305:). Availability of a system may also be increased by the strategy of focusing on increasing testability and maintainability and not on reliability. Improving maintainability is generally easier than reliability. Maintainability estimates (repair rates) are also generally more accurate. However, because the uncertainties in the reliability estimates are in most cases very large, it is likely to dominate the availability (prediction uncertainty) problem, even while maintainability levels are very high.
442:
commands that a computer card's voltage output be set high-low-high-low and so on at a rate of 1000 Hz. The operating system may choose to adjust the scheduling of each transition (high-low or low-high) based on an internal clock. The latency is the delay between the process instruction commanding the transition and the hardware actually transitioning the voltage from high to low or low to high.
438:
can be transmitted, as there is often a limit on the amount of information that is "in-flight" at any one moment. In the field of human-machine interaction, perceptible latency (delay between what the user commands and when the computer provides the results) has a strong effect on user satisfaction and usability.
472:
throughput of a computer network. The reason for this usage is that according to
Hartley's law, the maximum data rate of a physical communication link is proportional to its bandwidth in hertz, which is sometimes called frequency bandwidth, spectral bandwidth, RF bandwidth, signal bandwidth or analog bandwidth.
1140:
switches can change N and CPI—the benchmark executes faster if the new compiler can improve N or C without making the other worse, but often there is a trade-off between them—is it better, for example, to use a few complicated instructions that take a long time to execute, or to use instructions that
597:
Compression is useful because it helps reduce resource usage, such as data storage space or transmission capacity. Because compressed data must be decompressed to use, this extra processing imposes computational or other costs through decompression; this situation is far from being a free lunch. Data
698:
In software engineering, performance testing is, in general, conducted to determine how a system performs in terms of responsiveness and stability under a particular workload. It can also serve to investigate, measure, validate, or verify other quality attributes of the system, such as scalability,
1168:, etc.), while (hopefully) not sacrificing too much clock frequency—leading to a brainiac CPU design. For a given instruction set (and therefore fixed N) and semiconductor process, the maximum single-thread performance (1/t) requires a balance between brainiac techniques and speedracer techniques.
1031:
compiler. It cannot be determined from the number of lines of HLL source code. N is not affected by other processes running on the same processor. The significant point here is that hardware normally does not keep track of (or at least make easily available) a value of N for executed programs. The
437:
The precise definition of latency depends on the system being observed and the nature of stimulation. In communications, the lower limit of latency is determined by the medium being used for communications. In reliable two-way communication systems, latency limits the maximum rate that information
826:
The amount of time an application takes to start up, or a file to download, is not made faster by showing a startup screen (see Splash screen) or a file progress dialog box. However, it satisfies some human needs: it appears faster to the user as well as provides a visual cue to let them know the
433:
Latency is a time delay between the cause and the effect of some physical change in the system being observed. Latency is a result of the limited velocity with which any physical interaction can take place. This velocity is always lower or equal to speed of light. Therefore, every physical system
167:
Performance engineering within systems engineering encompasses the set of roles, skills, activities, practices, tools, and deliverables applied at every phase of the systems development life cycle which ensures that a solution will be designed, implemented, and operationally supported to meet the
441:
Computers run sets of instructions called a process. In operating systems, the execution of the process can be postponed if other processes are also executing. In addition, the operating system can schedule when to perform the action that the process is commanding. For example, suppose a process
471:
Bandwidth sometimes defines the net bit rate (aka. peak bit rate, information rate, or physical layer useful bit rate), channel capacity, or the maximum throughput of a logical or physical communication path in a digital communication system. For example, bandwidth tests measure the maximum
769:
performance. This is typically a computer application, but the same methods can be applied to economic markets, bureaucracies or other complex systems. The motivation for such activity is called a performance problem, which can be real or anticipated. Most systems will respond to increased
214:
designed to meet the challenges associated with application performance in increasingly distributed mobile, cloud and terrestrial IT environments. It includes the roles, skills, activities, practices, tools and deliverables applied at every phase of the application lifecycle that ensure an
1148:, and so cannot change N. Sometimes a designer focuses on improving performance by making significant improvements in f (with techniques such as deeper pipelines and faster caches), while (hopefully) not sacrificing too much C—leading to a
620:
The effect of computing on the environment, during manufacturing and recycling as well as during use. Measurements are taken with the objectives of reducing waste, reducing hazardous materials, and minimizing a computer's
187:
However, sometimes pushing one type of performance to an extreme leads to a CPU with worse overall performance, because other important aspects were sacrificed to get one impressive-looking number, for example, the chip's
830:
In most cases, increasing real performance increases perceived performance, but when real performance cannot be increased due to physical limitations, techniques can be used to increase perceived performance.
415:, defines the notion of channel capacity and provides a mathematical model by which one can compute it. The key result states that the capacity of the channel, as defined above, is given by the maximum of the
179:
with better overall performance by improving one of the aspects of performance, presented below, without sacrificing the CPU's performance in other areas. For example, building the CPU out of better, faster
363:
architecture) to be able to run a large base of pre-existing, pre-compiled software. Being relatively uninformed on computer benchmarks, some of them pick a particular CPU based on operating frequency (see
467:
In computer networking, bandwidth is a measurement of bit-rate of available or consumed data communication resources, expressed in bits per second or multiples of it (bit/s, kbit/s, Mbit/s, Gbit/s, etc.).
702:
Performance testing is a subset of performance engineering, an emerging computer science practice which strives to build performance into the implementation, design, and architecture of a system.
501:
in bit/s/Hz/area unit, bit/s/Hz/site or bit/s/Hz/cell, is the maximum system throughput (aggregate throughput) divided by the analog bandwidth and some measure of the system coverage area.
931:
887:
542:
Scalability is the ability of a system, network, or process to handle a growing amount of work in a capable manner or its ability to be enlarged to accommodate that growth.
1126:
1087:
971:
583:
For spaceflight computers, the processing speed per watt ratio is a more useful performance criterion than raw processing speed due to limited on-board resources of power.
134:
in computer performance means the same thing that performance means in other contexts, that is, it means "How well is the computer doing the work it is supposed to do?"
110:
The performance of any computer system can be evaluated in measurable, technical terms, using one or more of the metrics listed above. This way the performance can be
1052:
993:
520:, which is 'seconds per message' or 'seconds per output', throughput can be used to relate a computational device performing a dedicated function such as an
606:
This is an important performance feature of mobile systems, from the smart phones you keep in your pocket to the portable embedded systems in a spacecraft.
1182:
673:
319:
Response time is the total amount of time it takes to respond to a request for service. In computing, that service can be any unit of work from a simple
1419:
819:
Perceived performance, in computer engineering, refers to how quickly a software feature appears to perform its task. The concept applies mainly to
1192:
1023:(if available) or by estimation—itself based partly on estimated or actual frequency distribution of input variables and by examining generated
301:
Availability of a system is typically measured as a factor of its reliability - as reliability increases, so does availability (that is, less
804:
If the modification makes the performance better, adopt it. If the modification makes the performance worse, put it back to the way it was.
1318:
205:
1301:
751:). A number of different techniques may be used by profilers, such as event-based, statistical, instrumented, and simulation methods.
1152:
CPU design. Sometimes a designer focuses on improving performance by making significant improvements in CPI (with techniques such as
1374:
1390:
711:
337:
Transmission time – How long it takes to move the request to the computer doing the work and the response back to the requestor.
1197:
1028:
733:
725:
580:, pick CPUs based on their speed per watt of power, because the cost of powering the CPU outweighs the cost of the CPU itself.
494:
156:
1453:
693:
394:
122:
1401:
558:). This becomes especially important for systems with limited power sources such as solar, batteries, and human power.
215:
application will be designed, implemented and operationally supported to meet non-functional performance requirements.
508:
has a single input and a single output, and operates on discrete packets of information. Examples of such blocks are
1323:
Lecture notes for M.Sc. Data
Communication Networks and Distributed Systems D51 -- Basic Communications and Networks
1223:
498:
314:
232:
55:
50:
instructions. When it comes to high computer performance, one or more of the following factors might be involved:
1020:
729:
732:, or frequency and duration of function calls. The most common use of profiling information is to aid program
721:
577:
419:
between the input and output of the channel, where the maximization is with respect to the input distribution.
390:
996:
347:
276:
211:
176:
1032:
value can therefore only be accurately determined by instruction set simulation, which is rarely practiced.
121:
Whilst the above definition relates to a scientific, technical approach, the following definition given by
1417:
1269:
1177:
1153:
1129:
1090:
509:
1341:
171:
Performance engineering continuously deals with trade-offs between types of performance. Occasionally a
1213:
1165:
1137:
896:
852:
814:
792:
677:
662:
656:
486:
In general terms, throughput is the rate of production or the rate at which something can be processed.
462:
428:
320:
252:
240:
91:
820:
717:
622:
567:
489:
In communication networks, throughput is essentially synonymous to digital bandwidth consumption. In
268:
244:
148:
1434:
1274:
1233:
1187:
791:
Identify the part of the system that is critical for improving the performance. This is called the
446:
398:
334:
Wait time - How long the request has to wait for requests queued ahead of it before it gets to run.
260:
1098:
1059:
943:
1149:
774:
with some degree of decreasing performance. A system's ability to accept a higher load is called
760:
644:
573:
525:
416:
404:
69:
1326:
1297:
1228:
555:
517:
513:
505:
450:
408:
144:
98:
84:
1161:
771:
634:
592:
490:
380:
272:
236:
152:
76:
47:
371:
Some system designers building parallel computers pick CPUs based on the speed per dollar.
1423:
1208:
1202:
1145:
1004:
1000:
615:
224:
39:
1252:
Computer
Performance Analysis with Mathematica by Arnold O. Allen, Academic Press, 1994.
449:
systems want to guarantee worst-case response. That is easier to do when the CPU has low
1378:
397:, the channel capacity of a given channel is the limiting information rate (in units of
42:. Outside of specific contexts, computer performance is estimated in terms of accuracy,
551:
365:
193:
778:, and modifying a system to handle a higher load is synonymous to performance tuning.
17:
1447:
785:
Assess the problem and establish numeric values that categorize acceptable behavior.
1037:
1024:
978:
412:
296:
248:
228:
1293:
1157:
775:
740:
661:
Because there are so many programs to test a CPU on all aspects of performance,
537:
386:
264:
1141:
execute very quickly, although it takes more of them to execute the benchmark?
1136:
Even on one machine, a different compiler or the same compiler with different
640:
481:
401:
per unit time) that can be achieved with arbitrarily small error probability.
256:
189:
181:
172:
62:
43:
31:
434:
that has non-zero spatial dimensions will experience some sort of latency.
210:
Application
Performance Engineering (APE) is a specific methodology within
114:
Compared relative to other systems or the same system before/after changes
324:
302:
280:
1266:
Measuring
Program Similarity: Experiments with SPEC CPU Benchmark Suites
680:
benchmark developed by the
Embedded Microprocessor Benchmark Consortium
1218:
766:
720:, profiling ("program profiling", "software profiling") is a form of
669:
516:. Because the units of throughput are the reciprocal of the unit for
647:(IC). Transistor count is the most common measure of IC complexity.
681:
360:
357:
351:
521:
117:
In absolute terms, e.g. for fulfilling a contractual obligation
1160:
CPUs, larger caches, caches with improved hit rates, improved
284:
598:
compression is subject to a space–time complexity trade-off.
385:
Channel capacity is the tightest upper bound on the rate of
1375:"EEMBC -- the Embedded Microprocessor Benchmark Consortium"
1290:
The Every
Computer Performance Book, Chapter 3: Useful laws
1144:
A CPU designer is often required to implement a particular
788:
Measure the performance of the system before modification.
739:
Profiling is achieved by instrumenting either the program
528:
to a communications channel, simplifying system analysis.
331:
Service time - How long it takes to do the work requested.
801:
Measure the performance of the system after modification.
845:) required to execute a particular benchmark program is
798:
Modify that part of the system to remove the bottleneck.
724:
that measures, for example, the space (memory) or time
1101:
1062:
1040:
981:
946:
907:
863:
356:
Most consumers pick a computer architecture (normally
995:
is the number of instructions actually executed (the
899:
855:
1402:"The Next-Generation SC-7 RISC Spaceflight Computer"
743:
or its binary executable form using a tool called a
1416:Paul DeMone. "The Incredible Shrinking CPU". 2004.
168:performance requirements defined for the solution.
1120:
1081:
1046:
987:
965:
925:
881:
327:. The response time is the sum of three numbers:
973:is "the performance" in terms of time-to-execute
27:Amount of useful work accomplished by a computer
668:The most famous benchmarks are the SPECint and
38:is the amount of useful work accomplished by a
125:would be useful for a non-technical audience:
8:
1054:is the clock frequency in cycles per second.
1412:
1410:
1183:Computer performance by orders of magnitude
674:Standard Performance Evaluation Corporation
504:In integrated circuits, often a block in a
453:and when it has a deterministic response.
1273:
1108:
1100:
1069:
1061:
1039:
980:
953:
945:
906:
898:
862:
854:
765:Performance tuning is the improvement of
389:that can be reliably transmitted over a
1435:"Brainiacs, Speed Demons, and Farewell"
1359:Thomas M. Cover, Joy A. Thomas (2006).
1245:
1193:Latency oriented processor architecture
781:Systematic tuning follows these steps:
106:Technical and non-technical definitions
87:of the computing system or application.
7:
1400:D. J. Shirley; and M. K. McLelland.
1346:Information and Measurement, 2nd ed
206:Application performance engineering
200:Application performance engineering
1363:. John Wiley & Sons, New York.
926:{\displaystyle P={\tfrac {If}{N}}}
882:{\displaystyle t={\tfrac {NC}{f}}}
827:system is handling their request.
25:
699:reliability, and resource usage.
730:usage of particular instructions
712:Profiling (computer programming)
706:Profiling (performance analysis)
140:As an aspect of software quality
1198:Optimization (computer science)
495:cellular communication networks
151:response time, is an aspect of
1361:Elements of Information Theory
1229:Understanding Your PC Hardware
1:
1121:{\textstyle I={\frac {1}{C}}}
1082:{\textstyle C={\frac {1}{I}}}
966:{\textstyle P={\frac {1}{t}}}
694:Software performance testing
688:Software performance testing
395:noisy-channel coding theorem
227:(things to measure) include
157:human–computer interactions
1470:
1342:"Signals look like noise!"
1224:Cache replacement policies
841:The total amount of time (
812:
758:
709:
691:
654:
632:
613:
590:
572:System designers building
565:
535:
499:system spectral efficiency
479:
460:
445:System designers building
426:
378:
345:
315:Response time (technology)
312:
294:
287:benchmarks are available.
203:
147:performance, particularly
65:(rate of processing work).
58:for a given piece of work.
1254:$ 1.1 Introduction, pg 1.
1132:(IPC) for this benchmark.
1093:(CPI) for this benchmark.
1021:instruction set simulator
1015:can either be determined
175:can find a way to make a
75:Fast (or highly compact)
1268:, 2005, pp. 10–20,
722:dynamic program analysis
672:benchmarks developed by
997:instruction path length
726:complexity of a program
348:Instructions per second
277:instruction path length
212:performance engineering
163:Performance engineering
46:and speed of executing
1178:Algorithmic efficiency
1154:out-of-order execution
1130:instructions per cycle
1122:
1091:cycles per instruction
1083:
1048:
989:
967:
927:
883:
554:used by the computer (
391:communications channel
219:Aspects of performance
137:
18:Performance (software)
1288:Wescott, Bob (2013).
1214:Hardware acceleration
1166:speculative execution
1138:compiler optimization
1123:
1084:
1049:
990:
968:
928:
884:
815:Perceived performance
809:Perceived performance
657:Benchmark (computing)
614:Further information:
463:Bandwidth (computing)
429:Latency (engineering)
323:to loading a complex
223:Computer performance
155:that is important in
127:
1454:Computer performance
1099:
1060:
1038:
979:
944:
897:
853:
837:Performance Equation
718:software engineering
623:ecological footprint
610:Environmental impact
568:Performance per watt
562:Performance per watt
269:performance per watt
149:software application
36:computer performance
1234:Relative efficiency
1188:Network performance
447:real-time computing
261:relative efficiency
68:Low utilization of
1437:by Linley Gwennap
1422:2012-05-31 at the
1319:"Channel capacity"
1118:
1079:
1044:
985:
963:
923:
921:
879:
877:
761:Performance tuning
755:Performance tuning
678:Certification Mark
645:integrated circuit
574:parallel computers
526:embedded processor
514:binary multipliers
417:mutual information
405:Information theory
79:and decompression.
70:computing resource
1162:branch prediction
1116:
1077:
1007:strongly affects
961:
920:
890:, or equivalently
876:
587:Compression ratio
578:Google's hardware
556:power consumption
546:Power consumption
518:propagation delay
506:data flow diagram
491:wireless networks
451:interrupt latency
409:Claude E. Shannon
273:compression ratio
145:Computer software
99:data transmission
85:High availability
16:(Redirected from
1461:
1438:
1432:
1426:
1414:
1405:
1398:
1392:
1389:
1387:
1386:
1377:. Archived from
1371:
1365:
1364:
1356:
1350:
1349:
1337:
1331:
1330:
1325:. Archived from
1314:
1308:
1307:
1285:
1279:
1278:
1277:
1262:
1256:
1250:
1127:
1125:
1124:
1119:
1117:
1109:
1088:
1086:
1085:
1080:
1078:
1070:
1053:
1051:
1050:
1045:
994:
992:
991:
986:
972:
970:
969:
964:
962:
954:
932:
930:
929:
924:
922:
916:
908:
888:
886:
885:
880:
878:
872:
864:
665:were developed.
635:Transistor count
629:Transistor count
593:Data compression
381:Channel capacity
375:Channel capacity
342:Processing speed
237:channel capacity
153:software quality
77:data compression
48:computer program
21:
1469:
1468:
1464:
1463:
1462:
1460:
1459:
1458:
1444:
1443:
1442:
1441:
1433:
1429:
1424:Wayback Machine
1415:
1408:
1399:
1395:
1384:
1382:
1373:
1372:
1368:
1358:
1357:
1353:
1339:
1338:
1334:
1317:Saleem Bhatti.
1316:
1315:
1311:
1304:
1287:
1286:
1282:
1264:
1263:
1259:
1251:
1247:
1242:
1209:instruction set
1203:RAM update rate
1174:
1146:instruction set
1128:is the average
1097:
1096:
1089:is the average
1058:
1057:
1036:
1035:
1011:. The value of
1005:instruction set
977:
976:
942:
941:
909:
895:
894:
865:
851:
850:
839:
821:user acceptance
817:
811:
763:
757:
714:
708:
696:
690:
659:
653:
637:
631:
618:
616:Green computing
612:
604:
602:Size and weight
595:
589:
570:
564:
548:
540:
534:
484:
478:
465:
459:
431:
425:
407:, developed by
383:
377:
354:
346:Main articles:
344:
317:
311:
299:
293:
245:completion time
221:
208:
202:
165:
142:
108:
40:computer system
28:
23:
22:
15:
12:
11:
5:
1467:
1465:
1457:
1456:
1446:
1445:
1440:
1439:
1427:
1406:
1393:
1366:
1351:
1332:
1329:on 2007-08-21.
1309:
1303:978-1482657753
1302:
1280:
1275:10.1.1.123.501
1257:
1244:
1243:
1241:
1238:
1237:
1236:
1231:
1226:
1221:
1216:
1211:
1205:
1200:
1195:
1190:
1185:
1180:
1173:
1170:
1134:
1133:
1115:
1112:
1107:
1104:
1094:
1076:
1073:
1068:
1065:
1055:
1047:{\textstyle f}
1043:
1033:
988:{\textstyle N}
984:
974:
960:
957:
952:
949:
935:
934:
919:
915:
912:
905:
902:
891:
875:
871:
868:
861:
858:
838:
835:
813:Main article:
810:
807:
806:
805:
802:
799:
796:
789:
786:
759:Main article:
756:
753:
710:Main article:
707:
704:
692:Main article:
689:
686:
655:Main article:
652:
649:
639:The number of
633:Main article:
630:
627:
611:
608:
603:
600:
591:Main article:
588:
585:
566:Main article:
563:
560:
552:electric power
550:The amount of
547:
544:
536:Main article:
533:
530:
480:Main article:
477:
474:
461:Main article:
458:
455:
427:Main article:
424:
421:
379:Main article:
376:
373:
366:megahertz myth
343:
340:
339:
338:
335:
332:
313:Main article:
310:
307:
295:Main article:
292:
289:
220:
217:
204:Main article:
201:
198:
194:megahertz myth
164:
161:
141:
138:
119:
118:
115:
107:
104:
103:
102:
95:
88:
82:
81:
80:
66:
59:
26:
24:
14:
13:
10:
9:
6:
4:
3:
2:
1466:
1455:
1452:
1451:
1449:
1436:
1431:
1428:
1425:
1421:
1418:
1413:
1411:
1407:
1403:
1397:
1394:
1391:
1381:on 2005-03-27
1380:
1376:
1370:
1367:
1362:
1355:
1352:
1347:
1343:
1336:
1333:
1328:
1324:
1320:
1313:
1310:
1305:
1299:
1295:
1291:
1284:
1281:
1276:
1271:
1267:
1261:
1258:
1255:
1249:
1246:
1239:
1235:
1232:
1230:
1227:
1225:
1222:
1220:
1217:
1215:
1212:
1210:
1206:
1204:
1201:
1199:
1196:
1194:
1191:
1189:
1186:
1184:
1181:
1179:
1176:
1175:
1171:
1169:
1167:
1163:
1159:
1155:
1151:
1147:
1142:
1139:
1131:
1113:
1110:
1105:
1102:
1095:
1092:
1074:
1071:
1066:
1063:
1056:
1041:
1034:
1030:
1026:
1022:
1018:
1014:
1010:
1006:
1002:
998:
982:
975:
958:
955:
950:
947:
940:
939:
938:
933:
917:
913:
910:
903:
900:
892:
889:
873:
869:
866:
859:
856:
848:
847:
846:
844:
836:
834:
832:
828:
824:
822:
816:
808:
803:
800:
797:
794:
790:
787:
784:
783:
782:
779:
777:
773:
768:
762:
754:
752:
750:
749:code profiler
746:
742:
737:
735:
731:
727:
723:
719:
713:
705:
703:
700:
695:
687:
685:
683:
679:
675:
671:
666:
664:
658:
650:
648:
646:
642:
636:
628:
626:
624:
617:
609:
607:
601:
599:
594:
586:
584:
581:
579:
575:
569:
561:
559:
557:
553:
545:
543:
539:
531:
529:
527:
523:
519:
515:
511:
507:
502:
500:
496:
492:
487:
483:
475:
473:
469:
464:
456:
454:
452:
448:
443:
439:
435:
430:
422:
420:
418:
414:
410:
406:
402:
400:
396:
392:
388:
382:
374:
372:
369:
367:
362:
359:
353:
349:
341:
336:
333:
330:
329:
328:
326:
322:
316:
309:Response time
308:
306:
304:
298:
290:
288:
286:
282:
278:
274:
270:
266:
262:
258:
254:
250:
246:
242:
238:
234:
233:response time
230:
226:
218:
216:
213:
207:
199:
197:
195:
191:
185:
183:
178:
174:
169:
162:
160:
158:
154:
150:
146:
139:
136:
135:
131:
126:
124:
116:
113:
112:
111:
105:
100:
96:
93:
89:
86:
83:
78:
74:
73:
71:
67:
64:
60:
57:
56:response time
53:
52:
51:
49:
45:
41:
37:
33:
19:
1430:
1396:
1383:. Retrieved
1379:the original
1369:
1360:
1354:
1345:
1340:Jim Lesurf.
1335:
1327:the original
1322:
1312:
1289:
1283:
1265:
1260:
1253:
1248:
1143:
1135:
1025:machine code
1019:by using an
1016:
1012:
1008:
1001:code density
936:
893:
849:
842:
840:
833:
829:
825:
818:
780:
764:
748:
744:
738:
734:optimization
715:
701:
697:
667:
660:
638:
619:
605:
596:
582:
571:
549:
541:
503:
488:
485:
470:
466:
444:
440:
436:
432:
413:World War II
403:
384:
370:
355:
318:
300:
297:Availability
291:Availability
249:service time
229:availability
222:
209:
186:
173:CPU designer
170:
166:
143:
133:
129:
128:
123:Arnold Allen
120:
109:
35:
29:
1294:CreateSpace
1158:superscalar
1150:speed-demon
776:scalability
741:source code
641:transistors
538:Scalability
532:Scalability
512:modules or
399:information
387:information
265:scalability
182:transistors
132:performance
1385:2009-01-21
1240:References
793:bottleneck
663:benchmarks
651:Benchmarks
576:, such as
482:Throughput
476:Throughput
257:throughput
190:clock rate
63:throughput
44:efficiency
1270:CiteSeerX
1207:Complete
823:aspects.
457:Bandwidth
393:. By the
253:bandwidth
192:(see the
130:The word
92:bandwidth
32:computing
1448:Category
1420:Archived
1172:See also
1027:from an
745:profiler
676:and the
325:web page
303:downtime
281:speed up
1404:. p. 2.
1219:Speedup
1017:exactly
1003:of the
999:). The
423:Latency
411:during
321:disk IO
241:latency
225:metrics
1300:
1272:
937:where
767:system
728:, the
670:SPECfp
643:on an
497:, the
97:Short
54:Short
682:EEMBC
361:IA-32
358:Intel
352:FLOPS
101:time.
90:High
72:(s).
61:High
1298:ISBN
772:load
747:(or
522:ASIC
350:and
279:and
1029:HLL
716:In
524:or
510:FFT
493:or
368:).
285:CPU
196:).
177:CPU
30:In
1450::
1409:^
1344:.
1321:.
1296:.
1292:.
1164:,
1156:,
736:.
684:.
625:.
283:.
275:,
271:,
267:,
263:,
259:,
255:,
251:,
247:,
243:,
239:,
235:,
231:,
184:.
159:.
34:,
1388:.
1348:.
1306:.
1114:C
1111:1
1106:=
1103:I
1075:I
1072:1
1067:=
1064:C
1042:f
1013:N
1009:N
983:N
959:t
956:1
951:=
948:P
918:N
914:f
911:I
904:=
901:P
874:f
870:C
867:N
860:=
857:t
843:t
795:.
94:.
20:)
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.