #     # #   # #
1162% #     # #     # #          #     #     #     # #     # #    ##
1163% #     #  #####  #######    #    ###    #     # ####### #     #
1164
1165\subsection{Multi-monitor scheduling}
1166
1167External scheduling, like internal scheduling, becomes orders of magnitude more complex when we start introducing multi-monitor syntax. Even in the simplest possible case some new semantics need to be established :
1168\begin{lstlisting}
1169        accept( void f(mutex struct A & mutex this))
1170        mutex struct A {};
1171
1172        mutex struct B {};
1173
1174        void g(A & mutex a, B & mutex b) {
1175                accept(f); //ambiguous, which monitor
1176        }
1177\end{lstlisting}
1178
1179The obvious solution is to specify the correct monitor as follows :
1180
1181\begin{lstlisting}
1182        accept( void f(mutex struct A & mutex this))
1183        mutex struct A {};
1184
1185        mutex struct B {};
1186
1187        void g(A & mutex a, B & mutex b) {
1188                accept( f, b );
1189        }
1190\end{lstlisting}
1191
1192This is unambiguous. Both locks will be acquired and kept, when routine \code{f} is called the lock for monitor \code{a} will be temporarily transferred from \code{g} to \code{f} (while \code{g} still holds lock \code{b}). This behavior can be extended to multi-monitor accept statment as follows.
1193
1194\begin{lstlisting}
1195        accept( void f(mutex struct A & mutex, mutex struct A & mutex))
1196        mutex struct A {};
1197
1198        mutex struct B {};
1199
1200        void g(A & mutex a, B & mutex b) {
1201                accept( f, b, a );
1202        }
1203\end{lstlisting}
1204
1205Note that the set of monitors passed to the \code{accept} statement must be entirely contained in the set of monitor already acquired in the routine. \code{accept} used in any other context is Undefined Behaviour.
1206
1207% ######  ####### #######    #    ### #        #####
1208% #     # #          #      # #    #  #       #     #
1209% #     # #          #     #   #   #  #       #
1210% #     # #####      #    #     #  #  #        #####
1211% #     # #          #    #######  #  #             #
1212% #     # #          #    #     #  #  #       #     #
1213% ######  #######    #    #     # ### #######  #####
1214%
1215%                #####  #     # ####### #     # #######  #####
1216%             #     # #     # #       #     # #       #     #
1217%             #     # #     # #       #     # #       #
1218%    #####    #     # #     # #####   #     # #####    #####
1219%             #   # # #     # #       #     # #             #
1220%             #    #  #     # #       #     # #       #     #
1221%                #### #  #####  #######  #####  #######  #####
1222
1223
1224\subsection{Implementation Details: External scheduling queues}
1225To support multi-monitor external scheduling means that some kind of entry-queues must be used that is aware of both monitors. However, acceptable routines must be aware of the entry queues which means they must be stored inside at least one of the monitors that will be acquired. This in turn adds the requirement a systematic algorithm of disambiguating which queue is relavant regardless of user ordering. The proposed algorithm is to fall back on monitors lock ordering and specify that the monitor that is acquired first is the lock with the relevant entry queue. This assumes that the lock acquiring order is static for the lifetime of all concerned objects but that is a reasonnable constraint. This algorithm choice has two consequences, the entry queue of the highest priority monitor is no longer a true FIFO queue and the queue of the lowest priority monitor is both required and probably unused. The queue can no longer be a FIFO queue because instead of simply containing the waiting threads in order arrival, they also contain the second mutex. Therefore, another thread with the same highest priority monitor but a different lowest priority monitor may arrive first but enter the critical section after a thread with the correct pairing. Secondly, since it may not be known at compile time which monitor will be the lowest priority monitor, every monitor needs to have the correct queues even though it is probable that half the multi-monitor queues will go unused for the entire duration of the program.
1226
1227\section{Other concurrency tools}
1228TO BE CONTINUED...
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239% ######     #    ######     #    #       #       ####### #       ###  #####  #     #
1240% #     #   # #   #     #   # #   #       #       #       #        #  #     # ##   ##
1241% #     #  #   #  #     #  #   #  #       #       #       #        #  #       # # # #
1242% ######  #     # ######  #     # #       #       #####   #        #   #####  #  #  #
1243% #       ####### #   #   ####### #       #       #       #        #        # #     #
1244% #       #     # #    #  #     # #       #       #       #        #  #     # #     #
1245% #       #     # #     # #     # ####### ####### ####### ####### ###  #####  #     #
1246\chapter{Parallelism}
1247Historically, computer performance was about processor speeds and instructions count. However, with heat dissipation being a direct consequence of speed increase, parallelism has become the new source for increased performance~\cite{Sutter05, Sutter05b}. In this decade, it is not longer reasonnable to create a high-performance application without caring about parallelism. Indeed, parallelism is an important aspect of performance and more specifically throughput and hardware utilization. The lowest-level approach of parallelism is to use \glspl{kthread} in combination with semantics like \code{fork}, \code{join}, etc. However, since these have significant costs and limitations, \glspl{kthread} are now mostly used as an implementation tool rather than a user oriented one. There are several alternatives to solve these issues that all have strengths and weaknesses. While there are many variations of the presented paradigms, most of these variations do not actually change the guarantees or the semantics, they simply move costs in order to achieve better performance for certain workloads.
1248
1249\section{Paradigm}
1250\subsection{User-level threads}
1251A direct improvement on the \gls{kthread} approach is to use \glspl{uthread}. These threads offer most of the same features that the operating system already provide but can be used on a much larger scale. This approach is the most powerfull solution as it allows all the features of multi-threading, while removing several of the more expensives costs of using kernel threads. The down side is that almost none of the low-level threading problems are hidden, users still have to think about data races, deadlocks and synchronization issues. These issues can be somewhat alleviated by a concurrency toolkit with strong garantees but the parallelism toolkit offers very little to reduce complexity in itself.
1252
1253Examples of languages that support \glspl{uthread} are Erlang~\cite{Erlang} and \uC~\cite{uC++book}.
1254
1255\subsection{Fibers : user-level threads without preemption}
1256A popular varient of \glspl{uthread} is what is often reffered to as \glspl{fiber}. However, \glspl{fiber} do not present meaningful semantical differences with \glspl{uthread}. Advocates of \glspl{fiber} list their high performance and ease of implementation as majors strenghts of \glspl{fiber} but the performance difference between \glspl{uthread} and \glspl{fiber} is controversial and the ease of implementation, while true, is a weak argument in the context of language design. Therefore this proposal largely ignore fibers.
1257
1258An example of a language that uses fibers is Go~\cite{Go}
1259
1260\subsection{Jobs and thread pools}
1261The approach on the opposite end of the spectrum is to base parallelism on \glspl{pool}. Indeed, \glspl{pool} offer limited flexibility but at the benefit of a simpler user interface. In \gls{pool} based systems, users express parallelism as units of work and a dependency graph (either explicit or implicit) that tie them together. This approach means users need not worry about concurrency but significantly limits the interaction that can occur among jobs. Indeed, any \gls{job} that blocks also blocks the underlying worker, which effectively means the CPU utilization, and therefore throughput, suffers noticeably. It can be argued that a solution to this problem is to use more workers than available cores. However, unless the number of jobs and the number of workers are comparable, having a significant amount of blocked jobs always results in idles cores.
1262
1263The gold standard of this implementation is Intel's TBB library~\cite{TBB}.
1264
1265\subsection{Paradigm performance}
1266While the choice between the three paradigms listed above may have significant performance implication, it is difficult to pindown the performance implications of chosing a model at the language level. Indeed, in many situations one of these paradigms may show better performance but it all strongly depends on the workload. Having a large amount of mostly independent units of work to execute almost guarantess that the \gls{pool} based system has the best performance thanks to the lower memory overhead. However, interactions between jobs can easily exacerbate contention. User-level threads allow fine-grain context switching, which results in better resource utilisation, but context switches will be more expansive and the extra control means users need to tweak more variables to get the desired performance. Furthermore, if the units of uninterrupted work are large enough the paradigm choice is largely amorticised by the actual work done.
1267
1268%  #####  #######    #          ####### ######  ######
1269% #     # #         # #            #    #     # #     #
1270% #       #        #   #           #    #     # #     #
1271% #       #####   #     # #####    #    ######  ######
1272% #       #       #######          #    #     # #     #
1273% #     # #       #     #          #    #     # #     #
1274%  #####  #       #     #          #    ######  ######
1275
1276\section{\CFA 's Thread Building Blocks}
1277As a system-level language, \CFA should offer both performance and flexibilty as its primary goals, simplicity and user-friendliness being a secondary concern. Therefore, the core of parallelism in \CFA should prioritize power and efficiency. With this said, deconstructing popular paradigms in order to get simple building blocks yields \glspl{uthread} as the core parallelism block. \Glspl{pool} and other parallelism paradigms can then be built on top of the underlying threading model.
1278
1279\subsection{Coroutines : A stepping stone}\label{coroutine}
1280While the main focus of this proposal is concurrency and paralellism, it is important to adress coroutines which are actually a significant underlying aspect of the concurrency system. Indeed, while having nothing todo with parallelism and arguably very little to do with concurrency, coroutines need to deal with context-switchs and and other context management operations. Therefore, this proposal includes coroutines both as an intermediate step for the implementation of threads and a first class feature of \CFA.
1281
1282The core API of coroutines revolve around two features : independent stacks and \code{suspend}/\code{resume}.
1283Here is an example of a solution to the fibonnaci problem using \CFA coroutines :
1284\begin{lstlisting}
1285        struct Fibonacci {
1286              int fn; // used for communication
1287              coroutine_descriptor c;
1288        };
1289
1290        void ?{}(Fibonacci* this) {
1291              this->fn = 0;
1292        }
1293
1294        coroutine_descriptor* get_¶coroutine¶(Fibonacci* this) {
1295              return &this->c;
1296        }
1297
1298        void co_main(Fibonacci* this) {
1299                int fn1, fn2;           // retained between resumes
1300                this->fn = 0;
1301                fn1 = this->fn;
1302                suspend(this);          // return to last resume
1303
1304                this->fn = 1;
1305                fn2 = fn1;
1306                fn1 = this->fn;
1307                suspend(this);          // return to last resume
1308
1309                for ( ;; ) {
1310                        this->fn = fn1 + fn2;
1311                        fn2 = fn1;
1312                        fn1 = this->fn;
1313                        suspend(this);  // return to last resume
1314                }
1315        }
1316
1317        int next(Fibonacci* this) {
1318                resume(this); // transfer to last suspend
1319                return this.fn;
1320        }
1321
1322        void main() {
1323                Fibonacci f1, f2;
1324                for ( int i = 1; i <= 10; i += 1 ) {
1325                        sout | next(&f1) | '§\verb+ +§' | next(&f2) | endl;
1326                }
1327        }
1328\end{lstlisting}
1329
1330\subsubsection{Construction}
1331One important design challenge for coroutines and threads (shown in section \ref{threads}) is that the runtime system needs to run some code after the user-constructor runs. In the case of the coroutines this challenge is simpler since there is no loss of determinism brough by preemption or scheduling, however, the underlying challenge remains the same for coroutines and threads.
1332
1333The runtime system needs to create the coroutine's stack and more importantly prepare it for the first resumption. The timing of the creation is non trivial since users both expect to have fully constructed objects once the main is called and to be able to resume the coroutine from the main (Obviously we only solve cases where these two statements don't conflict). There are several solutions to this problem but the chosen options effectively forces the design of the coroutine.
1334
1335Furthermore, \CFA faces an extra challenge which is that polymorphique routines rely on invisible thunks when casted to non-polymorphic routines and these thunks have function scope, for example :
1336
1337TODO : Simple case where a thunk would be created.
1338
1339
1340
1341\subsubsection{Alternative: Inheritance}
1342One solution to this challenge would be to use inheritence,
1343
1344\begin{lstlisting}
1345        struct Fibonacci {
1346              int fn; // used for communication
1347              coroutine c;
1348        };
1349
1350        void ?{}(Fibonacci* this) {
1351              this->fn = 0;
1352                (&this->c){};
1353        }
1354\end{lstlisting}
1355
1356There are two downsides to the approach. The first, which is relatively minor, is that the base class needs to be made aware of the main routine pointer, regardless of whether we use a parameter or a virtual pointer, this means the coroutine data must be made larger to store a value that is actually a compile time constant (The address of the main routine). The second problem which is both subtle but significant, is that now can get the initialisation order of there coroutines wrong. Indeed, every field of a \CFA struct will be constructed but in the order of declaration, unless users explicitly write otherwise. This means that users who forget to initialize a the coroutine at the right time may resume the coroutine at with an uninitilized object. For coroutines, this is unlikely to be a problem, for threads however, this is a significant problem.
1357
1358\subsubsection{Alternative: Reserved keyword}
1359The next alternative is to use language support to annotate coroutines as follows :
1360
1361\begin{lstlisting}
1362        coroutine struct Fibonacci {
1363              int fn; // used for communication
1364        };
1365\end{lstlisting}
1366
1367This mean the compiler can solve problems by injecting code where needed. The downside of this approach is that it makes coroutine a special case in the language. Users who would want to extend coroutines or build their own for various reasons can only do so in ways offered by the language. Furthermore, implementing coroutines without language supports also displays the power of \CFA.
1368
1369\subsubsection{Alternative: Lamda Objects}
1370
1371Boost does not use objects...
1372TO BE CONTINUED...
1373
1374\subsubsection{Trait based coroutines}
1375
1376Finally the approach chosen, which is the one closest to \CFA idioms, is to use trait-based lazy coroutines, the approach shown in section \ref{coroutine}. This approach defines a coroutine as anything that satisfies the \code{is_coroutine} and is used as a coroutine is a coroutine. This entails the an object is not a coroutine until \code{resume} (and \code{prime}) is called on the object. Correspondingly, any object that is passed to \code{resume} is a coroutine since it must satisfy the \code{is_coroutine} trait to compile.
1377
1378% ####### #     # ######  #######    #    ######   #####
1379%    #    #     # #     # #         # #   #     # #     #
1380%    #    #     # #     # #        #   #  #     # #
1381%    #    ####### ######  #####   #     # #     #  #####
1382%    #    #     # #   #   #       ####### #     #       #
1383%    #    #     # #    #  #       #     # #     # #     #
1384%    #    #     # #     # ####### #     # ######   #####
1385
1386\subsection{Thread Interface}\label{threads}
1387The basic building blocks of \CFA are \glspl{cfathread}. By default these are implemented as \glspl{uthread}, and as such, offer a flexible and lightweight threading interface (lightweight compared to \glspl{kthread}). A thread can be declared using a struct declaration with prefix \code{thread} as follows :
1388
1389\begin{lstlisting}
1390        trait is_¶thread¶(dtype T) {
1391                void co_main(T* this);
1392                coroutine* get_coroutine(T* this);
1393        };
1394
1395        thread struct foo {};
1396\end{lstlisting}
1397
1398Obviously, for this thread implementation to be usefull it must run some user code. Several other threading interfaces use a function-pointer representation as the interface of threads (for example : \Csharp~\cite{Csharp} and Scala~\cite{Scala}). However, this proposal considers that statically tying a \code{main} routine to a thread superseeds this approach. Since the \code{main} routine is already a special routine in \CFA (where the program begins), the existing syntax for declaring routines names with special semantics can be extended, i.e. operator overloading. As such the \code{main} routine of a thread can be defined as :
1399\begin{lstlisting}
1400        thread struct foo {};
1401
1402        void ?main(foo* this) {
1403                sout | "Hello World!" | endl;
1404        }
1405\end{lstlisting}
1406
1407In this example, threads of type \code{foo} will start there execution in the \code{void ?main(foo*)} routine which in this case prints \code{"Hello World!"}. While this proposoal encourages this approach which is enforces strongly type programming. Users may prefer to use the routine based thread semantics for the sake of simplicity. With these semantics it is trivial to write a thread type that takes a function pointer as parameter and executes it on its stack asynchronously :
1408\begin{lstlisting}
1409        typedef void (*voidFunc)(void);
1410
1411        thread struct FuncRunner {
1412                voidFunc func;
1413        };
1414
1415        //ctor
1416        void ?{}(FuncRunner* this, voidFunc inFunc) {
1417                func = inFunc;
1418        }
1419
1420        //main
1421        void t_main(FuncRunner* this) {
1422                this->func();
1423        }
1424\end{lstlisting}
1425
1426Of course for threads to be useful, it must be possible to start and stop threads and wait for them to complete execution. While using an \acrshort{api} such as \code{fork} and \code{join} is relatively common in the literature, such an interface is unnecessary. Indeed, the simplest approach is to use \acrshort{raii} principles and have threads \code{fork} once the constructor has completed and \code{join} before the destructor runs.
1427\begin{lstlisting}
1428thread struct World; //FuncRunner declared above
1429
1430void ?main(thread World* this) {
1431        sout | "World!" | endl;
1432}
1433
1434void main() {
1435        World w;
1436        //Thread run forks here
1437
1438        //Print to "Hello " and "World!" will be run concurrently
1439        sout | "Hello " | endl;
1440
1441        //Implicit join at end of scope
1442}
1443\end{lstlisting}
1444This semantic has several advantages over explicit semantics : typesafety is guaranteed, a thread is always started and stopped exaclty once and users cannot make any progamming errors. However, one of the apparent drawbacks of this system is that threads now always form a lattice, that is they are always destroyed in opposite order of construction. While this seems like a significant limitation, existing \CFA semantics can solve this problem. Indeed, by using dynamic allocation to create threads will naturally let threads outlive the scope in which the thread was created much like dynamically allocating memory will let objects outlive the scope in which thy were created :
1445
1446\begin{lstlisting}
1447        thread struct MyThread {
1448                //...
1449        };
1450
1451        //ctor
1452        void ?{}(MyThread* this,
1453                     bool is_special = false) {
1454                //...
1455        }
1456
1457        //main
1458        void ?main(MyThread* this) {
1459                //...
1460        }
1461
1462        void foo() {
1463                MyThread* special_thread;
1464                {
1465                        MyThread thrds = {false};
1466                        //Start a thread at the beginning of the scope
1467
1468                        DoStuff();
1469
1470                        //create a other thread that will outlive the thread in this scope
1471                        special_thread = new MyThread{true};
1472
1473                        //Wait for the thread to finish
1474                }
1475                DoMoreStuff();
1476
1477                //Now wait for the special
1478        }
1479\end{lstlisting}
1480
1481Another advantage of this semantic is that it naturally scale to multiple threads meaning basic synchronisation is very simple :
1482
1483\begin{lstlisting}
1484        thread struct MyThread {
1485                //...
1486        };
1487
1488        //ctor
1489        void ?{}(MyThread* this) {}
1490
1491        //main
1492        void ?main(MyThread* this) {
1493                //...
1494        }
1495
1496        void foo() {
1497                MyThread thrds[10];
1498                //Start 10 threads at the beginning of the scope
1499
1500                DoStuff();
1501
1502                //Wait for the 10 threads to finish
1503        }
1504\end{lstlisting}
1505
1506\newpage
1507\bf{WORK IN PROGRESS}
1508\subsection{The \CFA Kernel : Processors, Clusters and Threads}\label{kernel}
1509
1510
1511\subsection{Paradigms}\label{cfaparadigms}
1512Given these building blocks we can then reproduce the all three of the popular paradigms. Indeed, we get \glspl{uthread} as the default paradigm in \CFA. However, disabling \glspl{preemption} on the \gls{cfacluster} means \glspl{cfathread} effectively become \glspl{fiber}. Since several \glspl{cfacluster} with different scheduling policy can coexist in the same application, this allows \glspl{fiber} and \glspl{uthread} to coexist in the runtime of an application.
1513
1514% \subsection{High-level options}\label{tasks}
1515%
1516% \subsubsection{Thread interface}
1517% constructors destructors
1518%       initializer lists
1519% monitors
1520%
1521% \subsubsection{Futures}
1522%
1523% \subsubsection{Implicit threading}
1524% Finally, simpler applications can benefit greatly from having implicit parallelism. That is, parallelism that does not rely on the user to write concurrency. This type of parallelism can be achieved both at the language level and at the system level.
1525%
1526% \begin{center}
1527% \begin{tabular}[t]{|c|c|c|}
1528% Sequential & System Parallel & Language Parallel \\
1529% \begin{lstlisting}
1530% void big_sum(int* a, int* b,
1531%                int* out,
1532%                size_t length)
1533% {
1534%       for(int i = 0; i < length; ++i ) {
1535%               out[i] = a[i] + b[i];
1536%       }
1537% }
1538%
1539%
1540%
1541%
1542%
1543% int* a[10000];
1544% int* b[10000];
1545% int* c[10000];
1546% //... fill in a and b ...
1547% big_sum(a, b, c, 10000);
1548% \end{lstlisting} &\begin{lstlisting}
1549% void big_sum(int* a, int* b,
1550%                int* out,
1551%                size_t length)
1552% {
1553%       range ar(a, a + length);
1554%       range br(b, b + length);
1555%       range or(out, out + length);
1556%       parfor( ai, bi, oi,
1557%       [](int* ai, int* bi, int* oi) {
1558%               oi = ai + bi;
1559%       });
1560% }
1561%
1562% int* a[10000];
1563% int* b[10000];
1564% int* c[10000];
1565% //... fill in a and b ...
1566% big_sum(a, b, c, 10000);
1567% \end{lstlisting}&\begin{lstlisting}
1568% void big_sum(int* a, int* b,
1569%                int* out,
1570%                size_t length)
1571% {
1572%       for (ai, bi, oi) in (a, b, out) {
1573%               oi = ai + bi;
1574%       }
1575% }
1576%
1577%
1578%
1579%
1580%
1581% int* a[10000];
1582% int* b[10000];
1583% int* c[10000];
1584% //... fill in a and b ...
1585% big_sum(a, b, c, 10000);
1586% \end{lstlisting}
1587% \end{tabular}
1588% \end{center}
1589%
1590% \subsection{Machine setup}\label{machine}
1591% Threads are all good and well but wee still some OS support to fully utilize available hardware.
1592%
1593% \textbf{\large{Work in progress...}} Do wee need something beyond specifying the number of kernel threads?
1594
1595%    #    #       #
1596%   # #   #       #
1597%  #   #  #       #
1598% #     # #       #
1599% ####### #       #
1600% #     # #       #
1601% #     # ####### #######
1602\chapter{Putting it all together}
1603
1604
1605
1606
1607
1608\chapter{Conclusion}
1609
1610
1611
1612
1613
1614
1615% ####### #     # ####### #     # ######  #######
1616% #       #     #    #    #     # #     # #
1617% #       #     #    #    #     # #     # #
1618% #####   #     #    #    #     # ######  #####
1619% #       #     #    #    #     # #   #   #
1620% #       #     #    #    #     # #    #  #
1621% #        #####     #     #####  #     # ######
1622\chapter{Future work}
1623Concurrency and parallelism is still a very active field that strongly benefits from hardware advances. As such certain features that aren't necessarily mature enough in their current state could become relevant in the lifetime of \CFA.
1624\subsection{Transactions}
1625
1626% ####### #     # ######
1627% #       ##    # #     #
1628% #       # #   # #     #
1629% #####   #  #  # #     #
1630% #       #   # # #     #
1631% #       #    ## #     #
1632% ####### #     # ######
1633\section*{Acknowledgements}
1634
1635\clearpage
1636\printglossary[type=\acronymtype]
1637\printglossary
1638
1639\clearpage
1640\bibliographystyle{plain}
1641\bibliography{cw92,distSharedMem,lfp92,mlw92,parallel,parallelIO,partheory,pl,pldi92,ps,realtime,techreportsPAB,visual,local}
1642
1643
1644\end{document}
Note: See TracBrowser for help on using the repository browser.