Ignore:
Timestamp:
Aug 1, 2022, 3:27:07 PM (2 years ago)
Author:
Peter A. Buhr <pabuhr@…>
Branches:
ADT, ast-experimental, master, pthread-emulation
Children:
3fe4acd
Parents:
30159e5
Message:

proofread abstract

File:
1 edited

Legend:

Unmodified
Added
Removed
  • doc/theses/thierry_delisle_PhD/thesis/text/front.tex

    r30159e5 r4e21942  
    106106% D E C L A R A T I O N   P A G E
    107107% -------------------------------
    108 % The following is a sample Delaration Page as provided by the GSO
     108% The following is a sample Declaration Page as provided by the GSO
    109109% December 13th, 2006.  It is designed for an electronic thesis.
    110110\noindent
     
    124124
    125125User-Level threading (M:N) is gaining popularity over kernel-level threading (1:1) in many programming languages.
    126 The user-level approach is often a better mechanism to express complex concurrent applications by efficiently running 10,000+ threads on multi-core systems.
    127 Indeed, over-partitioning into small work-units significantly eases load balancing while providing user threads for each unit of work offers greater freedom to the programmer.
     126The user threading approach is often a better mechanism to express complex concurrent applications by efficiently running 10,000+ threads on multi-core systems.
     127Indeed, over-partitioning into small work-units with user threading significantly eases load bal\-ancing, while simultaneously providing advanced synchronization and mutual exclusion mechanisms.
    128128To manage these high levels of concurrency, the underlying runtime must efficiently schedule many user threads across a few kernel threads;
    129 which begs of the question of how many kernel threads are needed and when should the need be re-evaliated.
    130 Furthermore, the scheduler must prevent kernel threads from blocking, otherwise user-thread parallelism drops, and put idle kernel-threads to sleep to avoid wasted resources.
     129which begs of the question of how many kernel threads are needed and should the number be dynamically reevaluated.
     130Furthermore, scheduling must prevent kernel threads from blocking, otherwise user-thread parallelism drops.
     131When user-threading parallelism does drop, how and when should idle kernel-threads be to sleep to avoid wasted CPU resources.
    131132Finally, the scheduling system must provide fairness to prevent a user thread from monopolizing a kernel thread;
    132 otherwise other user threads can experience short/long term starvation or kernel threads can deadlock waiting for events to occur.
     133otherwise other user threads can experience short/long term starvation or kernel threads can deadlock waiting for events to occur on busy kernel threads.
    133134
    134135This thesis analyses multiple scheduler systems, where each system attempts to fulfill the necessary requirements for user-level threading.
    135 The predominant technique for manage high levels of concurrency is sharding the ready-queue with one queue per kernel-threads and using some form of work stealing/sharing to dynamically rebalance workload shifts.
    136 Fairness can be handled through preemption or ad-hoc solutions, which leads to coarse-grained fairness and pathological cases.
     136The predominant technique for managing high levels of concurrency is sharding the ready-queue with one queue per kernel-thread and using some form of work stealing/sharing to dynamically rebalance workload shifts.
    137137Preventing kernel blocking is accomplish by transforming kernel locks and I/O operations into user-level operations that do not block the kernel thread or spin up new kernel threads to manage the blocking.
    138 
    139 After selecting specific approaches to these scheduling issues, a complete implementation was created and tested in the \CFA (C-for-all) runtime system.
     138Fairness is handled through preemption and/or ad-hoc solutions, which leads to coarse-grained fairness with some pathological cases.
     139
     140After testing and selecting specific approaches to these scheduling issues, a complete implementation was created and tested in the \CFA (C-for-all) runtime system.
    140141\CFA is a modern extension of C using user-level threading as its fundamental threading model.
    141142As one of its primary goals, \CFA aims to offer increased safety and productivity without sacrificing performance.
    142143The new scheduler achieves this goal by demonstrating equivalent performance to work-stealing schedulers while offering better fairness.
    143 This is achieved through several optimization that successfully eliminate the cost of the additional fairness, some of these optimization relying on interesting hardware optimizations present on most modern cpus.
    144 This work also includes support for user-level \io, allowing programmers to have many more user-threads blocking on \io operations than there are \glspl{kthrd}.
     144The implementation uses several optimizations that successfully balance the cost of fairness against performance;
     145some of these optimization rely on interesting hardware optimizations present on modern CPUs.
     146The new scheduler also includes support for implicit nonblocking \io, allowing applications to have more user-threads blocking on \io operations than there are \glspl{kthrd}.
    145147The implementation is based on @io_uring@, a recent addition to the Linux kernel, and achieves the same performance and fairness.
    146 To complete the picture, the idle sleep mechanism that goes along is presented.
    147 
    148 
     148To complete the scheduler, an idle sleep mechanism is implemented that significantly reduces wasted CPU cycles, which are then available outside of the application.
    149149
    150150\cleardoublepage
Note: See TracChangeset for help on using the changeset viewer.