31:
265:
273:
316:
357:. These coprocessors had a different instruction set from the Intel coprocessors, and used a different socket, which not all motherboards supported. The Weitek processors did not provide transcendental mathematics functions (for example, trigonometric functions) like the Intel x87 family, and required specific software libraries to support their functions.
157:. The mainframe would not require any I/O processing at all, instead would just set parameters for an input or output operation and then signal the channel processor to carry out the whole of the operation. By dedicating relatively simple sub-processors to handle time-consuming I/O formatting and processing, overall system performance was improved.
346:. The Intel 80486DX processor included floating-point hardware on the chip. Intel released a cost-reduced processor, the 80486SX, that had no floating-point hardware, and also sold an 80487SX coprocessor that essentially disabled the main processor when installed, since the 80487SX was a complete 80486DX with a different set of pin connections.
237:
As microprocessors developed, the cost of integrating the floating-point arithmetic functions into the processor declined. High processor speeds also made a closely integrated coprocessor difficult to implement. Separately packaged mathematics coprocessors are now uncommon in desktop computers. The
598:
computers. Rather than special-purpose graphics or arithmetic devices, these were general-purpose CPUs (such as 8086, Zilog Z80, or 6502) to which particular types of task were assigned by the operating system, off-loading them from the computer's main CPU and resulting in acceleration. In addition,
428:
and GPU do not have to perform these time-consuming calculations. It was designed for video games, although other mathematical uses could theoretically be developed for it. In 2008, Nvidia purchased the company and phased out the PhysX card line; the functionality was added through software allowing
327:
input/output coprocessor. It used the same programming technique as 8087 for input/output operations, such as transfer of data from memory to a peripheral device, and so reducing the load on the CPU. But IBM did not use it in IBM PC design and Intel stopped development of this type of coprocessor.
373:
coprocessors which provided similar floating-point speed acceleration as for the Intel processors. Computers using the 68000 family but not equipped with the hardware floating-point processor could trap and emulate the floating-point instructions in software, which, although slower, allowed one
349:
Intel processors later than the 80486 integrated floating-point hardware on the main processor chip; the advances in integration eliminated the cost advantage of selling the floating-point processor as an optional element. It would be very difficult to adapt circuit-board techniques adequate at
299:
or mathematics-intensive calculations. In that architecture, the coprocessor speeds up floating-point arithmetic on the order of fiftyfold. Users that only used the PC for word processing, for example, saved the high cost of the coprocessor, which would not have accelerated performance of text
311:
operation codes inserted in the 8088 instruction stream. An 8088 processor without an 8087 could not interpret these instructions, requiring separate versions of programs for FPU and non-FPU systems, or at least a test at run time to detect the FPU and select appropriate mathematical library
350:
75 MHz processor speed to meet the time-delay, power consumption, and radio-frequency interference standards required at gigahertz-range clock speeds. These on-chip floating-point processors are still referred to as coprocessors because they operate in parallel with the main CPU.
736:
620:
have become integrated on CPU dies. Nonetheless, specialized units remain popular away from desktop machines, and for additional power, and allow continued evolution independently of the main processor product lines.
740:
78:, coprocessors can accelerate system performance. Coprocessors allow a line of computers to be customized, so that customers who do not need the extra performance do not need to pay for it.
790:
168:
arithmetic operations. Where a coprocessor was supported, floating-point calculations could be carried out many times faster. Math coprocessors were popular purchases for users of
544:
supports up to four coprocessor units, used for memory management, floating-point arithmetic, and two undefined coprocessors for other tasks such as graphics accelerators.
608:
Over time CPUs have tended to grow to absorb the functionality of the most popular coprocessors. FPUs are now considered an integral part of a processors' main pipeline;
599:
a BBC Micro fitted with a coprocessor was able to run machine code software designed for other systems, such as CP/M and DOS which are written for 8086 processors.
1858:
551:(field-programmable gate arrays), custom coprocessors can be created for acceleration of particular processing tasks such as digital signal processing (e.g.
374:
binary version of the program to be distributed for both cases. The 68451 memory-management coprocessor was designed to work with the 68020 processor.
830:
193:
782:
1969:
1152:
583:
128 core PCI-e coprocessor is a proprietary accelerator that requires a CPU to run it, and has been employed in an upgrade of the 17,792 node
1671:
580:
164:
in the 1970s and became common throughout the 1980s and into the early 1990s. Early 8-bit and 16-bit processors used software to carry out
949:
1828:
1394:
1211:
394:
have been fitted with dedicated processors providing digital multichannel mixing and real-time DSP effects as early as 1990 to 1994 (the
432:
In 2006, BigFoot
Systems unveiled a PCI add-in card they christened the KillerNIC which ran its own special Linux kernel on a FreeScale
2182:
576:
chips can be programmed so that one of their processors is the primary processor, and the other processors are supporting coprocessors.
1174:
587:
supercomputer (2 Intel
Knights Bridge+ 2 Matrix 2000 each), now dubbed 2A, roughly doubling its speed at 95 petaflops, exceeding the
91:
1823:
1895:
466:
with the aim to support general-purpose CPUs, ATI/AMD and Nvidia GPUs (and other accelerators) with a single common language for
2177:
1648:
102:. Others are independent processors in their own right, capable of working asynchronously; they are still not optimized for
2592:
1716:
979:
823:
2602:
1743:
870:
653:
1910:
1738:
1711:
1090:
769:
723:
702:
681:
609:
429:
their GPUs to render PhysX on cores normally used for graphics processing, using their Nvidia PhysX engine software.
135:
2771:
2761:
2725:
2288:
1181:
1147:
1142:
1061:
1026:
2776:
2700:
2597:
1998:
1905:
1706:
927:
816:
99:
75:
1726:
1445:
880:
477:
as a coprocessor. Examples of coprocessors used for handling sensor integration in mobile devices include the
1900:
1748:
1721:
1582:
1196:
1157:
1014:
548:
498:
2337:
2099:
1575:
1536:
1191:
1186:
1120:
932:
531:), and as of 2018, such AI chips are in smartphones such as from Apple, and several Android phone vendors.
353:
During the era of 8- and 16-bit desktop computers another common source of floating-point coprocessors was
17:
1964:
1661:
1359:
1056:
569:; such accelerators used to be cards, but in modern times are instructions for crypto in mainstream CPUs.
2614:
2261:
1678:
1169:
1137:
907:
895:
875:
648:
516:
223:
2705:
2668:
2658:
1046:
486:
482:
288:
67:
51:
172:(CAD) software and scientific and engineering calculations. Some floating-point units, such as the
2720:
2127:
2063:
2040:
1890:
1852:
1688:
1638:
1633:
1110:
1004:
912:
617:
613:
383:
239:
212:
131:
494:
452:
2673:
2456:
2350:
2314:
2231:
2215:
2057:
1846:
1805:
1793:
1656:
1570:
1491:
1256:
917:
860:
437:
425:
421:
95:
47:
2479:
2451:
2361:
2326:
2075:
2069:
2051:
1785:
1779:
1683:
1587:
1478:
1417:
1279:
922:
573:
520:
219:
2766:
2653:
2562:
2308:
2020:
1838:
1597:
1565:
1523:
1435:
1236:
1051:
1041:
1031:
1021:
991:
974:
839:
448:
296:
200:
169:
111:
2683:
2619:
2205:
1927:
1817:
1764:
1296:
1009:
865:
847:
403:
115:
103:
8:
2730:
2715:
2535:
2386:
2368:
2332:
2320:
1974:
1921:
1698:
1614:
1496:
1351:
1246:
1105:
399:
292:
87:
30:
2587:
2579:
2431:
2406:
2210:
2085:
1609:
1550:
1430:
1162:
890:
566:
407:
150:
43:
2540:
2507:
2423:
2355:
2256:
2246:
2236:
2167:
2162:
2157:
2080:
2009:
1915:
1875:
1508:
1458:
1408:
1384:
1266:
1206:
1201:
1083:
999:
765:
719:
698:
677:
588:
541:
528:
395:
343:
71:
63:
59:
55:
2710:
2643:
2629:
2484:
2391:
2345:
2152:
2147:
2142:
2137:
2132:
2122:
1992:
1959:
1870:
1865:
1774:
1626:
1621:
1604:
1592:
1531:
1095:
1073:
959:
937:
855:
562:
556:
490:
161:
138:
737:"Intel Delivers New Architecture for Discovery with Intel® Xeon Phi™ Coprocessors"
2624:
2609:
2557:
2461:
2436:
2273:
2266:
2117:
2112:
2107:
2046:
1954:
1944:
1666:
1501:
1453:
1216:
1100:
1068:
969:
964:
885:
630:
177:
173:
107:
783:"China's Tianhe-2A will Use Proprietary Accelerator and Boast 95 Petaflops Peak"
515:
As of 2016, various companies are developing coprocessors aimed at accelerating
2735:
2569:
2552:
2545:
2441:
2298:
2035:
1949:
1880:
1463:
1425:
1374:
1369:
1364:
1078:
902:
509:
467:
370:
366:
335:
189:
165:
127:
2755:
2530:
2446:
1486:
1468:
1261:
954:
459:
387:
153:
processor time, input/output tasks were delegated to separate systems called
123:
1389:
242:
has grown, however, particularly due to the increasing demand for realistic
199:
Another form of coprocessor was a video display coprocessor, as used in the
2740:
2678:
2494:
2471:
2283:
2004:
942:
308:
119:
645:
framework for writing programs that execute across heterogeneous platforms
424:. PhysX was designed to perform complex physics computations so that the
74:
with peripheral devices. By offloading processor-intensive tasks from the
2525:
2489:
2200:
2172:
808:
444:
332:
243:
154:
612:
units gave multimedia its acceleration, taking over the role of various
2411:
2401:
2396:
2378:
2278:
2251:
1513:
1346:
1316:
1036:
474:
433:
391:
324:
304:
285:
247:
185:
2502:
2499:
2241:
1311:
1289:
595:
524:
2517:
1336:
639:, an initiative to implement coprocessor support for AMD processors
636:
584:
478:
204:
264:
184:
FPUs were treated as peripheral devices, while others such as the
1326:
1284:
272:
227:
447:
is a media-oriented add-in card with a coprocessor based on the
323:
Another coprocessor for the 8086/8088 central processor was the
1341:
1306:
1271:
716:
Assembly language and systems programming for the M68000 family
642:
463:
354:
315:
281:
231:
181:
46:
used to supplement the functions of the primary processor (the
473:
In 2010s, some mobile computation devices had implemented the
1799:
1331:
1301:
633:, the use of two or more CPUs within a single computer system
505:
418:
414:
339:
216:
160:
Coprocessors for floating-point arithmetic first appeared in
86:
Coprocessors vary in their degree of autonomy. Some (such as
2663:
1811:
1731:
1321:
552:
417:
announced an add-in card for computers that it called the
1251:
1241:
295:) which was a popular option for people using the PC for
259:
208:
436:
running at 400 MHz, calling the FreeScale chip a
342:) to perform floating-point operations directly in
118:(DMA), with the host processor (a CPU) building a
50:). Operations performed by the coprocessor may be
594:A range of coprocessors were available for Acorn
382:As of 2001, dedicated Graphics Processing Units (
2753:
739:. Newsroom.intel.com. 2012-11-12. Archived from
319:Intel 80386DX CPU with 80387DX math coprocessor
106:, or they are incapable of it due to a limited
824:
1829:Computer performance by orders of magnitude
695:Upgrading and Repairing PCs, Second Edition
519:for vision and other cognitive tasks (e.g.
196:were more closely integrated with the CPU.
838:
831:
817:
338:used an optional "math" coprocessor (the
303:The 8087 was tightly integrated with the
114:. It is common for these to be driven by
687:
666:
314:
271:
263:
29:
14:
2754:
377:
307:/8088 and responded to floating-point
812:
708:
534:
1800:Floating-point operations per second
793:from the original on 1 December 2020
718:Jones & Bartlett Learning, 1992
455:are themselves vector coprocessors.
141:capable of both modes of operation.
676:15th edition, Que Publishing, 2003
402:being typical examples), while the
390:are commonplace. Certain models of
211:home computers, which were called "
24:
268:i8087 and i80287 microarchitecture
222:includes such a unit known as the
25:
2788:
559:cores with FPGA on a single die).
2726:Semiconductor device fabrication
81:
2701:History of general-purpose CPUs
928:Nondeterministic Turing machine
487:Qualcomm Snapdragon Sensor Core
34:AM9511-1 arithmetic coprocessor
881:Deterministic finite automaton
775:
762:The MIPS programmer's handbook
754:
729:
714:William Ford, William R. Topp
240:dedicated graphics coprocessor
13:
1:
1672:Simultaneous and heterogenous
760:Erin Farquhar, Philip Bunce,
659:
589:world's fastest supercomputer
90:) rely on direct control via
2356:Integrated memory controller
2338:Translation lookaside buffer
1537:Memory dependence prediction
980:Random-access stored program
933:Probabilistic Turing machine
674:Upgrading and repairing PCs
616:accelerator cards; and even
7:
1812:Synaptic updates per second
624:
495:Holographic Processing Unit
360:
112:accelerating specific tasks
10:
2793:
2216:Heterogeneous architecture
1138:Orthogonal instruction set
908:Alternating Turing machine
896:Quantum cellular automaton
649:Asymmetric multiprocessing
517:artificial neural networks
410:are more recent examples.
284:included a socket for the
257:
144:
27:Type of computer processor
2706:Microprocessor chronology
2693:
2669:Dynamic frequency scaling
2642:
2578:
2516:
2470:
2422:
2377:
2297:
2224:
2193:
2098:
2019:
1983:
1937:
1837:
1824:Cache performance metrics
1763:
1697:
1647:
1558:
1549:
1522:
1477:
1444:
1416:
1407:
1227:
1130:
1119:
990:
846:
603:
300:manipulation operations.
213:Video Display Controllers
52:floating-point arithmetic
2721:Hardware security module
2064:Digital signal processor
2041:Graphics processing unit
1853:Graphics processing unit
764:, Morgan Kaufmann, 1994
276:i80387 microarchitecture
253:
234:manipulation in memory.
149:To make the best use of
92:coprocessor instructions
2772:Heterogeneous computing
2762:Central processing unit
2674:Dynamic voltage scaling
2457:Memory address register
2351:Branch target predictor
2315:Address generation unit
2058:Physics processing unit
1847:Central processing unit
1806:Transactions per second
1794:Instructions per second
1717:Array processing (SIMT)
861:Stored-program computer
697:, Que Publishing, 1992
521:vision processing units
451:microarchitecture. The
438:Network Processing Unit
2777:OpenCL compute devices
2480:Hardwired control unit
2362:Memory management unit
2327:Memory management unit
2076:Secure cryptoprocessor
2070:Tensor Processing Unit
2052:Vision processing unit
1786:Cycles per instruction
1780:Instructions per cycle
1727:Associative processing
1418:Instruction pipelining
840:Processor technologies
772:, appendix A3 page 330
320:
277:
269:
35:
2563:Sum-addressed decoder
2309:Arithmetic logic unit
1436:Classic RISC pipeline
1390:Epiphany architecture
1237:Motorola 68000 series
789:. 25 September 2017.
318:
297:computer-aided design
275:
267:
201:Atari 8-bit computers
170:computer-aided design
130:contained an unusual
33:
2684:Performance per watt
2262:replacement policies
1928:Package on a package
1818:Performance per watt
1722:Pipelined processing
1492:Tomasulo's algorithm
1297:Clipper architecture
1153:Application-specific
866:Finite-state machine
563:TLS/SSL accelerators
404:Sound Blaster Audigy
116:direct memory access
104:general-purpose code
2716:Digital electronics
2369:Instruction decoder
2321:Floating-point unit
1975:Soft microprocessor
1922:System in a package
1497:Reservation station
1027:Transport-triggered
483:motion coprocessors
400:Sound Blaster AWE32
378:Modern coprocessors
2588:Integrated circuit
2432:Processor register
2086:Baseband processor
1431:Operand forwarding
891:Cellular automaton
535:Other coprocessors
499:Microsoft HoloLens
408:Sound Blaster X-Fi
321:
278:
270:
151:mainframe computer
100:instruction stream
94:, embedded in the
44:computer processor
36:
2749:
2748:
2638:
2637:
2257:Instruction cache
2247:Scratchpad memory
2094:
2093:
2081:Network processor
2010:Network on a chip
1965:Ultra-low-voltage
1916:Multi-chip module
1759:
1758:
1545:
1544:
1532:Branch prediction
1509:Register renaming
1403:
1402:
1385:VISC architecture
1207:Quantum computing
1202:VISC architecture
1084:Secondary storage
1000:Microarchitecture
960:Register machines
542:MIPS architecture
396:Gravis Ultrasound
386:) in the form of
291:coprocessor (aka
230:for accelerating
162:desktop computers
64:string processing
60:signal processing
16:(Redirected from
2784:
2711:Processor design
2603:Power management
2485:Instruction unit
2346:Branch predictor
2295:
2294:
1993:System on a chip
1935:
1934:
1775:Transistor count
1699:Flynn's taxonomy
1556:
1555:
1414:
1413:
1217:Addressing modes
1128:
1127:
1074:Memory hierarchy
938:Hypercomputation
856:Abstract machine
833:
826:
819:
810:
809:
803:
802:
800:
798:
779:
773:
758:
752:
751:
749:
748:
733:
727:
726:page 892 and ff.
712:
706:
691:
685:
670:
491:Qualcomm Hexagon
21:
2792:
2791:
2787:
2786:
2785:
2783:
2782:
2781:
2752:
2751:
2750:
2745:
2731:Tick–tock model
2689:
2645:
2634:
2574:
2558:Address decoder
2512:
2466:
2462:Program counter
2437:Status register
2418:
2373:
2333:Load–store unit
2300:
2293:
2220:
2189:
2090:
2047:Image processor
2022:
2015:
1985:
1979:
1955:Microcontroller
1945:Embedded system
1933:
1833:
1766:
1755:
1693:
1643:
1541:
1518:
1502:Re-order buffer
1473:
1454:Data dependency
1440:
1399:
1229:
1223:
1122:
1121:Instruction set
1115:
1101:Multiprocessing
1069:Cache hierarchy
1062:Register/memory
986:
886:Queue automaton
842:
837:
807:
806:
796:
794:
781:
780:
776:
759:
755:
746:
744:
735:
734:
730:
713:
709:
693:Scott Mueller,
692:
688:
684:, pages 108–110
672:Scott Mueller,
671:
667:
662:
631:Multiprocessing
627:
606:
537:
468:compute kernels
380:
369:family had the
363:
262:
256:
226:, as well as a
178:Intel 8231/8232
147:
108:instruction set
84:
72:I/O interfacing
28:
23:
22:
15:
12:
11:
5:
2790:
2780:
2779:
2774:
2769:
2764:
2747:
2746:
2744:
2743:
2738:
2736:Pin grid array
2733:
2728:
2723:
2718:
2713:
2708:
2703:
2697:
2695:
2691:
2690:
2688:
2687:
2681:
2676:
2671:
2666:
2661:
2656:
2650:
2648:
2640:
2639:
2636:
2635:
2633:
2632:
2627:
2622:
2617:
2612:
2607:
2606:
2605:
2600:
2595:
2584:
2582:
2576:
2575:
2573:
2572:
2570:Barrel shifter
2567:
2566:
2565:
2560:
2553:Binary decoder
2550:
2549:
2548:
2538:
2533:
2528:
2522:
2520:
2514:
2513:
2511:
2510:
2505:
2497:
2492:
2487:
2482:
2476:
2474:
2468:
2467:
2465:
2464:
2459:
2454:
2449:
2444:
2442:Stack register
2439:
2434:
2428:
2426:
2420:
2419:
2417:
2416:
2415:
2414:
2409:
2399:
2394:
2389:
2383:
2381:
2375:
2374:
2372:
2371:
2366:
2365:
2364:
2353:
2348:
2343:
2342:
2341:
2335:
2324:
2318:
2312:
2305:
2303:
2292:
2291:
2286:
2281:
2276:
2271:
2270:
2269:
2264:
2259:
2254:
2249:
2244:
2234:
2228:
2226:
2222:
2221:
2219:
2218:
2213:
2208:
2203:
2197:
2195:
2191:
2190:
2188:
2187:
2186:
2185:
2175:
2170:
2165:
2160:
2155:
2150:
2145:
2140:
2135:
2130:
2125:
2120:
2115:
2110:
2104:
2102:
2096:
2095:
2092:
2091:
2089:
2088:
2083:
2078:
2073:
2067:
2061:
2055:
2049:
2044:
2038:
2036:AI accelerator
2033:
2027:
2025:
2017:
2016:
2014:
2013:
2007:
2002:
1999:Multiprocessor
1996:
1989:
1987:
1981:
1980:
1978:
1977:
1972:
1967:
1962:
1957:
1952:
1950:Microprocessor
1947:
1941:
1939:
1938:By application
1932:
1931:
1925:
1919:
1913:
1908:
1903:
1898:
1893:
1888:
1883:
1881:Tile processor
1878:
1873:
1868:
1863:
1862:
1861:
1850:
1843:
1841:
1835:
1834:
1832:
1831:
1826:
1821:
1815:
1809:
1803:
1797:
1791:
1790:
1789:
1777:
1771:
1769:
1761:
1760:
1757:
1756:
1754:
1753:
1752:
1751:
1741:
1736:
1735:
1734:
1729:
1724:
1719:
1709:
1703:
1701:
1695:
1694:
1692:
1691:
1686:
1681:
1676:
1675:
1674:
1669:
1667:Hyperthreading
1659:
1653:
1651:
1649:Multithreading
1645:
1644:
1642:
1641:
1636:
1631:
1630:
1629:
1619:
1618:
1617:
1612:
1602:
1601:
1600:
1595:
1585:
1580:
1579:
1578:
1573:
1562:
1560:
1553:
1547:
1546:
1543:
1542:
1540:
1539:
1534:
1528:
1526:
1520:
1519:
1517:
1516:
1511:
1506:
1505:
1504:
1499:
1489:
1483:
1481:
1475:
1474:
1472:
1471:
1466:
1461:
1456:
1450:
1448:
1442:
1441:
1439:
1438:
1433:
1428:
1426:Pipeline stall
1422:
1420:
1411:
1405:
1404:
1401:
1400:
1398:
1397:
1392:
1387:
1382:
1379:
1378:
1377:
1375:z/Architecture
1372:
1367:
1362:
1354:
1349:
1344:
1339:
1334:
1329:
1324:
1319:
1314:
1309:
1304:
1299:
1294:
1293:
1292:
1287:
1282:
1274:
1269:
1264:
1259:
1254:
1249:
1244:
1239:
1233:
1231:
1225:
1224:
1222:
1221:
1220:
1219:
1209:
1204:
1199:
1194:
1189:
1184:
1179:
1178:
1177:
1167:
1166:
1165:
1155:
1150:
1145:
1140:
1134:
1132:
1125:
1117:
1116:
1114:
1113:
1108:
1103:
1098:
1093:
1088:
1087:
1086:
1081:
1079:Virtual memory
1071:
1066:
1065:
1064:
1059:
1054:
1049:
1039:
1034:
1029:
1024:
1019:
1018:
1017:
1007:
1002:
996:
994:
988:
987:
985:
984:
983:
982:
977:
972:
967:
957:
952:
947:
946:
945:
940:
935:
930:
925:
920:
915:
910:
903:Turing machine
900:
899:
898:
893:
888:
883:
878:
873:
863:
858:
852:
850:
844:
843:
836:
835:
828:
821:
813:
805:
804:
774:
753:
728:
707:
686:
664:
663:
661:
658:
657:
656:
654:AI accelerator
651:
646:
640:
634:
626:
623:
605:
602:
601:
600:
592:
577:
570:
560:
545:
536:
533:
510:Intel Xeon Phi
508:announced the
388:graphics cards
379:
376:
367:Motorola 68000
362:
359:
336:microprocessor
289:floating-point
258:Main article:
255:
252:
248:computer games
220:custom chipset
194:National 32081
190:Motorola 68881
166:floating-point
146:
143:
128:Emotion Engine
83:
80:
76:main processor
26:
9:
6:
4:
3:
2:
2789:
2778:
2775:
2773:
2770:
2768:
2765:
2763:
2760:
2759:
2757:
2742:
2739:
2737:
2734:
2732:
2729:
2727:
2724:
2722:
2719:
2717:
2714:
2712:
2709:
2707:
2704:
2702:
2699:
2698:
2696:
2692:
2685:
2682:
2680:
2677:
2675:
2672:
2670:
2667:
2665:
2662:
2660:
2657:
2655:
2652:
2651:
2649:
2647:
2641:
2631:
2628:
2626:
2623:
2621:
2618:
2616:
2613:
2611:
2608:
2604:
2601:
2599:
2596:
2594:
2591:
2590:
2589:
2586:
2585:
2583:
2581:
2577:
2571:
2568:
2564:
2561:
2559:
2556:
2555:
2554:
2551:
2547:
2544:
2543:
2542:
2539:
2537:
2534:
2532:
2531:Demultiplexer
2529:
2527:
2524:
2523:
2521:
2519:
2515:
2509:
2506:
2504:
2501:
2498:
2496:
2493:
2491:
2488:
2486:
2483:
2481:
2478:
2477:
2475:
2473:
2469:
2463:
2460:
2458:
2455:
2453:
2452:Memory buffer
2450:
2448:
2447:Register file
2445:
2443:
2440:
2438:
2435:
2433:
2430:
2429:
2427:
2425:
2421:
2413:
2410:
2408:
2405:
2404:
2403:
2400:
2398:
2395:
2393:
2390:
2388:
2387:Combinational
2385:
2384:
2382:
2380:
2376:
2370:
2367:
2363:
2360:
2359:
2357:
2354:
2352:
2349:
2347:
2344:
2339:
2336:
2334:
2331:
2330:
2328:
2325:
2322:
2319:
2316:
2313:
2310:
2307:
2306:
2304:
2302:
2296:
2290:
2287:
2285:
2282:
2280:
2277:
2275:
2272:
2268:
2265:
2263:
2260:
2258:
2255:
2253:
2250:
2248:
2245:
2243:
2240:
2239:
2238:
2235:
2233:
2230:
2229:
2227:
2223:
2217:
2214:
2212:
2209:
2207:
2204:
2202:
2199:
2198:
2196:
2192:
2184:
2181:
2180:
2179:
2176:
2174:
2171:
2169:
2166:
2164:
2161:
2159:
2156:
2154:
2151:
2149:
2146:
2144:
2141:
2139:
2136:
2134:
2131:
2129:
2126:
2124:
2121:
2119:
2116:
2114:
2111:
2109:
2106:
2105:
2103:
2101:
2097:
2087:
2084:
2082:
2079:
2077:
2074:
2071:
2068:
2065:
2062:
2059:
2056:
2053:
2050:
2048:
2045:
2042:
2039:
2037:
2034:
2032:
2029:
2028:
2026:
2024:
2018:
2011:
2008:
2006:
2003:
2000:
1997:
1994:
1991:
1990:
1988:
1982:
1976:
1973:
1971:
1968:
1966:
1963:
1961:
1958:
1956:
1953:
1951:
1948:
1946:
1943:
1942:
1940:
1936:
1929:
1926:
1923:
1920:
1917:
1914:
1912:
1909:
1907:
1904:
1902:
1899:
1897:
1894:
1892:
1889:
1887:
1884:
1882:
1879:
1877:
1874:
1872:
1869:
1867:
1864:
1860:
1857:
1856:
1854:
1851:
1848:
1845:
1844:
1842:
1840:
1836:
1830:
1827:
1825:
1822:
1819:
1816:
1813:
1810:
1807:
1804:
1801:
1798:
1795:
1792:
1787:
1784:
1783:
1781:
1778:
1776:
1773:
1772:
1770:
1768:
1762:
1750:
1747:
1746:
1745:
1742:
1740:
1737:
1733:
1730:
1728:
1725:
1723:
1720:
1718:
1715:
1714:
1713:
1710:
1708:
1705:
1704:
1702:
1700:
1696:
1690:
1687:
1685:
1682:
1680:
1677:
1673:
1670:
1668:
1665:
1664:
1663:
1660:
1658:
1655:
1654:
1652:
1650:
1646:
1640:
1637:
1635:
1632:
1628:
1625:
1624:
1623:
1620:
1616:
1613:
1611:
1608:
1607:
1606:
1603:
1599:
1596:
1594:
1591:
1590:
1589:
1586:
1584:
1581:
1577:
1574:
1572:
1569:
1568:
1567:
1564:
1563:
1561:
1557:
1554:
1552:
1548:
1538:
1535:
1533:
1530:
1529:
1527:
1525:
1521:
1515:
1512:
1510:
1507:
1503:
1500:
1498:
1495:
1494:
1493:
1490:
1488:
1487:Scoreboarding
1485:
1484:
1482:
1480:
1476:
1470:
1469:False sharing
1467:
1465:
1462:
1460:
1457:
1455:
1452:
1451:
1449:
1447:
1443:
1437:
1434:
1432:
1429:
1427:
1424:
1423:
1421:
1419:
1415:
1412:
1410:
1406:
1396:
1393:
1391:
1388:
1386:
1383:
1380:
1376:
1373:
1371:
1368:
1366:
1363:
1361:
1358:
1357:
1355:
1353:
1350:
1348:
1345:
1343:
1340:
1338:
1335:
1333:
1330:
1328:
1325:
1323:
1320:
1318:
1315:
1313:
1310:
1308:
1305:
1303:
1300:
1298:
1295:
1291:
1288:
1286:
1283:
1281:
1278:
1277:
1275:
1273:
1270:
1268:
1265:
1263:
1262:Stanford MIPS
1260:
1258:
1255:
1253:
1250:
1248:
1245:
1243:
1240:
1238:
1235:
1234:
1232:
1226:
1218:
1215:
1214:
1213:
1210:
1208:
1205:
1203:
1200:
1198:
1195:
1193:
1190:
1188:
1185:
1183:
1180:
1176:
1173:
1172:
1171:
1168:
1164:
1161:
1160:
1159:
1156:
1154:
1151:
1149:
1146:
1144:
1141:
1139:
1136:
1135:
1133:
1129:
1126:
1124:
1123:architectures
1118:
1112:
1109:
1107:
1104:
1102:
1099:
1097:
1094:
1092:
1091:Heterogeneous
1089:
1085:
1082:
1080:
1077:
1076:
1075:
1072:
1070:
1067:
1063:
1060:
1058:
1055:
1053:
1050:
1048:
1045:
1044:
1043:
1042:Memory access
1040:
1038:
1035:
1033:
1030:
1028:
1025:
1023:
1020:
1016:
1013:
1012:
1011:
1008:
1006:
1003:
1001:
998:
997:
995:
993:
989:
981:
978:
976:
975:Random-access
973:
971:
968:
966:
963:
962:
961:
958:
956:
955:Stack machine
953:
951:
948:
944:
941:
939:
936:
934:
931:
929:
926:
924:
921:
919:
916:
914:
911:
909:
906:
905:
904:
901:
897:
894:
892:
889:
887:
884:
882:
879:
877:
874:
872:
871:with datapath
869:
868:
867:
864:
862:
859:
857:
854:
853:
851:
849:
845:
841:
834:
829:
827:
822:
820:
815:
814:
811:
792:
788:
784:
778:
771:
770:1-55860-297-6
767:
763:
757:
743:on 2013-06-03
742:
738:
732:
725:
724:0-7637-0357-5
721:
717:
711:
705:, pp. 412-413
704:
703:0-88022-856-3
700:
696:
690:
683:
682:0-7897-2974-1
679:
675:
669:
665:
655:
652:
650:
647:
644:
641:
638:
635:
632:
629:
628:
622:
619:
615:
611:
597:
593:
590:
586:
582:
578:
575:
571:
568:
564:
561:
558:
554:
550:
546:
543:
539:
538:
532:
530:
526:
522:
518:
513:
512:coprocessor.
511:
507:
502:
500:
496:
492:
488:
484:
480:
476:
471:
469:
465:
462:released the
461:
460:Khronos Group
456:
454:
450:
446:
441:
439:
435:
430:
427:
423:
420:
416:
411:
409:
405:
401:
397:
393:
389:
385:
375:
372:
368:
358:
356:
351:
347:
345:
341:
337:
334:
329:
326:
317:
313:
310:
306:
301:
298:
294:
290:
287:
283:
280:The original
274:
266:
261:
251:
249:
245:
241:
238:demand for a
235:
233:
229:
225:
221:
218:
214:
210:
206:
202:
197:
195:
191:
187:
183:
179:
175:
171:
167:
163:
158:
156:
152:
142:
140:
137:
133:
129:
125:
124:PlayStation 2
121:
117:
113:
109:
105:
101:
97:
93:
89:
82:Functionality
79:
77:
73:
69:
65:
61:
57:
53:
49:
45:
41:
32:
19:
2767:Coprocessors
2741:Chip carrier
2679:Clock gating
2598:Mixed-signal
2495:Write buffer
2472:Control unit
2284:Clock signal
2030:
2023:accelerators
2005:Cypress PSoC
1885:
1662:Simultaneous
1479:Out-of-order
1111:Neuromorphic
992:Architecture
950:Belt machine
943:Zeno machine
876:Hierarchical
795:. Retrieved
786:
777:
761:
756:
745:. Retrieved
741:the original
731:
715:
710:
694:
689:
673:
668:
607:
514:
503:
472:
457:
442:
431:
412:
381:
364:
352:
348:
330:
322:
309:machine code
302:
279:
236:
198:
159:
148:
120:command list
85:
68:cryptography
39:
37:
18:Coprocessors
2526:Multiplexer
2490:Data buffer
2201:Single-core
2173:bit slicing
2031:Coprocessor
1886:Coprocessor
1767:performance
1689:Cooperative
1679:Speculative
1639:Distributed
1598:Superscalar
1583:Instruction
1551:Parallelism
1524:Speculative
1356:System/3x0
1228:Instruction
1005:Von Neumann
918:Post–Turing
787:hpcwire.com
581:Matrix 2000
555:, combines
445:SpursEngine
392:sound cards
371:68881/68882
333:Intel 80386
312:functions.
244:3D graphics
155:Channel I/O
139:vector unit
110:focused on
40:coprocessor
2756:Categories
2646:management
2541:Multiplier
2402:Logic gate
2392:Sequential
2299:Functional
2279:Clock rate
2252:Data cache
2225:Components
2206:Multi-core
2194:Core count
1684:Preemptive
1588:Pipelining
1571:Bit-serial
1514:Wide-issue
1459:Structural
1381:Tilera ISA
1347:MicroBlaze
1317:ETRAX CRIS
1212:Comparison
1057:Load–store
1037:Endianness
747:2013-06-16
660:References
574:multi-core
565:, used on
493:, and the
475:sensor hub
434:PowerQUICC
286:Intel 8087
186:Intel 8087
2580:Circuitry
2500:Microcode
2424:Registers
2267:coherence
2242:CPU cache
2100:Word size
1765:Processor
1409:Execution
1312:DEC Alpha
1290:Power ISA
1106:Cognitive
913:Universal
596:BBC Micro
525:TrueNorth
504:In 2012,
458:In 2008,
413:In 2006,
2518:Datapath
2211:Manycore
2183:variable
2021:Hardware
1657:Temporal
1337:OpenRISC
1032:Cellular
1022:Dataflow
1015:modified
791:Archived
637:Torrenza
625:See also
585:Tianhe-2
579:China's
497:for the
479:Apple M7
440:or NPU.
406:and the
361:Motorola
344:hardware
205:TI-99/4A
174:AMD 9511
56:graphics
2694:Related
2625:Quantum
2615:Digital
2610:Boolean
2508:Counter
2407:Quantum
2168:512-bit
2163:256-bit
2158:128-bit
2001:(MPSoC)
1986:on chip
1984:Systems
1802:(FLOPS)
1615:Process
1464:Control
1446:Hazards
1332:Itanium
1327:Unicore
1285:PowerPC
1010:Harvard
970:Pointer
965:Counter
923:Quantum
797:7 April
567:servers
481:and M8
228:blitter
215:". The
145:History
2630:Switch
2620:Analog
2358:(IMC)
2329:(MMU)
2178:others
2153:64-bit
2148:48-bit
2143:32-bit
2138:24-bit
2133:16-bit
2128:15-bit
2123:12-bit
1960:Mobile
1876:Stream
1871:Barrel
1866:Vector
1855:(GPU)
1814:(SUPS)
1782:(IPC)
1634:Memory
1627:Vector
1610:Thread
1593:Scalar
1395:Others
1342:RISC-V
1307:SuperH
1276:Power
1272:MIPS-X
1247:PDP-11
1096:Fabric
848:Models
768:
722:
701:
680:
643:OpenCL
604:Trends
547:Using
529:Zeroth
527:, and
485:, the
464:OpenCL
355:Weitek
282:IBM PC
232:bitmap
224:Copper
207:, and
182:Weitek
134:-like
122:. The
2686:(PPW)
2644:Power
2536:Adder
2412:Array
2379:Logic
2340:(TLB)
2323:(FPU)
2317:(AGU)
2311:(ALU)
2301:units
2237:Cache
2118:8-bit
2113:4-bit
2108:1-bit
2072:(TPU)
2066:(DSP)
2060:(PPU)
2054:(VPU)
2043:(GPU)
2012:(NoC)
1995:(SoC)
1930:(PoP)
1924:(SiP)
1918:(MCM)
1859:GPGPU
1849:(CPU)
1839:Types
1820:(PPW)
1808:(TPS)
1796:(IPS)
1788:(CPI)
1559:Level
1370:S/390
1365:S/370
1360:S/360
1302:SPARC
1280:POWER
1163:TRIPS
1131:Types
572:Some
506:Intel
419:PhysX
415:AGEIA
340:80387
254:Intel
217:Amiga
42:is a
2664:ACPI
2397:Glue
2289:FIFO
2232:Core
1970:ASIP
1911:CPLD
1906:FPOA
1901:FPGA
1896:ASIC
1749:SPMD
1744:MIMD
1739:MISD
1732:SWAR
1712:SIMD
1707:SISD
1622:Data
1605:Task
1576:Word
1322:M32R
1267:MIPS
1230:sets
1197:ZISC
1192:NISC
1187:OISC
1182:MISC
1175:EPIC
1170:VLIW
1158:EDGE
1148:RISC
1143:CISC
1052:HUMA
1047:NUMA
799:2018
766:ISBN
720:ISBN
699:ISBN
678:ISBN
618:GPUs
610:SIMD
553:Zynq
549:FPGA
540:The
489:and
453:SPUs
449:Cell
443:The
398:and
384:GPUs
365:The
331:The
325:8089
305:8086
192:and
180:and
136:SIMD
88:FPUs
2659:APM
2654:PMU
2546:CPU
2503:ROM
2274:Bus
1891:PAL
1566:Bit
1352:LMC
1257:ARM
1252:x86
1242:VAX
614:DSP
557:ARM
426:CPU
422:PPU
293:FPU
260:x87
246:in
209:MSX
132:DSP
126:'s
98:'s
96:CPU
70:or
48:CPU
2758::
2593:3D
785:.
523:,
501:.
470:.
250:.
203:,
188:,
176:,
66:,
62:,
58:,
54:,
38:A
832:e
825:t
818:v
801:.
750:.
591:.
20:)
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.