294:). 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.
431:
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.
427:
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.
461:
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.
1129:
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
586:
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
687:
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,
1157:, 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.
1020:
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
426:
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
815:
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
422:
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
156:
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
430:
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
460:
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
758:
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
203:
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
1137:, 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
609:
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
176:
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
819:
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.
404:, 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
168:
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
352:
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
456:
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.).
691:
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.
490:
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.
920:
876:
531:
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.
1115:
1076:
960:
572:
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.
123:
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?"
99:
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
1041:
982:
509:, 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
595:
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.
666:
308:
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
1408:
233:
808:
Perceived performance, in computer engineering, refers to how quickly a software feature appears to perform its task. The concept applies mainly to
1012:(if available) or by estimation—itself based partly on estimated or actual frequency distribution of input variables and by examining generated
290:
Availability of a system is typically measured as a factor of its reliability - as reliability increases, so does availability (that is, less
1138:
793:
If the modification makes the performance better, adopt it. If the modification makes the performance worse, put it back to the way it was.
1307:
1171:
662:
740:). A number of different techniques may be used by profilers, such as event-based, statistical, instrumented, and simulation methods.
1141:
CPU design. Sometimes a designer focuses on improving performance by making significant improvements in CPI (with techniques such as
1363:
1379:
1181:
326:
Transmission time – How long it takes to move the request to the computer doing the work and the response back to the requestor.
569:, pick CPUs based on their speed per watt of power, because the cost of powering the CPU outweighs the cost of the CPU itself.
194:
1390:
1290:
566:
547:). This becomes especially important for systems with limited power sources such as solar, batteries, and human power.
204:
application will be designed, implemented and operationally supported to meet non-functional performance requirements.
497:
has a single input and a single output, and operates on discrete packets of information. Examples of such blocks are
1312:
Lecture notes for M.Sc. Data
Communication Networks and Distributed Systems D51 -- Basic Communications and Networks
700:
39:
instructions. When it comes to high computer performance, one or more of the following factors might be involved:
1186:
1017:
722:
714:
483:
721:, or frequency and duration of function calls. The most common use of profiling information is to aid program
408:
between the input and output of the channel, where the maximization is with respect to the input distribution.
145:
1442:
1021:
value can therefore only be accurately determined by instruction set simulation, which is rarely practiced.
682:
383:
111:
110:
Whilst the above definition relates to a scientific, technical approach, the following definition given by
1406:
1330:
160:
Performance engineering continuously deals with trade-offs between types of performance. Occasionally a
885:
841:
475:
In general terms, throughput is the rate of production or the rate at which something can be processed.
1212:
487:
303:
221:
44:
1009:
718:
478:
In communication networks, throughput is essentially synonymous to digital bandwidth consumption. In
1423:
1263:
780:
Identify the part of the system that is critical for improving the performance. This is called the
710:
379:
323:
Wait time - How long the request has to wait for requests queued ahead of it before it gets to run.
1087:
1048:
932:
985:
763:
with some degree of decreasing performance. A system's ability to accept a higher load is called
336:
265:
200:
165:
1315:
1258:
1166:
1142:
1118:
1079:
498:
1217:
989:
1202:
1154:
1126:
803:
781:
651:
645:
451:
417:
309:
241:
229:
80:
809:
706:
611:
556:
360:
Some system designers building parallel computers pick CPUs based on the speed per dollar.
257:
137:
8:
1241:
Computer
Performance Analysis with Mathematica by Arnold O. Allen, Academic Press, 1994.
1222:
1176:
438:
systems want to guarantee worst-case response. That is easier to do when the CPU has low
435:
387:
249:
1367:
386:, the channel capacity of a given channel is the limiting information rate (in units of
31:. Outside of specific contexts, computer performance is estimated in terms of accuracy,
749:
633:
562:
514:
405:
393:
58:
767:, and modifying a system to handle a higher load is synonymous to performance tuning.
1286:
544:
506:
502:
494:
439:
397:
133:
87:
73:
774:
Assess the problem and establish numeric values that categorize acceptable behavior.
1150:
1026:
967:
760:
623:
581:
479:
369:
261:
225:
141:
65:
36:
1412:
1197:
1191:
1134:
993:
604:
213:
28:
650:
Because there are so many programs to test a CPU on all aspects of performance,
1130:
execute very quickly, although it takes more of them to execute the benchmark?
540:
354:
182:
1125:
Even on one machine, a different compiler or the same compiler with different
390:
per unit time) that can be achieved with arbitrarily small error probability.
1436:
1013:
423:
that has non-zero spatial dimensions will experience some sort of latency.
401:
285:
237:
217:
199:
Application
Performance Engineering (APE) is a specific methodology within
103:
Compared relative to other systems or the same system before/after changes
1282:
1146:
764:
729:
526:
375:
253:
1255:
Measuring
Program Similarity: Experiments with SPEC CPU Benchmark Suites
669:
benchmark developed by the
Embedded Microprocessor Benchmark Consortium
629:
470:
245:
178:
170:
161:
51:
32:
709:, profiling ("program profiling", "software profiling") is a form of
505:. Because the units of throughput are the reciprocal of the unit for
20:
636:(IC). Transistor count is the most common measure of IC complexity.
313:
291:
269:
1207:
755:
658:
106:
In absolute terms, e.g. for fulfilling a contractual obligation
1347:
1149:
CPUs, larger caches, caches with improved hit rates, improved
670:
587:
compression is subject to a space–time complexity trade-off.
349:
346:
340:
374:
Channel capacity is the tightest upper bound on the rate of
1364:"EEMBC -- the Embedded Microprocessor Benchmark Consortium"
1279:
The Every
Computer Performance Book, Chapter 3: Useful laws
1133:
A CPU designer is often required to implement a particular
510:
777:
Measure the performance of the system before modification.
728:
Profiling is achieved by instrumenting either the program
517:
to a communications channel, simplifying system analysis.
320:
Service time - How long it takes to do the work requested.
94:
790:
Measure the performance of the system after modification.
273:
834:) required to execute a particular benchmark program is
787:
Modify that part of the system to remove the bottleneck.
713:
that measures, for example, the space (memory) or time
1090:
1051:
1029:
970:
935:
896:
852:
345:
Most consumers pick a computer architecture (normally
188:
984:
is the number of instructions actually executed (the
888:
844:
1391:"The Next-Generation SC-7 RISC Spaceflight Computer"
732:
or its binary executable form using a tool called a
1405:Paul DeMone. "The Incredible Shrinking CPU". 2004.
694:
157:performance requirements defined for the solution.
128:
1109:
1070:
1035:
976:
954:
914:
870:
316:. The response time is the sum of three numbers:
1434:
962:is "the performance" in terms of time-to-execute
16:Amount of useful work accomplished by a computer
676:
657:The most famous benchmarks are the SPECint and
27:is the amount of useful work accomplished by a
114:would be useful for a non-technical audience:
1043:is the clock frequency in cycles per second.
1401:
1399:
1172:Computer performance by orders of magnitude
663:Standard Performance Evaluation Corporation
493:In integrated circuits, often a block in a
442:and when it has a deterministic response.
151:
1305:
1262:
754:Performance tuning is the improvement of
207:
1396:
797:
378:that can be reliably transmitted over a
1424:"Brainiacs, Speed Demons, and Farewell"
1348:Thomas M. Cover, Joy A. Thomas (2006).
1276:
1182:Latency oriented processor architecture
825:
770:Systematic tuning follows these steps:
598:
550:
95:Technical and non-technical definitions
76:of the computing system or application.
1435:
1328:
743:
1389:D. J. Shirley; and M. K. McLelland.
575:
534:
1335:Information and Measurement, 2nd ed
617:
363:
330:
195:Application performance engineering
189:Application performance engineering
13:
1352:. John Wiley & Sons, New York.
915:{\displaystyle P={\tfrac {If}{N}}}
871:{\displaystyle t={\tfrac {NC}{f}}}
816:system is handling their request.
590:
14:
1454:
688:reliability, and resource usage.
719:usage of particular instructions
701:Profiling (computer programming)
695:Profiling (performance analysis)
297:
129:As an aspect of software quality
1417:
1187:Optimization (computer science)
484:cellular communication networks
279:
140:response time, is an aspect of
1383:
1356:
1350:Elements of Information Theory
1341:
1322:
1299:
1270:
1247:
1235:
1218:Understanding Your PC Hardware
520:
1:
1228:
1110:{\textstyle I={\frac {1}{C}}}
1071:{\textstyle C={\frac {1}{I}}}
955:{\textstyle P={\frac {1}{t}}}
639:
464:
683:Software performance testing
677:Software performance testing
445:
384:noisy-channel coding theorem
216:(things to measure) include
7:
1160:
146:human–computer interactions
10:
1459:
1331:"Signals look like noise!"
1213:Cache replacement policies
830:The total amount of time (
801:
747:
698:
680:
643:
621:
602:
579:
561:System designers building
554:
524:
488:system spectral efficiency
468:
449:
434:System designers building
415:
411:
367:
334:
304:Response time (technology)
301:
283:
276:benchmarks are available.
192:
136:performance, particularly
54:(rate of processing work).
47:for a given piece of work.
1243:$ 1.1 Introduction, pg 1.
1121:(IPC) for this benchmark.
1082:(CPI) for this benchmark.
1010:instruction set simulator
1004:can either be determined
164:can find a way to make a
64:Fast (or highly compact)
1257:, 2005, pp. 10–20,
711:dynamic program analysis
661:benchmarks developed by
986:instruction path length
715:complexity of a program
337:Instructions per second
266:instruction path length
201:performance engineering
152:Performance engineering
35:and speed of executing
1167:Algorithmic efficiency
1143:out-of-order execution
1119:instructions per cycle
1111:
1080:cycles per instruction
1072:
1037:
978:
956:
916:
872:
543:used by the computer (
380:communications channel
208:Aspects of performance
126:
1277:Wescott, Bob (2013).
1203:Hardware acceleration
1155:speculative execution
1127:compiler optimization
1112:
1073:
1038:
979:
957:
917:
873:
804:Perceived performance
798:Perceived performance
646:Benchmark (computing)
603:Further information:
452:Bandwidth (computing)
418:Latency (engineering)
312:to loading a complex
212:Computer performance
144:that is important in
116:
1443:Computer performance
1088:
1049:
1027:
968:
933:
886:
842:
826:Performance Equation
707:software engineering
612:ecological footprint
599:Environmental impact
557:Performance per watt
551:Performance per watt
258:performance per watt
138:software application
25:computer performance
1223:Relative efficiency
1177:Network performance
436:real-time computing
250:relative efficiency
57:Low utilization of
1426:by Linley Gwennap
1411:2012-05-31 at the
1308:"Channel capacity"
1107:
1068:
1033:
974:
952:
912:
910:
868:
866:
750:Performance tuning
744:Performance tuning
667:Certification Mark
634:integrated circuit
563:parallel computers
515:embedded processor
503:binary multipliers
406:mutual information
394:Information theory
68:and decompression.
59:computing resource
1151:branch prediction
1105:
1066:
996:strongly affects
950:
909:
879:, or equivalently
865:
576:Compression ratio
567:Google's hardware
545:power consumption
535:Power consumption
507:propagation delay
495:data flow diagram
480:wireless networks
440:interrupt latency
398:Claude E. Shannon
262:compression ratio
134:Computer software
88:data transmission
74:High availability
1450:
1427:
1421:
1415:
1403:
1394:
1387:
1381:
1378:
1376:
1375:
1366:. Archived from
1360:
1354:
1353:
1345:
1339:
1338:
1326:
1320:
1319:
1314:. Archived from
1303:
1297:
1296:
1274:
1268:
1267:
1266:
1251:
1245:
1239:
1116:
1114:
1113:
1108:
1106:
1098:
1077:
1075:
1074:
1069:
1067:
1059:
1042:
1040:
1039:
1034:
983:
981:
980:
975:
961:
959:
958:
953:
951:
943:
921:
919:
918:
913:
911:
905:
897:
877:
875:
874:
869:
867:
861:
853:
654:were developed.
624:Transistor count
618:Transistor count
582:Data compression
370:Channel capacity
364:Channel capacity
331:Processing speed
226:channel capacity
142:software quality
66:data compression
37:computer program
1458:
1457:
1453:
1452:
1451:
1449:
1448:
1447:
1433:
1432:
1431:
1430:
1422:
1418:
1413:Wayback Machine
1404:
1397:
1388:
1384:
1373:
1371:
1362:
1361:
1357:
1346:
1342:
1327:
1323:
1306:Saleem Bhatti.
1304:
1300:
1293:
1275:
1271:
1253:
1252:
1248:
1240:
1236:
1231:
1198:instruction set
1192:RAM update rate
1163:
1135:instruction set
1117:is the average
1097:
1089:
1086:
1085:
1078:is the average
1058:
1050:
1047:
1046:
1028:
1025:
1024:
1000:. The value of
994:instruction set
969:
966:
965:
942:
934:
931:
930:
898:
895:
887:
884:
883:
854:
851:
843:
840:
839:
828:
810:user acceptance
806:
800:
752:
746:
703:
697:
685:
679:
648:
642:
626:
620:
607:
605:Green computing
601:
593:
591:Size and weight
584:
578:
559:
553:
537:
529:
523:
473:
467:
454:
448:
420:
414:
396:, developed by
372:
366:
343:
335:Main articles:
333:
306:
300:
288:
282:
234:completion time
210:
197:
191:
154:
131:
97:
29:computer system
17:
12:
11:
5:
1456:
1446:
1445:
1429:
1428:
1416:
1395:
1382:
1355:
1340:
1321:
1318:on 2007-08-21.
1298:
1292:978-1482657753
1291:
1269:
1264:10.1.1.123.501
1246:
1233:
1232:
1230:
1227:
1226:
1225:
1220:
1215:
1210:
1205:
1200:
1194:
1189:
1184:
1179:
1174:
1169:
1162:
1159:
1123:
1122:
1104:
1101:
1096:
1093:
1083:
1065:
1062:
1057:
1054:
1044:
1036:{\textstyle f}
1032:
1022:
977:{\textstyle N}
973:
963:
949:
946:
941:
938:
924:
923:
908:
904:
901:
894:
891:
880:
864:
860:
857:
850:
847:
827:
824:
802:Main article:
799:
796:
795:
794:
791:
788:
785:
778:
775:
748:Main article:
745:
742:
699:Main article:
696:
693:
681:Main article:
678:
675:
644:Main article:
641:
638:
628:The number of
622:Main article:
619:
616:
600:
597:
592:
589:
580:Main article:
577:
574:
555:Main article:
552:
549:
541:electric power
539:The amount of
536:
533:
525:Main article:
522:
519:
469:Main article:
466:
463:
450:Main article:
447:
444:
416:Main article:
413:
410:
368:Main article:
365:
362:
355:megahertz myth
332:
329:
328:
327:
324:
321:
302:Main article:
299:
296:
284:Main article:
281:
278:
209:
206:
193:Main article:
190:
187:
183:megahertz myth
153:
150:
130:
127:
108:
107:
104:
96:
93:
92:
91:
84:
77:
71:
70:
69:
55:
48:
15:
9:
6:
4:
3:
2:
1455:
1444:
1441:
1440:
1438:
1425:
1420:
1414:
1410:
1407:
1402:
1400:
1392:
1386:
1380:
1370:on 2005-03-27
1369:
1365:
1359:
1351:
1344:
1336:
1332:
1325:
1317:
1313:
1309:
1302:
1294:
1288:
1284:
1280:
1273:
1265:
1260:
1256:
1250:
1244:
1238:
1234:
1224:
1221:
1219:
1216:
1214:
1211:
1209:
1206:
1204:
1201:
1199:
1195:
1193:
1190:
1188:
1185:
1183:
1180:
1178:
1175:
1173:
1170:
1168:
1165:
1164:
1158:
1156:
1152:
1148:
1144:
1140:
1136:
1131:
1128:
1120:
1102:
1099:
1094:
1091:
1084:
1081:
1063:
1060:
1055:
1052:
1045:
1030:
1023:
1019:
1015:
1011:
1007:
1003:
999:
995:
991:
987:
971:
964:
947:
944:
939:
936:
929:
928:
927:
922:
906:
902:
899:
892:
889:
881:
878:
862:
858:
855:
848:
845:
837:
836:
835:
833:
823:
821:
817:
813:
811:
805:
792:
789:
786:
783:
779:
776:
773:
772:
771:
768:
766:
762:
757:
751:
741:
739:
738:code profiler
735:
731:
726:
724:
720:
716:
712:
708:
702:
692:
689:
684:
674:
672:
668:
664:
660:
655:
653:
647:
637:
635:
631:
625:
615:
613:
606:
596:
588:
583:
573:
570:
568:
564:
558:
548:
546:
542:
532:
528:
518:
516:
512:
508:
504:
500:
496:
491:
489:
485:
481:
476:
472:
462:
458:
453:
443:
441:
437:
432:
428:
424:
419:
409:
407:
403:
399:
395:
391:
389:
385:
381:
377:
371:
361:
358:
356:
351:
348:
342:
338:
325:
322:
319:
318:
317:
315:
311:
305:
298:Response time
295:
293:
287:
277:
275:
271:
267:
263:
259:
255:
251:
247:
243:
239:
235:
231:
227:
223:
222:response time
219:
215:
205:
202:
196:
186:
184:
180:
174:
172:
167:
163:
158:
149:
147:
143:
139:
135:
125:
124:
120:
115:
113:
105:
102:
101:
100:
89:
85:
82:
78:
75:
72:
67:
63:
62:
60:
56:
53:
49:
46:
45:response time
42:
41:
40:
38:
34:
30:
26:
22:
1419:
1385:
1372:. Retrieved
1368:the original
1358:
1349:
1343:
1334:
1329:Jim Lesurf.
1324:
1316:the original
1311:
1301:
1278:
1272:
1254:
1249:
1242:
1237:
1132:
1124:
1014:machine code
1008:by using an
1005:
1001:
997:
990:code density
925:
882:
838:
831:
829:
822:
818:
814:
807:
769:
753:
737:
733:
727:
723:optimization
704:
690:
686:
656:
649:
627:
608:
594:
585:
571:
560:
538:
530:
492:
477:
474:
459:
455:
433:
429:
425:
421:
402:World War II
392:
373:
359:
344:
307:
289:
286:Availability
280:Availability
238:service time
218:availability
211:
198:
175:
162:CPU designer
159:
155:
132:
122:
118:
117:
112:Arnold Allen
109:
98:
24:
18:
1283:CreateSpace
1147:superscalar
1139:speed-demon
765:scalability
730:source code
630:transistors
527:Scalability
521:Scalability
501:modules or
388:information
376:information
254:scalability
171:transistors
121:performance
1374:2009-01-21
1229:References
782:bottleneck
652:benchmarks
640:Benchmarks
565:, such as
471:Throughput
465:Throughput
246:throughput
179:clock rate
52:throughput
33:efficiency
1259:CiteSeerX
1196:Complete
812:aspects.
446:Bandwidth
382:. By the
242:bandwidth
181:(see the
119:The word
81:bandwidth
21:computing
1437:Category
1409:Archived
1161:See also
1016:from an
734:profiler
665:and the
314:web page
292:downtime
270:speed up
1393:. p. 2.
1208:Speedup
1006:exactly
992:of the
988:). The
412:Latency
400:during
310:disk IO
230:latency
214:metrics
1289:
1261:
926:where
756:system
717:, the
659:SPECfp
632:on an
486:, the
86:Short
43:Short
671:EEMBC
350:IA-32
347:Intel
341:FLOPS
90:time.
79:High
61:(s).
50:High
1287:ISBN
761:load
736:(or
511:ASIC
339:and
268:and
1018:HLL
705:In
513:or
499:FFT
482:or
357:).
274:CPU
185:).
166:CPU
19:In
1439::
1398:^
1333:.
1310:.
1285:.
1281:.
1153:,
1145:,
725:.
673:.
614:.
272:.
264:,
260:,
256:,
252:,
248:,
244:,
240:,
236:,
232:,
228:,
224:,
220:,
173:.
148:.
23:,
1377:.
1337:.
1295:.
1103:C
1100:1
1095:=
1092:I
1064:I
1061:1
1056:=
1053:C
1031:f
1002:N
998:N
972:N
948:t
945:1
940:=
937:P
907:N
903:f
900:I
893:=
890:P
863:f
859:C
856:N
849:=
846:t
832:t
784:.
83:.
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.