Message boards :
Number crunching :
ubuntu Install
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5
Author | Message |
---|---|
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Guess we should be running the SSE41 app on the BLC05 cpu tasks. keith@Darksider:~/Downloads/KWSN-Bench-Linux-MBv7_v2.01.08$ ./benchmark KWSN-Linux-MBbench v2.1.08 Running on Darksider at Fri 12 Jan 2018 10:44:33 PM UTC ---------------------------------------------------------------- Starting benchmark run... ---------------------------------------------------------------- Listing wu-file(s) in /testWUs : blc05_2bit_guppi_57976_78659_HIP61801_0041.32033.0.22.45.190.vlar.wu Listing executable(s) in /APPS : MBv8_8.05r3345_avx_linux64 Listing executable in /REF_APPS : MBv8_8.22r3711_sse41_x86_64-pc-linux-gnu ---------------------------------------------------------------- Current WU: blc05_2bit_guppi_57976_78659_HIP61801_0041.32033.0.22.45.190.vlar.wu ---------------------------------------------------------------- Running default app with command :... MBv8_8.22r3711_sse41_x86_64-pc-linux-gnu Elapsed Time: ....................... 1909 seconds ---------------------------------------------------------------- Running app with command : .......... MBv8_8.05r3345_avx_linux64 Elapsed Time : ...................... 1899 seconds Speed compared to default : ......... 100 % ----------------- Comparing results Result : Strongly similar, Q= 99.96% ---------------------------------------------------------------- Done with blc05_2bit_guppi_57976_78659_HIP61801_0041.32033.0.22.45.190.vlar.wu ==================================================================== Done with Benchmark run! Removing temporary files! keith@Darksider:~/Downloads/KWSN-Bench-Linux-MBv7_v2.01.08$ ./benchmark KWSN-Linux-MBbench v2.1.08 Running on Darksider at Fri 12 Jan 2018 11:58:00 PM UTC ---------------------------------------------------------------- Starting benchmark run... ---------------------------------------------------------------- Listing wu-file(s) in /testWUs : blc05_2bit_guppi_57976_78659_HIP61801_0041.32033.0.22.45.190.vlar.wu Listing executable(s) in /APPS : MBv8_8.05r3345_avx_linux64 Listing executable in /REF_APPS : MBv8_8.22r3712_avx2_x86_64-pc-linux-gnu ---------------------------------------------------------------- Current WU: blc05_2bit_guppi_57976_78659_HIP61801_0041.32033.0.22.45.190.vlar.wu ---------------------------------------------------------------- Running default app with command :... MBv8_8.22r3712_avx2_x86_64-pc-linux-gnu Elapsed Time: ....................... 2185 seconds ---------------------------------------------------------------- Running app with command : .......... MBv8_8.05r3345_avx_linux64 Elapsed Time : ...................... 1772 seconds Speed compared to default : ......... 123 % ----------------- Comparing results Result : Strongly similar, Q= 99.95% ---------------------------------------------------------------- Done with blc05_2bit_guppi_57976_78659_HIP61801_0041.32033.0.22.45.190.vlar.wu ==================================================================== Done with Benchmark run! Removing temporary files! keith@Darksider:~/Downloads/KWSN-Bench-Linux-MBv7_v2.01.08$ Its a wash between the r3345 AVX app and the r3711 SSE41 apps. But the r3345 AVX app is 23% faster than the r3712 AVX2 app. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
RueiKe Send message Joined: 14 Feb 16 Posts: 492 Credit: 378,512,430 RAC: 785 |
From my experience it depends on the tasks and the CPU load. I have definitely found that to be the case. My approach now is to free up 1 core on my machine for the benchmark runs and make sure it doesn't stop BOINCmgr from continuing to run tasks. I am still concerned that results may be influence by what app type is running on the rest of the cores (will tesing AVX be invluence by other cores running SSE42?) GitHub: Ricks-Lab Instagram: ricks_labs |
RueiKe Send message Joined: 14 Feb 16 Posts: 492 Credit: 378,512,430 RAC: 785 |
Guess we should be running the SSE41 app on the BLC05 cpu tasks. I have confirmed that my actual performance had improved when I switched from AVX to SSE42 as my benchmarks indicated. Probably a good idea to re-validate optimization when WU characteristics change. GitHub: Ricks-Lab Instagram: ricks_labs |
RueiKe Send message Joined: 14 Feb 16 Posts: 492 Credit: 378,512,430 RAC: 785 |
Hi RueiKe! That was so long ago that I can not remember the details of what I was doing. But definitely all of the issues I was having were fixed by Tbar's latest Linux build of BOINCmgr. GitHub: Ricks-Lab Instagram: ricks_labs |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
Guess we should be running the SSE41 app on the BLC05 cpu tasks I made few runs in real life with the SSE4.1 and the times almost match the AVX2 times. About 48 min to crunch a blc05 WU Not see the differences you are talking about. Maybe the big gain is a Ryzen thing only. Can anyone with more knowledge check my host tell me if i'm wrong? I can't see any real difference between them in my Intel host. |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Thanks for the posts, Rick. I guess I should look at that SSE4.2 app. From my reading, there seemed to be VERY little difference between the capabilities of SSE 4.1 and SSE4.2, only a couple more instruction calls. Wikipedia being my source. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Juan, it very well could be a "Ryzen thing" .... don't know. Haven't a Intel processor to check things out with. I'm seeing about a 2 minute improvement so far over the AVX2 app with the SSE41 app. Have no idea if the improvement will stick with Arecibo resends or whatever flavor of BLC tasks the splitters cough out next. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13736 Credit: 208,696,464 RAC: 304 |
Juan, it very well could be a "Ryzen thing" .... don't know. Haven't a Intel processor to check things out with. AMD have had a lot more time to work on their SSE4.x implementation, compared to AVX and the much more recent AVX2 instruction sets. Grant Darwin NT |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Looks like its time to benchmark the SSE42 app on the BLC05 tasks Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
The r3306 SSE42 app was 3% faster than the r3711 SSE41 app on my Ryzen 1800X. Not sure if the difference is worth changing over to. The advantage of any one app can and probably will change depending on the work mix coming from the project. YMMV. Juan tested the SSE42 app against the SSE41 app on his Intel and the SSE42 app was 10% slower. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Mike Send message Joined: 17 Feb 01 Posts: 34258 Credit: 79,922,639 RAC: 80 |
The r3306 SSE42 app was 3% faster than the r3711 SSE41 app on my Ryzen 1800X. Not sure if the difference is worth changing over to. The advantage of any one app can and probably will change depending on the work mix coming from the project. YMMV. Thats correct Keith. There is never just one specific app being fastest on my Lunatics tests because i always test different AR`s and type of WU`s. So long its just BLC it doesn`t really matter if you use SSE4.1 or SSE4.2 the difference is just a few seconds. With each crime and every kindness we birth our future. |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Juan has tested both SSE4 apps today against the AVX and AVX2 and the SSE4x apps do particularly bad against the AVX2 on his Intel cpu. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
©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.