mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Hardware > GPU Computing

Reply
 
Thread Tools
Old 2016-02-19, 06:00   #12
airsquirrels
 
airsquirrels's Avatar
 
"David"
Jul 2015
Ohio

11·47 Posts
Default

Can you post the fft bench results from both 6.5 and 7.5?

My testing has all been choosing exponents of 2048K and 4096K to stay in the highest performance range of the cards. Other power combinations may have been tuned differently between 6.5 and 7.5
airsquirrels is offline   Reply With Quote
Old 2016-02-19, 06:30   #13
Brain
 
Brain's Avatar
 
Dec 2009
Peine, Germany

331 Posts
Default cufftbench

Quick runs with 6.5 / 7.5, may need more samples

6.5:
CUDALucas2.05.1-CUDA6.5-Windows-x64.exe -cufftbench 4096 8192 2
Code:
Device              GeForce GTX TITAN
Compatibility       3.5
clockRate (MHz)     875
memClockRate (MHz)  2600

  fft    max exp  ms/iter
 4096   75846319   3.0762
 4320   79902611   3.8299
 4374   80879779   3.8564
 4500   83158811   3.9610
 4536   83809729   3.9791
 5184   95507747   4.0854
 5488  100984691   4.5723
 5600  103000823   5.1283
 5832  107174381   5.1384
 6000  110194363   5.3512
 6125  112440191   5.4362
 6250  114685037   5.4624
 6272  115080019   5.4955
 6400  117377567   5.5779
 6480  118813021   5.6794
 6561  120266023   5.7351
 6750  123654943   5.8944
 7776  142017539   5.9808
 8000  146019329   6.2703
 8192  149447533   6.3158
7.5:
CUDALucas2.05.1-CUDA7.5-Windows-x64.exe -cufftbench 4096 8192 2
Code:
Device              GeForce GTX TITAN
Compatibility       3.5
clockRate (MHz)     875
memClockRate (MHz)  2600

  fft    max exp  ms/iter
 4096   75846319   3.0423
 4320   79902611   3.9058
 4374   80879779   4.0908
 4536   83809729   4.1059
 4608   85111207   4.1071
 5184   95507747   4.1222
 5488  100984691   4.7986
 5600  103000823   5.0615
 5832  107174381   5.4100
 6048  111056879   5.4940
 6144  112781477   5.5065
 6250  114685037   5.7932
 6400  117377567   5.8289
 6480  118813021   5.8704
 6561  120266023   6.1004
 6912  126558077   6.2039
 7776  142017539   6.3066
 8192  149447533   6.4384
Both runs under same conditions.
Attached Thumbnails
Click image for larger version

Name:	Performance.PNG
Views:	85
Size:	22.7 KB
ID:	13917  
Brain is offline   Reply With Quote
Old 2016-02-19, 15:53   #14
airsquirrels
 
airsquirrels's Avatar
 
"David"
Jul 2015
Ohio

11·47 Posts
Default

Quote:
Originally Posted by Brain View Post
Quick runs with 6.5 / 7.5, may need more samples

6.5:
CUDALucas2.05.1-CUDA6.5-Windows-x64.exe -cufftbench 4096 8192 2
Code:
Device              GeForce GTX TITAN
Compatibility       3.5
clockRate (MHz)     875
memClockRate (MHz)  2600

  fft    max exp  ms/iter
 4096   75846319   3.0762
 4320   79902611   3.8299
 4374   80879779   3.8564
 4500   83158811   3.9610
 4536   83809729   3.9791
 5184   95507747   4.0854
 5488  100984691   4.5723
 5600  103000823   5.1283
 5832  107174381   5.1384
 6000  110194363   5.3512
 6125  112440191   5.4362
 6250  114685037   5.4624
 6272  115080019   5.4955
 6400  117377567   5.5779
 6480  118813021   5.6794
 6561  120266023   5.7351
 6750  123654943   5.8944
 7776  142017539   5.9808
 8000  146019329   6.2703
 8192  149447533   6.3158
7.5:
CUDALucas2.05.1-CUDA7.5-Windows-x64.exe -cufftbench 4096 8192 2
Code:
Device              GeForce GTX TITAN
Compatibility       3.5
clockRate (MHz)     875
memClockRate (MHz)  2600

  fft    max exp  ms/iter
 4096   75846319   3.0423
 4320   79902611   3.9058
 4374   80879779   4.0908
 4536   83809729   4.1059
 4608   85111207   4.1071
 5184   95507747   4.1222
 5488  100984691   4.7986
 5600  103000823   5.0615
 5832  107174381   5.4100
 6048  111056879   5.4940
 6144  112781477   5.5065
 6250  114685037   5.7932
 6400  117377567   5.8289
 6480  118813021   5.8704
 6561  120266023   6.1004
 6912  126558077   6.2039
 7776  142017539   6.3066
 8192  149447533   6.4384
Both runs under same conditions.
That supports my experience, the power of 2 FFT sizes I use seem to have improved in performance although other sizes had their performance reduced.
airsquirrels is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
CUDALucas (a.k.a. MaclucasFFTW/CUDA 2.3/CUFFTW) msft GPU Computing 2817 2020-07-03 19:43
Don't DC/LL them with CudaLucas LaurV Data 131 2017-05-02 18:41
CUDALucas gives all-zero residues fivemack GPU Computing 4 2016-07-21 15:49
Performance of cuda-ecm on newer hardware? fivemack GMP-ECM 14 2015-02-12 20:10
cuFFT on multiple GPUs HHfromG GPU Computing 2 2014-05-11 19:57

All times are UTC. The time now is 18:31.

Wed Aug 12 18:31:14 UTC 2020 up 26 days, 14:18, 0 users, load averages: 3.86, 2.93, 2.59

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.