323:(if any), and are thus relatively cheap to create and destroy. Thread switching is also relatively cheap: it requires a context switch (saving and restoring registers and stack pointer), but does not change virtual memory and is thus cache-friendly (leaving TLB valid). The kernel can assign one or more software threads to each core in a CPU (it being able to assign itself multiple software threads depending on its support for multithreading), and can swap out threads that get blocked. However, kernel threads take much longer than user threads to be swapped.
61:
1248:(GIL). The GIL is a mutual exclusion lock held by the interpreter that can prevent the interpreter from simultaneously interpreting the application's code on two or more threads at once. This effectively limits the parallelism on multiple core systems. It also limits performance for processor-bound threads (which require the processor), but doesn't effect I/O-bound or network-bound ones as much. Other implementations of interpreted programming languages, such as
153:
50:
2654:
801:" threading systems are more complex to implement than either kernel or user threads, because changes to both kernel and user-space code are required. In the M:N implementation, the threading library is responsible for scheduling user threads on the available schedulable entities; this makes context switching of threads very fast, as it avoids system calls. However, this increases complexity and the likelihood of
1291:
742::1 model implies that all application-level threads map to one kernel-level scheduled entity; the kernel has no knowledge of the application threads. With this approach, context switching can be done very quickly and, in addition, it can be implemented even on simple kernels which do not support threading. One of the major drawbacks, however, is that it cannot benefit from the hardware acceleration on
363:
problem is when performing I/O: most programs are written to perform I/O synchronously. When an I/O operation is initiated, a system call is made, and does not return until the I/O operation has been completed. In the intervening period, the entire process is "blocked" by the kernel and cannot run, which starves other user threads and fibers in the same process from executing.
1135:. In general, multithreaded programs are non-deterministic, and as a result, are untestable. In other words, a multithreaded program can easily have bugs which never manifest on a test system, manifesting only in production. This can be alleviated by restricting inter-thread communications to certain well-defined patterns (such as message-passing).
885:
implement LWPs as kernel threads (1:1 model). SunOS 5.2 through SunOS 5.8 as well as NetBSD 2 to NetBSD 4 implemented a two level model, multiplexing one or more user level threads on each kernel thread (M:N model). SunOS 5.9 and later, as well as NetBSD 5 eliminated user threads support, returning
366:
A common solution to this problem (used, in particular, by many green threads implementations) is providing an I/O API that implements an interface that blocks the calling thread, rather than the entire process, by using non-blocking I/O internally, and scheduling another user thread or fiber while
358:
As user thread implementations are typically entirely in userspace, context switching between user threads within the same process is extremely efficient because it does not require any interaction with the kernel at all: a context switch can be performed by locally saving the CPU registers used by
1032:
where a set number of threads are created at startup that then wait for a task to be assigned. When a new task arrives, it wakes up, completes the task and goes back to waiting. This avoids the relatively expensive thread creation and destruction functions for every task performed and takes thread
921:
Multithreading is mainly found in multitasking operating systems. Multithreading is a widespread programming and execution model that allows multiple threads to exist within the context of one process. These threads share the process's resources, but are able to execute independently. The threaded
1128:
has written: "Although threads seem to be a small step from sequential computation, in fact, they represent a huge step. They discard the most essential and appealing properties of sequential computation: understandability, predictability, and determinism. Threads, as a model of computation, are
933:
Multithreading libraries tend to provide a function call to create a new thread, which takes a function as a parameter. A concurrent thread is then created which starts running the passed function and ends when the function returns. The thread libraries also offer data synchronization functions.
362:
However, the use of blocking system calls in user threads (as opposed to kernel threads) can be problematic. If a user thread or a fiber performs a system call that blocks, the other user threads and fibers in the process are unable to run until the system call returns. A typical example of this
367:
the I/O operation is in progress. Similar solutions can be provided for other blocking system calls. Alternatively, the program can be written to avoid the use of synchronous I/O or other blocking system calls (in particular, using non-blocking I/O, including lambda continuations and/or async/
1067:: applications looking to use multicore or multi-CPU systems can use multithreading to split data and tasks into parallel subtasks and let the underlying architecture manage how the threads run, either concurrently on one core or in parallel on multiple cores. GPU computing environments like
233:, which share the process's resources, such as memory and file handles – a process is a unit of resources, while a thread is a unit of scheduling and execution. Kernel scheduling is typically uniformly done preemptively or, less commonly, cooperatively. At the user level a process such as a
400:. A fiber can be scheduled to run in any thread in the same process. This permits applications to gain performance improvements by managing scheduling themselves, instead of relying on the kernel scheduler (which may not be tuned for the application). Some research implementations of the
1212:
languages, expose threading to developers while abstracting the platform specific differences in threading implementations in the runtime. Several other programming languages and language extensions also try to abstract the concept of concurrency and threading from the developer fully
302:
is a "lightweight" unit of kernel scheduling. At least one kernel thread exists within each process. If multiple kernel threads exist within a process, then they share the same memory and file resources. Kernel threads are preemptively multitasked if the operating system's process
194:
The use of threads in software applications became more common in the early 2000s as CPUs began to utilize multiple cores. Applications wishing to take advantage of multiple cores for performance advantages were required to employ concurrency to utilize the multiple cores.
886:
to a 1:1 model. FreeBSD 5 implemented M:N model. FreeBSD 6 supported both 1:1 and M:N, users could choose which one should be used with a given program using /etc/libmap.conf. Starting with FreeBSD 7, the 1:1 became the default. FreeBSD 8 no longer supports the M:N model.
1052:
that runs concurrently with the main execution thread, it is possible for the application to remain responsive to user input while executing tasks in the background. On the other hand, in most cases multithreading is not the only way to keep a program responsive, with
359:
the currently executing user thread or fiber and then loading the registers required by the user thread or fiber to be executed. Since scheduling occurs in userspace, the scheduling policy can be more easily tailored to the requirements of the program's workload.
962:
if they require more than one CPU instruction to update: two threads may end up attempting to update the data structure at the same time and find it unexpectedly changing underfoot. Bugs caused by race conditions can be very difficult to reproduce and isolate.
1141:. As thread context switch on modern CPUs can cost up to 1 million CPU cycles, it makes writing efficient multithreading programs difficult. In particular, special attention has to be paid to avoid inter-thread synchronization from being too frequent.
651:
usually occurs frequently enough that users perceive the threads or tasks as running in parallel (for popular server/desktop operating systems, maximum time slice of a thread, when other threads are waiting, is often limited to 100–200ms). On a
1048:: multithreading can allow an application to remain responsive to input. In a one-thread program, if the main execution thread blocks on a long-running task, the entire application can appear to freeze. By moving such long-running tasks to a
1267:
in parallel using only its ID to find its data in memory. In essence, the application must be designed so that each thread performs the same operation on different segments of memory so that they can operate in parallel and use the GPU
524:: due to threads sharing the same address space, an illegal operation performed by a thread can crash the entire process; therefore, one misbehaving thread can disrupt the processing of all the other threads in the application.
750:
computers: there is never more than one thread being scheduled at the same time. For example: If one of the threads needs to execute an I/O request, the whole process is blocked and the threading advantage cannot be used. The
982:
data structures against concurrent access. On uniprocessor systems, a thread running into a locked mutex must sleep and hence trigger a context switch. On multi-processor systems, the thread may instead poll the mutex in a
282:. Creating or destroying a process is relatively expensive, as resources must be acquired or released. Processes are typically preemptively multitasked, and process switching is relatively expensive, beyond basic cost of
1162:) as early as in the late 1960s, and this was continued in the Optimizing Compiler and later versions. The IBM Enterprise PL/I compiler introduced a new model "thread" API. Neither version was part of the PL/I standard.
1177:(Pthreads), which is a set of C-function library calls. OS vendors are free to implement the interface as desired, but the application developer should be able to use the same interface across multiple platforms. Most
278:, and do not share address spaces or file resources except through explicit methods such as inheriting file handles or shared memory segments, or mapping the same file in a shared way – see
1252:
using the Thread extension, avoid the GIL limit by using an
Apartment model where data and code must be explicitly "shared" between threads. In Tcl each thread has one or more interpreters.
237:
can itself schedule multiple threads of execution. If these do not share data, as in Erlang, they are usually analogously called processes, while if they share data they are usually called
1980:
1543:
339:. The kernel is unaware of them, so they are managed and scheduled in userspace. Some implementations base their user threads on top of several kernel threads, to benefit from
247:; different processes may schedule user threads differently. User threads may be executed by kernel threads in various ways (one-to-one, many-to-one, many-to-many). The term "
2744:
286:, due to issues such as cache flushing (in particular, process switching changes virtual memory addressing, causing invalidation and thus flushing of an untagged
2070:
1083:. This, in turn, enables better system utilization, and (provided that synchronization costs don't eat the benefits up), can provide faster program execution.
1922:
2691:
1033:
management out of the application developer's hand and leaves it to a library or the operating system that is better suited to optimize thread management.
119:
922:
programming model provides developers with a useful abstraction of concurrent execution. Multithreading can also be applied to one process to enable
1116:) to prevent common data from being read or overwritten in one thread while being modified by another. Careless use of such primitives can lead to
1173:
support threading, and provide access to the native threading APIs of the operating system. A standardized interface for thread implementation is
122:, while different processes do not share these resources. In particular, the threads of a process share its executable code and the values of its
683:
Threads created by the user in a 1:1 correspondence with schedulable entities in the kernel are the simplest possible threading implementation.
2051:
28:
1660:
3020:
2991:
1726:
1695:
1581:
1520:
2318:
805:, as well as suboptimal scheduling without extensive (and expensive) coordination between the userland scheduler and the kernel scheduler.
1225:(MPI)). Some languages are designed for sequential parallelism instead (especially using GPUs), without requiring concurrency or threads (
2341:
2230:
1325:
1310:
743:
258:
is a "heavyweight" unit of kernel scheduling, as creating, destroying, and switching processes is relatively expensive. Processes own
598:, although threads were still used on such computers because switching between threads was generally still quicker than full-process
2336:
2313:
1894:
1880:
1866:
1852:
1838:
1644:
1438:
967:
2795:
2739:
1915:
1557:
Ferat, Manuel; Pereira, Romain; Roussel, Adrien; Carribault, Patrick; Steffenel, Luiz-Angelo; Gautier, Thierry (September 2022).
2714:
2684:
2308:
2123:
1193:
2907:
2805:
2415:
2278:
1360:
856:
504:
of threads: using threads, an application can operate using fewer resources than it would need when using multiple processes.
3206:
3185:
2734:
2719:
2639:
2473:
2091:
2011:
1271:
1205:
583:
3211:
2780:
2765:
2724:
1568:. IWOMP 2022: 18th International Workshop on OpenMP. Lecture Notes in Computer Science. Vol. 13527. pp. 3–16.
907:
571:
491:
408:, with the distinction being that coroutines are a language-level construct, while fibers are a system-level construct.
287:
789:
number of kernel entities, or "virtual processors." This is a compromise between kernel-level ("1:1") and user-level ("
2946:
2893:
2658:
2604:
2064:
1908:
1345:
1296:
1201:
1117:
955:
951:
700:
607:
579:
515:
453:
440:
259:
2961:
2800:
2677:
2583:
2378:
2263:
2225:
2075:
1965:
1558:
832:
708:
567:
279:
558:. However, preemptive scheduling may context-switch threads at moments unanticipated by programmers, thus causing
2996:
2815:
2775:
2770:
2729:
2599:
2578:
2523:
2410:
2400:
2373:
2235:
1305:
1260:
1222:
1011:
988:
971:
551:
3039:
2926:
2790:
2553:
2179:
2118:
2031:
1166:
852:
543:
432:
389:
385:
380:
243:
208:
60:
1685:
2468:
1104:
and other non-intuitive behaviors. In order for data to be correctly manipulated, threads will often need to
2785:
2614:
2609:
2059:
1245:
1113:
1007:
992:
979:
640:
251:" variously refers to user threads or to kernel mechanisms for scheduling user threads onto kernel threads.
40:
1181:
platforms, including Linux, support
Pthreads. Microsoft Windows has its own set of thread functions in the
3173:
3112:
3001:
2981:
2930:
2888:
2353:
2285:
2189:
2081:
2036:
1537:
1330:
1320:
1278:
have a different threading model that supports extremely large numbers of threads (for modeling hardware).
915:
636:
539:
393:
304:
204:
115:
100:
77:
73:
2143:
2956:
2922:
2824:
2760:
2445:
2405:
2358:
2348:
2086:
2006:
1945:
1315:
814:
417:
267:
184:
96:
54:
1486:
1385:
664:, with every processor or core executing a separate thread simultaneously; on a processor or core with
1811:
3153:
3127:
2385:
2273:
2268:
2258:
2245:
2041:
1718:
1632:
1458:
895:
873:
828:
752:
575:
320:
248:
127:
1522:
Eight Ways to Handle Non-blocking
Returns in Message-passing Programs: from C++98 via C++11 to C++20
1244:
for Python) which support threading and concurrency but not parallel execution of threads, due to a
1129:
wildly non-deterministic, and the job of the programmer becomes one of pruning that nondeterminism."
462:
between threads in the same process typically occurs faster than context switching between processes
3122:
3074:
2951:
2548:
2503:
2329:
2324:
2303:
2169:
1796:
1791:
1771:
1753:
1748:
1667:
1426:
1350:
1029:
1023:
999:
950:
Threads in the same process share the same address space. This allows concurrently running code to
903:
435:
information than threads, whereas multiple threads within a process share process state as well as
421:
220:
188:
137:
108:
68:
1606:
482:
processes; in other operating systems there is not so great a difference except in the cost of an
3059:
2573:
2422:
2395:
2220:
2184:
1975:
1955:
1950:
1931:
1587:
1408:
1335:
1105:
923:
802:
661:
563:
312:
263:
2133:
594:
Until the early 2000s, most desktop computers had only one single-core CPU, with no support for
1559:"Enhancing MPI+OpenMP Task Based Applications for Heterogeneous Architectures with GPU support"
404:
parallel programming model implement their tasks through fibers. Closely related to fibers are
3168:
3117:
3049:
3006:
2847:
2619:
2295:
2253:
2148:
1890:
1876:
1862:
1848:
1834:
1786:
1722:
1712:
1691:
1640:
1577:
1434:
914:
can be used differently to mean "backtracking within a single thread", which is common in the
283:
275:
123:
396:" to allow another fiber to run, which makes their implementation much easier than kernel or
46:
Smallest sequence of programmed instructions that can be managed independently by a scheduler
3148:
2700:
2629:
2428:
2363:
2210:
2026:
2021:
2016:
1985:
1569:
1400:
1109:
1096:
complexity and related bugs: when using shared resources typical for threaded programs, the
1076:
1054:
1003:
670:, separate software threads can also be executed concurrently by separate hardware threads.
547:
436:
241:, particularly if preemptively scheduled. Cooperatively scheduled user threads are known as
104:
88:
64:
1505:
99:
is the smallest sequence of programmed instructions that can be managed independently by a
3092:
3054:
3025:
2493:
2433:
2368:
2215:
2205:
2138:
1970:
1960:
927:
747:
666:
616:
595:
511:
348:
340:
316:
130:
2128:
1614:. The 28th International Conference on Parallel Architectures and Compilation Techniques.
17:
3178:
3102:
3064:
2936:
2624:
2440:
2097:
1990:
1381:
1264:
1209:
1197:
1121:
1101:
959:
653:
648:
599:
555:
459:
234:
203:
Scheduling can be done at the kernel level or user level, and multitasking can be done
81:
262:
allocated by the operating system. Resources include memory (for both code and data),
152:
49:
3200:
3044:
2883:
2837:
2513:
2390:
1591:
1454:
1386:"How to Make a Multiprocessor Computer That Correctly Executes Multiprocess Programs"
1355:
1174:
1125:
945:
882:
817:
used by older versions of the NetBSD native POSIX threads library implementation (an
756:
696:
483:
446:
352:
36:
1108:
in time in order to process the data in the correct order. Threads may also require
958:. When shared between threads, however, even simple data structures become prone to
2971:
2113:
1412:
1340:
1058:
704:
629:
27:
This article is about the concurrency concept. For multithreading in hardware, see
1605:
Iwasaki, Shintaro; Amer, Abdelhalim; Taura, Kenjiro; Seo, Sangmin; Balaji, Pavan.
1481:
1041:
Multithreaded applications have the following advantages vs single-threaded ones:
586:
other threads by not yielding control of execution during intensive computation.
570:
relies on threads to relinquish control of execution, thus ensuring that threads
3097:
3079:
2862:
2852:
2842:
2634:
1573:
1477:
1290:
1186:
954:
tightly and conveniently exchange data without the overhead or complexity of an
559:
428:
processes are typically independent, while threads exist as subsets of a process
518:(IPC), threads can communicate through data, code and files they already share.
1286:
1097:
846:
657:
467:
332:
308:
32:
3034:
2941:
2867:
2832:
2508:
2483:
1404:
1182:
621:
611:
405:
1482:"The Free Lunch Is Over: A Fundamental Turn Toward Concurrency in Software"
1743:
3163:
2558:
2538:
2463:
1237:
1226:
1080:
984:
859:
uses lightweight threads which are scheduled on operating system threads.
1075:
use the multithreading model where dozens to hundreds of threads run in
3158:
3087:
2857:
2563:
2543:
2518:
2153:
1566:
OpenMP in a Modern World: From Multi-device
Support to Meta Programming
1275:
1241:
716:
635:
Systems with a single processor generally implement multithreading by
2533:
2528:
1900:
1218:
1072:
878:
712:
401:
2669:
1236:
A few interpreted programming languages have implementations (e.g.,
1170:
825:
model as opposed to a 1:1 kernel or userspace implementation model)
3107:
975:
868:
720:
692:
688:
603:
368:
48:
2568:
2498:
2488:
1608:
BOLT: Optimizing OpenMP Parallel
Regions with User-Level Threads
1256:
1230:
1214:
1178:
1155:
1068:
1037:
Multithreaded programs vs single-threaded programs pros and cons
684:
574:. This can cause problems if a cooperatively multitasked thread
471:
307:
is preemptive. Kernel threads do not own resources except for a
2673:
1904:
1787:"Multi-threading at Business-logic Level is Considered Harmful"
1150:
Many programming languages support threading in some capacity.
991:(SMP) systems to contend for the memory bus, especially if the
3143:
2478:
2455:
1249:
839:
724:
625:
497:
Advantages and disadvantages of threads vs processes include:
487:
183:
Threads made an early appearance under the name of "tasks" in
147:
118:(via multithreading capabilities), sharing resources such as
987:. Both of these may sap performance and force processors in
1028:
A popular programming pattern involving threads is that of
1843:
Bradford
Nichols, Dick Buttlar, Jacqueline Proulx Farell:
1087:
Multithreaded applications have the following drawbacks:
1639:(9th ed.). Hoboken, N.J.: Wiley. pp. 170–171.
114:
The multiple threads of a given process may be executed
53:
A process with two threads of execution, running on one
185:
OS/360 Multiprogramming with a
Variable Number of Tasks
164:
554:
for its finer-grained control over execution time via
1661:"Multithreading in the Solaris Operating Environment"
906:
at a time. In the formal analysis of the variables'
3136:
3073:
3019:
2980:
2915:
2906:
2876:
2823:
2814:
2753:
2707:
2592:
2454:
2294:
2244:
2198:
2162:
2106:
2050:
1999:
1938:
1525:. CPPCON. Archived from the original on 2020-11-25
388:are an even lighter unit of scheduling which are
1542:: CS1 maint: bot: original URL status unknown (
1460:Traffic Control in a Multiplexed Computer System
1158:(F) included support for multithreading (called
452:processes interact only through system-provided
1711:O'Hearn, Peter William; Tennent, R. D. (1997).
1061:being available for obtaining similar results.
510:of threads: unlike processes, which require a
107:. In many cases, a thread is a component of a
31:. For the form of code consisting entirely of
2685:
1916:
1690:. Mike Murach & Associates. p. 512.
486:switch, which on some architectures (notably
211:. This yields a variety of related concepts.
8:
1627:
1625:
1623:
1621:
691:used this approach from the start, while on
864:History of threading models in Unix systems
449:, whereas threads share their address space
2912:
2820:
2692:
2678:
2670:
1923:
1909:
1901:
1635:; Galvin, Peter Baer; Gagne, Greg (2013).
538:Operating systems schedule threads either
266:, sockets, device handles, windows, and a
890:Single-threaded vs multithreaded programs
620:; in 2005, they introduced the dual-core
1433:. Prentice-Hall International Editions.
785:number of application threads onto some
660:system, multiple threads can execute in
59:
1744:"Single-Threading: Back to the Future?"
1466:(Doctor of Science thesis). p. 20.
1373:
1765:
1763:
1687:Murach's CICS for the COBOL Programmer
1535:
566:, or other side-effects. In contrast,
514:or shared memory mechanism to perform
187:(MVT) in 1967. Saltzer (1966) credits
29:Multithreading (computer architecture)
1812:"Operation Costs in CPU Clock Cycles"
344:
331:Threads are sometimes implemented in
7:
1873:Multithreading Applications in Win32
1810:'No Bugs' Hare (12 September 2016).
1112:operations (often implemented using
1092:
534:Preemptive vs cooperative scheduling
508:Simplified sharing and communication
140:differs between operating systems.
1785:Ignatchenko, Sergey (August 2015).
1742:Ignatchenko, Sergey (August 2010).
1684:Menéndez, Raúl; Lowe, Doug (2001).
1185:interface for multithreading, like
998:Other synchronization APIs include
755:uses User-level threading, as does
392:: a running fiber must explicitly "
103:, which is typically a part of the
1326:Multithreading (computer hardware)
1311:Communicating sequential processes
968:application programming interfaces
699:implements this approach (via the
590:Single- vs multi-processor systems
431:processes carry considerably more
347:). User threads as implemented by
136:The implementation of threads and
25:
1887:Unix Internals: the New Frontiers
1200:) programming languages, such as
707:). This approach is also used by
643:(CPU) switches between different
397:
2796:Object-oriented operating system
2653:
2652:
1770:Lee, Edward (January 10, 2006).
1289:
938:Threads and data synchronization
416:Threads differ from traditional
151:
2124:Analysis of parallel algorithms
2806:Supercomputer operating system
1871:Jim Beveridge, Robert Wiener:
1831:Programming with POSIX Threads
1393:IEEE Transactions on Computers
1361:Win32 Thread Information Block
1272:Hardware description languages
1255:In programming models such as
831:used by older versions of the
809:Hybrid implementation examples
1:
2071:Simultaneous and heterogenous
1847:, O'Reilly & Associates,
793::1") threading. In general, "
2781:Just enough operating system
2766:Distributed operating system
2659:Category: Parallel computing
1146:Programming language support
995:of the locking is too fine.
910:and process state, the term
679:1:1 (kernel-level threading)
548:Multi-user operating systems
492:translation lookaside buffer
288:translation lookaside buffer
2894:User space and kernel space
1574:10.1007/978-3-031-15922-0_1
1346:Simultaneous multithreading
1297:Computer programming portal
966:To prevent this, threading
608:simultaneous multithreading
516:inter-process communication
454:inter-process communication
3228:
2801:Real-time operating system
1966:High-performance computing
1772:"The Problem with Threads"
1263:, an array of threads run
1021:
972:synchronization primitives
943:
614:processor, under the name
568:cooperative multithreading
502:Lower resource consumption
378:
280:interprocess communication
218:
26:
2997:Multilevel feedback queue
2992:Fixed-priority preemptive
2776:Hobbyist operating system
2771:Embedded operating system
2648:
2600:Automatic parallelization
2236:Application checkpointing
1637:Operating system concepts
1306:Clone (Linux system call)
1261:data parallel computation
1223:Message Passing Interface
1100:must be careful to avoid
989:symmetric multiprocessing
902:is the processing of one
734::1 (user-level threading)
628:introduced the dual-core
552:preemptive multithreading
18:Thread (computer science)
3040:General protection fault
2791:Network operating system
2745:User features comparison
1431:Modern Operating Systems
1165:Many implementations of
853:Glasgow Haskell Compiler
522:Thread crashes a process
445:processes have separate
381:Fiber (computer science)
290:(TLB), notably on x86).
191:with the term "thread".
2786:Mobile operating system
2615:Embarrassingly parallel
2610:Deterministic algorithm
1859:Java Thread Programming
1506:"Erlang: 3.1 Processes"
1405:10.1109/tc.1979.1675439
1246:global interpreter lock
855:(GHC) for the language
641:central processing unit
390:cooperatively scheduled
335:libraries, thus called
225:At the kernel level, a
41:Thread (disambiguation)
2889:Loadable kernel module
2330:Associative processing
2286:Non-blocking algorithm
2092:Clustered multi-thread
1666:. 2002. Archived from
1455:Saltzer, Jerome Howard
1331:Non-blocking algorithm
1321:Multi-core (computing)
916:functional programming
874:light-weight processes
829:Light-weight processes
630:Athlon 64 X2
84:
57:
39:. For other uses, see
2957:Process control block
2923:Computer multitasking
2761:Disk operating system
2446:Hardware acceleration
2359:Superscalar processor
2349:Dataflow architecture
1946:Distributed computing
1673:on February 26, 2009.
1633:Silberschatz, Abraham
1519:Ignatchenko, Sergey.
1316:Computer multitasking
1139:Synchronization costs
1081:large number of cores
815:Scheduler activations
268:process control block
229:contains one or more
124:dynamically allocated
63:
52:
3207:Concurrent computing
3128:Virtual tape library
2720:Forensic engineering
2325:Pipelined processing
2274:Explicit parallelism
2269:Implicit parallelism
2259:Dataflow programming
1845:Pthreads Programming
1714:ALGOL-like languages
1427:Tanenbaum, Andrew S.
1077:parallel across data
896:computer programming
845:The OS for the Tera-
753:GNU Portable Threads
412:Threads vs processes
321:thread-local storage
249:light-weight process
133:at any given time.
3212:Threads (computing)
3137:Supporting concepts
3123:Virtual file system
2549:Parallel Extensions
2354:Pipelined processor
1829:David R. Butenhof:
1351:Thread pool pattern
1124:over resources. As
1024:Thread pool pattern
1000:condition variables
221:Process (computing)
189:Victor A. Vyssotsky
3060:Segmentation fault
2908:Process management
2423:Massively parallel
2401:distributed shared
2221:Cache invalidation
2185:Instruction window
1976:Manycore processor
1956:Massively parallel
1951:Parallel computing
1932:Parallel computing
1875:, Addison-Wesley,
1833:, Addison-Wesley,
1487:Dr. Dobb's Journal
1384:(September 1979).
1336:Priority inversion
1110:mutually exclusive
924:parallel execution
803:priority inversion
770:(hybrid threading)
606:added support for
564:priority inversion
163:. You can help by
126:variables and non-
85:
58:
3194:
3193:
3050:Memory protection
3021:Memory management
3015:
3014:
3007:Shortest job next
2902:
2901:
2701:Operating systems
2667:
2666:
2620:Parallel slowdown
2254:Stream processing
2144:Karp–Flatt metric
1889:, Prentice Hall,
1728:978-0-8176-3937-2
1719:Birkhäuser Verlag
1697:978-1-890774-09-7
1583:978-3-031-15921-3
1004:critical sections
649:context switching
572:run to completion
556:context switching
474:are said to have
460:context switching
424:in several ways:
420:operating-system
284:context switching
276:process isolation
181:
180:
82:Context Switching
16:(Redirected from
3219:
3149:Computer network
2913:
2821:
2694:
2687:
2680:
2671:
2656:
2655:
2630:Software lockout
2429:Computer cluster
2364:Vector processor
2319:Array processing
2304:Flynn's taxonomy
2211:Memory coherence
1986:Computer network
1925:
1918:
1911:
1902:
1816:
1815:
1807:
1801:
1800:
1782:
1776:
1775:
1767:
1758:
1757:
1739:
1733:
1732:
1708:
1702:
1701:
1681:
1675:
1674:
1672:
1665:
1657:
1651:
1650:
1629:
1616:
1615:
1613:
1602:
1596:
1595:
1563:
1554:
1548:
1547:
1541:
1533:
1531:
1530:
1516:
1510:
1509:
1502:
1496:
1495:
1474:
1468:
1467:
1465:
1451:
1445:
1444:
1423:
1417:
1416:
1390:
1378:
1299:
1294:
1293:
1133:Being untestable
1055:non-blocking I/O
912:single threading
900:single-threading
871:4.x implemented
838:Marcel from the
835:operating system
674:Threading models
667:hardware threads
645:software threads
600:context switches
596:hardware threads
578:by waiting on a
550:generally favor
466:Systems such as
351:are also called
349:virtual machines
311:, a copy of the
270:. Processes are
199:Related concepts
176:
173:
155:
148:
131:global variables
105:operating system
89:computer science
21:
3227:
3226:
3222:
3221:
3220:
3218:
3217:
3216:
3197:
3196:
3195:
3190:
3132:
3093:Defragmentation
3078:
3069:
3055:Protection ring
3024:
3011:
2983:
2976:
2898:
2872:
2810:
2749:
2703:
2698:
2668:
2663:
2644:
2588:
2494:Coarray Fortran
2450:
2434:Beowulf cluster
2290:
2240:
2231:Synchronization
2216:Cache coherence
2206:Multiprocessing
2194:
2158:
2139:Cost efficiency
2134:Gustafson's law
2102:
2046:
1995:
1971:Multiprocessing
1961:Cloud computing
1934:
1929:
1899:
1885:Uresh Vahalia:
1825:
1823:Further reading
1820:
1819:
1809:
1808:
1804:
1784:
1783:
1779:
1769:
1768:
1761:
1741:
1740:
1736:
1729:
1721:. p. 157.
1717:. Vol. 2.
1710:
1709:
1705:
1698:
1683:
1682:
1678:
1670:
1663:
1659:
1658:
1654:
1647:
1631:
1630:
1619:
1611:
1604:
1603:
1599:
1584:
1561:
1556:
1555:
1551:
1534:
1528:
1526:
1518:
1517:
1513:
1504:
1503:
1499:
1476:
1475:
1471:
1463:
1453:
1452:
1448:
1441:
1425:
1424:
1420:
1388:
1382:Lamport, Leslie
1380:
1379:
1375:
1370:
1365:
1295:
1288:
1285:
1148:
1120:, livelocks or
1102:race conditions
1093:Synchronization
1065:Parallelization
1039:
1026:
1020:
960:race conditions
948:
940:
928:multiprocessing
892:
866:
811:
788:
784:
772:
748:multi-processor
736:
681:
676:
617:hyper-threading
592:
536:
531:
512:message passing
490:) results in a
414:
383:
377:
341:multi-processor
329:
317:program counter
296:
223:
217:
201:
177:
171:
168:
161:needs expansion
146:
72:
47:
44:
23:
22:
15:
12:
11:
5:
3225:
3223:
3215:
3214:
3209:
3199:
3198:
3192:
3191:
3189:
3188:
3183:
3182:
3181:
3179:User interface
3176:
3166:
3161:
3156:
3151:
3146:
3140:
3138:
3134:
3133:
3131:
3130:
3125:
3120:
3115:
3110:
3105:
3103:File attribute
3100:
3095:
3090:
3084:
3082:
3071:
3070:
3068:
3067:
3065:Virtual memory
3062:
3057:
3052:
3047:
3042:
3037:
3031:
3029:
3017:
3016:
3013:
3012:
3010:
3009:
3004:
2999:
2994:
2988:
2986:
2978:
2977:
2975:
2974:
2969:
2964:
2959:
2954:
2949:
2944:
2939:
2937:Context switch
2934:
2919:
2917:
2910:
2904:
2903:
2900:
2899:
2897:
2896:
2891:
2886:
2880:
2878:
2874:
2873:
2871:
2870:
2865:
2860:
2855:
2850:
2845:
2840:
2835:
2829:
2827:
2818:
2812:
2811:
2809:
2808:
2803:
2798:
2793:
2788:
2783:
2778:
2773:
2768:
2763:
2757:
2755:
2751:
2750:
2748:
2747:
2742:
2737:
2732:
2727:
2722:
2717:
2711:
2709:
2705:
2704:
2699:
2697:
2696:
2689:
2682:
2674:
2665:
2664:
2662:
2661:
2649:
2646:
2645:
2643:
2642:
2637:
2632:
2627:
2625:Race condition
2622:
2617:
2612:
2607:
2602:
2596:
2594:
2590:
2589:
2587:
2586:
2581:
2576:
2571:
2566:
2561:
2556:
2551:
2546:
2541:
2536:
2531:
2526:
2521:
2516:
2511:
2506:
2501:
2496:
2491:
2486:
2481:
2476:
2471:
2466:
2460:
2458:
2452:
2451:
2449:
2448:
2443:
2438:
2437:
2436:
2426:
2420:
2419:
2418:
2413:
2408:
2403:
2398:
2393:
2383:
2382:
2381:
2376:
2369:Multiprocessor
2366:
2361:
2356:
2351:
2346:
2345:
2344:
2339:
2334:
2333:
2332:
2327:
2322:
2311:
2300:
2298:
2292:
2291:
2289:
2288:
2283:
2282:
2281:
2276:
2271:
2261:
2256:
2250:
2248:
2242:
2241:
2239:
2238:
2233:
2228:
2223:
2218:
2213:
2208:
2202:
2200:
2196:
2195:
2193:
2192:
2187:
2182:
2177:
2172:
2166:
2164:
2160:
2159:
2157:
2156:
2151:
2146:
2141:
2136:
2131:
2126:
2121:
2116:
2110:
2108:
2104:
2103:
2101:
2100:
2098:Hardware scout
2095:
2089:
2084:
2079:
2073:
2068:
2062:
2056:
2054:
2052:Multithreading
2048:
2047:
2045:
2044:
2039:
2034:
2029:
2024:
2019:
2014:
2009:
2003:
2001:
1997:
1996:
1994:
1993:
1991:Systolic array
1988:
1983:
1978:
1973:
1968:
1963:
1958:
1953:
1948:
1942:
1940:
1936:
1935:
1930:
1928:
1927:
1920:
1913:
1905:
1898:
1897:
1883:
1869:
1855:
1841:
1826:
1824:
1821:
1818:
1817:
1802:
1777:
1774:. UC Berkeley.
1759:
1734:
1727:
1703:
1696:
1676:
1652:
1645:
1617:
1597:
1582:
1549:
1511:
1497:
1480:(March 2005).
1469:
1446:
1439:
1418:
1399:(9): 690–691.
1372:
1371:
1369:
1366:
1364:
1363:
1358:
1353:
1348:
1343:
1338:
1333:
1328:
1323:
1318:
1313:
1308:
1302:
1301:
1300:
1284:
1281:
1280:
1279:
1269:
1253:
1234:
1210:.NET Framework
1198:cross-platform
1190:
1163:
1147:
1144:
1143:
1142:
1136:
1130:
1085:
1084:
1062:
1046:Responsiveness
1038:
1035:
1022:Main article:
1019:
1016:
944:Main article:
939:
936:
891:
888:
865:
862:
861:
860:
849:
843:
836:
826:
810:
807:
786:
782:
771:
761:
746:processors or
735:
729:
680:
677:
675:
672:
654:multiprocessor
624:processor and
622:Pentium D
612:Pentium 4
591:
588:
535:
532:
530:
527:
526:
525:
519:
505:
464:
463:
457:
450:
447:address spaces
443:
429:
413:
410:
379:Main article:
376:
373:
328:
325:
315:including the
295:
294:Kernel threads
292:
239:(user) threads
235:runtime system
231:kernel threads
219:Main article:
216:
213:
200:
197:
179:
178:
158:
156:
145:
142:
45:
24:
14:
13:
10:
9:
6:
4:
3:
2:
3224:
3213:
3210:
3208:
3205:
3204:
3202:
3187:
3184:
3180:
3177:
3175:
3172:
3171:
3170:
3167:
3165:
3162:
3160:
3157:
3155:
3152:
3150:
3147:
3145:
3142:
3141:
3139:
3135:
3129:
3126:
3124:
3121:
3119:
3116:
3114:
3111:
3109:
3106:
3104:
3101:
3099:
3096:
3094:
3091:
3089:
3086:
3085:
3083:
3081:
3076:
3072:
3066:
3063:
3061:
3058:
3056:
3053:
3051:
3048:
3046:
3045:Memory paging
3043:
3041:
3038:
3036:
3033:
3032:
3030:
3027:
3022:
3018:
3008:
3005:
3003:
3000:
2998:
2995:
2993:
2990:
2989:
2987:
2985:
2979:
2973:
2970:
2968:
2965:
2963:
2960:
2958:
2955:
2953:
2950:
2948:
2945:
2943:
2940:
2938:
2935:
2932:
2928:
2924:
2921:
2920:
2918:
2914:
2911:
2909:
2905:
2895:
2892:
2890:
2887:
2885:
2884:Device driver
2882:
2881:
2879:
2875:
2869:
2866:
2864:
2861:
2859:
2856:
2854:
2851:
2849:
2846:
2844:
2841:
2839:
2836:
2834:
2831:
2830:
2828:
2826:
2825:Architectures
2822:
2819:
2817:
2813:
2807:
2804:
2802:
2799:
2797:
2794:
2792:
2789:
2787:
2784:
2782:
2779:
2777:
2774:
2772:
2769:
2767:
2764:
2762:
2759:
2758:
2756:
2752:
2746:
2743:
2741:
2738:
2736:
2733:
2731:
2728:
2726:
2723:
2721:
2718:
2716:
2713:
2712:
2710:
2706:
2702:
2695:
2690:
2688:
2683:
2681:
2676:
2675:
2672:
2660:
2651:
2650:
2647:
2641:
2638:
2636:
2633:
2631:
2628:
2626:
2623:
2621:
2618:
2616:
2613:
2611:
2608:
2606:
2603:
2601:
2598:
2597:
2595:
2591:
2585:
2582:
2580:
2577:
2575:
2572:
2570:
2567:
2565:
2562:
2560:
2557:
2555:
2552:
2550:
2547:
2545:
2542:
2540:
2537:
2535:
2532:
2530:
2527:
2525:
2522:
2520:
2517:
2515:
2514:Global Arrays
2512:
2510:
2507:
2505:
2502:
2500:
2497:
2495:
2492:
2490:
2487:
2485:
2482:
2480:
2477:
2475:
2472:
2470:
2467:
2465:
2462:
2461:
2459:
2457:
2453:
2447:
2444:
2442:
2441:Grid computer
2439:
2435:
2432:
2431:
2430:
2427:
2424:
2421:
2417:
2414:
2412:
2409:
2407:
2404:
2402:
2399:
2397:
2394:
2392:
2389:
2388:
2387:
2384:
2380:
2377:
2375:
2372:
2371:
2370:
2367:
2365:
2362:
2360:
2357:
2355:
2352:
2350:
2347:
2343:
2340:
2338:
2335:
2331:
2328:
2326:
2323:
2320:
2317:
2316:
2315:
2312:
2310:
2307:
2306:
2305:
2302:
2301:
2299:
2297:
2293:
2287:
2284:
2280:
2277:
2275:
2272:
2270:
2267:
2266:
2265:
2262:
2260:
2257:
2255:
2252:
2251:
2249:
2247:
2243:
2237:
2234:
2232:
2229:
2227:
2224:
2222:
2219:
2217:
2214:
2212:
2209:
2207:
2204:
2203:
2201:
2197:
2191:
2188:
2186:
2183:
2181:
2178:
2176:
2173:
2171:
2168:
2167:
2165:
2161:
2155:
2152:
2150:
2147:
2145:
2142:
2140:
2137:
2135:
2132:
2130:
2127:
2125:
2122:
2120:
2117:
2115:
2112:
2111:
2109:
2105:
2099:
2096:
2093:
2090:
2088:
2085:
2083:
2080:
2077:
2074:
2072:
2069:
2066:
2063:
2061:
2058:
2057:
2055:
2053:
2049:
2043:
2040:
2038:
2035:
2033:
2030:
2028:
2025:
2023:
2020:
2018:
2015:
2013:
2010:
2008:
2005:
2004:
2002:
1998:
1992:
1989:
1987:
1984:
1982:
1979:
1977:
1974:
1972:
1969:
1967:
1964:
1962:
1959:
1957:
1954:
1952:
1949:
1947:
1944:
1943:
1941:
1937:
1933:
1926:
1921:
1919:
1914:
1912:
1907:
1906:
1903:
1896:
1895:0-13-101908-2
1892:
1888:
1884:
1882:
1881:0-201-44234-5
1878:
1874:
1870:
1868:
1867:0-672-31585-8
1864:
1860:
1856:
1854:
1853:1-56592-115-1
1850:
1846:
1842:
1840:
1839:0-201-63392-2
1836:
1832:
1828:
1827:
1822:
1813:
1806:
1803:
1798:
1794:
1793:
1788:
1781:
1778:
1773:
1766:
1764:
1760:
1755:
1751:
1750:
1745:
1738:
1735:
1730:
1724:
1720:
1716:
1715:
1707:
1704:
1699:
1693:
1689:
1688:
1680:
1677:
1669:
1662:
1656:
1653:
1648:
1646:9781118063330
1642:
1638:
1634:
1628:
1626:
1624:
1622:
1618:
1610:
1609:
1601:
1598:
1593:
1589:
1585:
1579:
1575:
1571:
1567:
1560:
1553:
1550:
1545:
1539:
1538:cite AV media
1524:
1523:
1515:
1512:
1507:
1501:
1498:
1493:
1489:
1488:
1483:
1479:
1473:
1470:
1462:
1461:
1457:(July 1966).
1456:
1450:
1447:
1442:
1440:0-13-595752-4
1436:
1432:
1428:
1422:
1419:
1414:
1410:
1406:
1402:
1398:
1394:
1387:
1383:
1377:
1374:
1367:
1362:
1359:
1357:
1356:Thread safety
1354:
1352:
1349:
1347:
1344:
1342:
1339:
1337:
1334:
1332:
1329:
1327:
1324:
1322:
1319:
1317:
1314:
1312:
1309:
1307:
1304:
1303:
1298:
1292:
1287:
1282:
1277:
1273:
1270:
1268:architecture.
1266:
1265:the same code
1262:
1259:designed for
1258:
1254:
1251:
1247:
1243:
1239:
1235:
1232:
1228:
1224:
1220:
1216:
1211:
1207:
1203:
1199:
1196:(and usually
1195:
1191:
1188:
1184:
1180:
1176:
1175:POSIX Threads
1172:
1168:
1164:
1161:
1157:
1153:
1152:
1151:
1145:
1140:
1137:
1134:
1131:
1127:
1126:Edward A. Lee
1123:
1119:
1115:
1111:
1107:
1103:
1099:
1095:
1094:
1090:
1089:
1088:
1082:
1078:
1074:
1070:
1066:
1063:
1060:
1056:
1051:
1050:worker thread
1047:
1044:
1043:
1042:
1036:
1034:
1031:
1025:
1017:
1015:
1013:
1009:
1005:
1001:
996:
994:
990:
986:
981:
977:
973:
970:(APIs) offer
969:
964:
961:
957:
953:
947:
946:Thread safety
942:
937:
935:
931:
929:
925:
919:
917:
913:
909:
905:
901:
897:
889:
887:
884:
883:DragonFly BSD
880:
876:
875:
870:
863:
858:
854:
850:
848:
844:
841:
837:
834:
830:
827:
824:
820:
816:
813:
812:
808:
806:
804:
800:
796:
792:
780:
776:
769:
765:
762:
760:
758:
757:State Threads
754:
749:
745:
744:multithreaded
741:
733:
730:
728:
726:
722:
718:
714:
710:
706:
702:
698:
697:GNU C Library
694:
690:
686:
678:
673:
671:
669:
668:
663:
659:
655:
650:
646:
642:
638:
633:
631:
627:
623:
619:
618:
613:
609:
605:
601:
597:
589:
587:
585:
581:
577:
573:
569:
565:
561:
557:
553:
549:
545:
544:cooperatively
541:
533:
528:
523:
520:
517:
513:
509:
506:
503:
500:
499:
498:
495:
494:(TLB) flush.
493:
489:
485:
484:address-space
481:
477:
473:
469:
461:
458:
455:
451:
448:
444:
442:
438:
434:
430:
427:
426:
425:
423:
419:
411:
409:
407:
403:
399:
395:
391:
387:
382:
374:
372:
371:primitives).
370:
364:
360:
356:
354:
353:green threads
350:
346:
342:
338:
334:
326:
324:
322:
318:
314:
310:
306:
301:
300:kernel thread
293:
291:
289:
285:
281:
277:
273:
269:
265:
261:
257:
252:
250:
246:
245:
240:
236:
232:
228:
222:
214:
212:
210:
209:cooperatively
206:
198:
196:
192:
190:
186:
175:
172:February 2021
166:
162:
159:This section
157:
154:
150:
149:
143:
141:
139:
134:
132:
129:
125:
121:
117:
112:
110:
106:
102:
98:
94:
90:
83:
79:
75:
70:
66:
62:
56:
51:
42:
38:
37:Threaded code
34:
30:
19:
3080:file systems
2972:Time-sharing
2966:
2199:Coordination
2174:
2129:Amdahl's law
2065:Simultaneous
1886:
1872:
1858:
1844:
1830:
1805:
1790:
1780:
1747:
1737:
1713:
1706:
1686:
1679:
1668:the original
1655:
1636:
1607:
1600:
1565:
1552:
1527:. Retrieved
1521:
1514:
1500:
1491:
1485:
1478:Sutter, Herb
1472:
1459:
1449:
1430:
1421:
1396:
1392:
1376:
1341:Protothreads
1194:higher level
1160:multitasking
1159:
1149:
1138:
1132:
1091:
1086:
1064:
1059:Unix signals
1049:
1045:
1040:
1030:thread pools
1027:
1018:Thread pools
997:
965:
949:
941:
932:
920:
911:
899:
893:
872:
867:
822:
818:
798:
794:
790:
778:
774:
773:
767:
763:
739:
737:
731:
705:LinuxThreads
682:
665:
644:
637:time slicing
634:
615:
593:
540:preemptively
537:
521:
507:
501:
496:
479:
478:threads and
475:
465:
418:multitasking
415:
398:user threads
384:
365:
361:
357:
337:user threads
336:
330:
327:User threads
299:
297:
271:
264:file handles
255:
253:
242:
238:
230:
226:
224:
205:preemptively
202:
193:
182:
169:
165:adding to it
160:
135:
128:thread-local
116:concurrently
113:
92:
86:
3098:Device file
3088:Boot loader
3002:Round-robin
2927:Cooperative
2863:Rump kernel
2853:Multikernel
2843:Microkernel
2740:Usage share
2635:Scalability
2396:distributed
2279:Concurrency
2246:Programming
2087:Cooperative
2076:Speculative
2012:Instruction
1857:Paul Hyde:
1187:beginthread
993:granularity
918:community.
632:processor.
602:. In 2002,
560:lock convoy
71:vs. Thread
35:calls, see
3201:Categories
3028:protection
2984:algorithms
2982:Scheduling
2931:Preemptive
2877:Components
2848:Monolithic
2715:Comparison
2640:Starvation
2379:asymmetric
2114:PRAM model
2082:Preemptive
1529:2020-11-24
1368:References
1240:for Ruby,
1106:rendezvous
1098:programmer
1008:semaphores
881:2.x+, and
847:Cray MTA-2
781:maps some
658:multi-core
529:Scheduling
468:Windows NT
456:mechanisms
439:and other
406:coroutines
343:machines (
78:Preemption
74:Scheduling
33:subroutine
3118:Partition
3035:Bus error
2962:Real-time
2942:Interrupt
2868:Unikernel
2833:Exokernel
2374:symmetric
2119:PEM model
1592:251692327
1183:process.h
1118:deadlocks
908:semantics
877:or LWPs.
703:or older
582:or if it
480:expensive
441:resources
422:processes
345:M:N model
333:userspace
313:registers
305:scheduler
260:resources
215:Processes
138:processes
101:scheduler
97:execution
55:processor
3164:Live USB
3026:resource
2916:Concepts
2754:Variants
2735:Timeline
2605:Deadlock
2593:Problems
2559:pthreads
2539:OpenHMPP
2464:Ateji PX
2425:computer
2296:Hardware
2163:Elements
2149:Slowdown
2060:Temporal
2042:Pipeline
1861:, Sams,
1792:Overload
1756:: 16–19.
1749:Overload
1429:(1992).
1283:See also
1274:such as
1238:Ruby MRI
1227:Ateji PX
1012:monitors
985:spinlock
974:such as
930:system.
842:project.
662:parallel
580:resource
272:isolated
3159:Live CD
3113:Journal
3077:access,
3075:Storage
2952:Process
2858:vkernel
2725:History
2708:General
2564:RaftLib
2544:OpenACC
2519:GPUOpen
2509:C++ AMP
2484:Charm++
2226:Barrier
2170:Process
2154:Speedup
1939:General
1795:(128).
1413:5679366
1276:Verilog
1242:CPython
1114:mutexes
1057:and/or
976:mutexes
904:command
857:Haskell
833:Solaris
717:FreeBSD
709:Solaris
647:. This
610:to the
584:starves
256:process
227:process
144:History
109:process
69:Process
65:Program
2967:Thread
2838:Hybrid
2816:Kernel
2657:
2534:OpenCL
2529:OpenMP
2474:Chapel
2391:shared
2386:Memory
2321:(SIMT)
2264:Models
2175:Thread
2107:Theory
2078:(SpMT)
2032:Memory
2017:Thread
2000:Levels
1893:
1879:
1865:
1851:
1837:
1799:: 4–7.
1752:(97).
1725:
1694:
1643:
1590:
1580:
1437:
1411:
1219:OpenMP
1208:, and
1206:Python
1073:OpenCL
1010:, and
952:couple
879:NetBSD
723:, and
713:NetBSD
639:: the
576:blocks
437:memory
402:OpenMP
386:Fibers
375:Fibers
319:, and
244:fibers
120:memory
93:thread
3169:Shell
3108:Inode
2504:Dryad
2469:Boost
2190:Array
2180:Fiber
2094:(CMT)
2067:(SMT)
1981:GPGPU
1671:(PDF)
1664:(PDF)
1612:(PDF)
1588:S2CID
1562:(PDF)
1464:(PDF)
1409:S2CID
1389:(PDF)
1192:Some
1122:races
1079:on a
926:on a
869:SunOS
721:macOS
693:Linux
689:Win32
604:Intel
476:cheap
433:state
394:yield
369:await
309:stack
2730:List
2569:ROCm
2499:CUDA
2489:Cilk
2456:APIs
2416:COMA
2411:NUMA
2342:MIMD
2337:MISD
2314:SIMD
2309:SISD
2037:Loop
2027:Data
2022:Task
1891:ISBN
1877:ISBN
1863:ISBN
1849:ISBN
1835:ISBN
1797:ACCU
1754:ACCU
1723:ISBN
1692:ISBN
1641:ISBN
1578:ISBN
1544:link
1494:(3).
1435:ISBN
1397:C-28
1257:CUDA
1231:CUDA
1215:Cilk
1202:Java
1179:Unix
1169:and
1156:PL/I
1154:IBM
1071:and
1069:CUDA
980:lock
851:The
701:NPTL
695:the
687:and
685:OS/2
472:OS/2
470:and
91:, a
67:vs.
3186:PXE
3174:CLI
3154:HAL
3144:API
2947:IPC
2584:ZPL
2579:TBB
2574:UPC
2554:PVM
2524:MPI
2479:HPX
2406:UMA
2007:Bit
1570:doi
1401:doi
1250:Tcl
1171:C++
978:to
956:IPC
894:In
840:PM2
738:An
725:iOS
656:or
626:AMD
542:or
488:x86
274:by
207:or
167:.
111:.
95:of
87:In
3203::
2929:,
1789:.
1762:^
1746:.
1620:^
1586:.
1576:.
1564:.
1540:}}
1536:{{
1492:30
1490:.
1484:.
1407:.
1395:.
1391:.
1233:).
1229:,
1221:,
1217:,
1204:,
1014:.
1006:,
1002:,
898:,
759:.
727:.
719:,
715:,
711:,
562:,
546:.
355:.
298:A
254:A
80:,
76:,
3023:,
2933:)
2925:(
2693:e
2686:t
2679:v
1924:e
1917:t
1910:v
1814:.
1731:.
1700:.
1649:.
1594:.
1572::
1546:)
1532:.
1508:.
1443:.
1415:.
1403::
1213:(
1189:.
1167:C
823:N
821::
819:M
799:N
797::
795:M
791:N
787:N
783:M
779:N
777::
775:M
768:N
766::
764:M
740:M
732:M
174:)
170:(
43:.
20:)
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.
