Questions and Answers :
Windows :
Benchmark result problems
Message board moderation
Author | Message |
---|---|
Chris Send message Joined: 27 Jun 04 Posts: 5 Credit: 1,112 RAC: 0 |
win XP pro sp1, dual amd athlon mp2400+ 512mb ddr Problem is with boinc 3.19 --- - 2004-07-12 10:45:15 - Number of CPUs: 2 --- - 2004-07-12 10:45:15 - 2426 double precision MIPS (Whetstone) per CPU --- - 2004-07-12 10:45:15 - 1376 integer MIPS (Dhrystone) per CPU --- - 2004-07-12 10:45:15 - Finished CPU benchmarks what is going on? why is mips so low? it conistantly occurs with version 3.19. If i use version 3.18, most of the time i get reasonable results --- - 2004-07-12 10:39:51 - Benchmark results: --- - 2004-07-12 10:39:51 - Number of CPUs: 2 --- - 2004-07-12 10:39:51 - 2457 double precision MIPS (Whetstone) per CPU --- - 2004-07-12 10:39:51 - 4321 integer MIPS (Dhrystone) per CPU --- - 2004-07-12 10:39:51 - Finished CPU benchmarks --- - 2004-07-12 10:33:51 - Running CPU benchmarks --- - 2004-07-12 10:34:53 - Benchmark results: --- - 2004-07-12 10:34:53 - Number of CPUs: 2 --- - 2004-07-12 10:34:53 - 2449 double precision MIPS (Whetstone) per CPU --- - 2004-07-12 10:34:53 - 4294 integer MIPS (Dhrystone) per CPU --- - 2004-07-12 10:34:53 - Finished CPU benchmarks but even this version occasionally yields crazy results. Is this a program fault or my hardware? |
Heffed Send message Joined: 19 Mar 02 Posts: 1856 Credit: 40,736 RAC: 0 |
Don't use 3.18. Always use the most recent version. The main reason being, the benchmarks have changed from 3.18 to 3.19. <a> [/url] |
Thierry Van Driessche Send message Joined: 20 Aug 02 Posts: 3083 Credit: 150,096 RAC: 0 |
I did a lot of trials on benchmarks during the Beta Test as well as with the new version we are using now. The values I got with v3.19 are not that far away from each other. Your low integer MIPS (Dhrystone) number is definitely something strange. I do not know what could be the possible reason for this. Hardware problem or processes that are running in the background maybe? Try to do another benchmark test to see what result you obtain. Greetings from Belgium. |
Jaaku Send message Joined: 29 Oct 02 Posts: 494 Credit: 346,224 RAC: 0 |
|
Chris Send message Joined: 27 Jun 04 Posts: 5 Credit: 1,112 RAC: 0 |
> Don't use 3.18. Always use the most recent version. > > The main reason being, the benchmarks have changed from 3.18 to 3.19. > > <a> [/url] > yes i figured, but i was curious as to the difference. Why should 3.19 give me shocking results like this? i tried sis a several times and it gives normal results SiSoftware Sandra Benchmark Results Dhrystone ALU : 15104MIPS Whetstone FPU : 6281MFLOPS ok, not the best program... but why should BOINC be giving crazy benchmark results? back on 3.19 tried it without anything but system processes running, and boinc respectively. Any other software i can try for free that will test the Dhrystone, and Whetstone on my dual rig to rule out hardware problems? Also just ran benchmark using boinc 3.20.... same problems as with 3.19 =( |
Chris Send message Joined: 27 Jun 04 Posts: 5 Credit: 1,112 RAC: 0 |
Ok keeping you up to date. just levelled my system and installed windows server 2003, nothing but boinc 3.20 installed. after first running the benchmark when boinc first runs --- - 2004-07-12 14:08:34 - Benchmark results: --- - 2004-07-12 14:08:34 - Number of CPUs: 2 --- - 2004-07-12 14:08:34 - 2449 double precision MIPS (Whetstone) per CPU --- - 2004-07-12 14:08:34 - 4331 integer MIPS (Dhrystone) per CPU --- - 2004-07-12 14:08:34 - Finished CPU benchmarks more like. but.... running benchmark again straight after --- - 2004-07-12 14:10:38 - Benchmark results: --- - 2004-07-12 14:10:38 - Number of CPUs: 2 --- - 2004-07-12 14:10:38 - 2450 double precision MIPS (Whetstone) per CPU --- - 2004-07-12 14:10:38 - 1361 integer MIPS (Dhrystone) per CPU --- - 2004-07-12 14:10:38 - Finished CPU benchmarks =( now... if i exit boinc, start it again and run the benchmark i get --- - 2004-07-12 14:12:08 - Benchmark results: --- - 2004-07-12 14:12:08 - Number of CPUs: 2 --- - 2004-07-12 14:12:08 - 2452 double precision MIPS (Whetstone) per CPU --- - 2004-07-12 14:12:08 - 4306 integer MIPS (Dhrystone) per CPU --- - 2004-07-12 14:12:08 - Finished CPU benchmarks exit and start again... --- - 2004-07-12 14:13:38 - Benchmark results: --- - 2004-07-12 14:13:38 - Number of CPUs: 2 --- - 2004-07-12 14:13:38 - 2451 double precision MIPS (Whetstone) per CPU --- - 2004-07-12 14:13:38 - 4308 integer MIPS (Dhrystone) per CPU --- - 2004-07-12 14:13:38 - Finished CPU benchmarks exit and start again... --- - 2004-07-12 14:16:09 - Benchmark results: --- - 2004-07-12 14:16:09 - Number of CPUs: 2 --- - 2004-07-12 14:16:09 - 2451 double precision MIPS (Whetstone) per CPU --- - 2004-07-12 14:16:09 - 4308 integer MIPS (Dhrystone) per CPU --- - 2004-07-12 14:16:09 - Finished CPU benchmarks consistent results. if i run boinc benchmark any time at all on winxppro i get really low dhrystone results. Its like after running a benchmark or running any processing once.... it clogs up and throws the benchmarks off. This is going into areas i don't fully understand so if some one can help me out it would be appreciated. |
Thierry Van Driessche Send message Joined: 20 Aug 02 Posts: 3083 Credit: 150,096 RAC: 0 |
>if i run boinc benchmark any time at all on winxppro i get > really low dhrystone results. Its like after running a benchmark or running > any processing once.... it clogs up and throws the benchmarks off. This is > going into areas i don't fully understand so if some one can help me out it > would be appreciated. Benchmark results with Boinc are very dependent on OS. The problem of low benchmark with XP has been improved during the different versions of Boinc. Some best benchmark results with v3.05 --- - 2004-05-30 11:51:42 - Benchmark results: 354 million floating point ops/sec --- - 2004-05-30 11:51:42 - Benchmark results: 436 million integer ops/sec Same PC but using v3.19 --- - 2004-07-12 13:32:33 - 1580 double precision MIPS (Whetstone) per CPU --- - 2004-07-12 13:32:33 - 1823 integer MIPS (Dhrystone) per CPU These results are still at a too low level compared to other OS as well as to other benchmark systems. I hope in a newer version it will be improved once again. Greetings from Belgium. |
Chris Send message Joined: 27 Jun 04 Posts: 5 Credit: 1,112 RAC: 0 |
So it is more likely just the buggy program than my hardware? |
belgix Send message Joined: 7 Apr 00 Posts: 40 Credit: 1,951,701 RAC: 0 |
Guys, no great improvement in BOINC 3.20. Works fine for my P4 laptop but still over estimate speed of my AMD-K6 II desktop by a factor 2x. |
Sven Lapczynski Send message Joined: 25 Oct 99 Posts: 46 Credit: 78,668 RAC: 0 |
|
Chris Send message Joined: 27 Jun 04 Posts: 5 Credit: 1,112 RAC: 0 |
ye, but at least your measured integer speed isn't less than the floating point speed! lol this is just crazy |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.