mersenneforum.org  

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

Reply
 
Thread Tools
Old 2020-09-26, 16:18   #3323
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

2×3×521 Posts
Default

Quote:
Originally Posted by Neutron3529 View Post
using 3080 is enough for TF
I borrowed a 3080 card and test its performance..
If you have a completed TF run could you please submit the data so I can calculate the values for RTX 3080?
https://www.mersenne.ca/mfaktc.php#benchmark
James Heinrich is online now   Reply With Quote
Old 2020-09-26, 22:23   #3324
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

110318 Posts
Default

Quote:
Originally Posted by Neutron3529 View Post
using 3080 is enough for TF
I borrowed a 3080 card and test its performance..
it could brings 2/3 TF performance of A100 but using only 1/10 price.[/CODE]
sieveprimes=82485 is consistent with a default mfaktc.ini.
Did you try any performance tuning?
kriesel is offline   Reply With Quote
Old 2020-09-27, 00:50   #3325
storm5510
Random Account
 
storm5510's Avatar
 
Aug 2009
U.S.A.

2×811 Posts
Default

Quote:
# Default: GPUSievePrimes=82486
I believe the above is what you are referring to. Sometime in the past, I changed mine to 100,000, although I do not remember doing it. Whether this helps or hinders, I am not sure. I suppose I may have been experimenting. Oh, well.
storm5510 is offline   Reply With Quote
Old 2020-09-27, 09:49   #3326
Neutron3529
 
Neutron3529's Avatar
 
Dec 2018
China

23·5 Posts
Default

Quote:
Originally Posted by James Heinrich View Post
If you have a completed TF run could you please submit the data so I can calculate the values for RTX 3080?
https://www.mersenne.ca/mfaktc.php#benchmark
I'm sorry that I had returned the GPU back and most of the results are not stored since I wasn't asked to do so.


the only saved result is four NF report and what I wrote here.
Neutron3529 is offline   Reply With Quote
Old 2020-09-27, 10:46   #3327
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

2×3×521 Posts
Default

Quote:
Originally Posted by Neutron3529 View Post
I'm sorry that I had returned the GPU back and most of the results are not stored since I wasn't asked to do so. The only saved result is four NF report and what I wrote here.
A NF result is probably what I'm looking for. Ideally I'd want to know the clockspeed the GPU was running at during the run as well, but a completed NF run is a good start.
https://www.mersenne.ca/mfaktc.php#benchmark
James Heinrich is online now   Reply With Quote
Old 2020-09-27, 15:29   #3328
Neutron3529
 
Neutron3529's Avatar
 
Dec 2018
China

23·5 Posts
Default

Quote:
Originally Posted by James Heinrich View Post
A NF result is probably what I'm looking for. Ideally I'd want to know the clockspeed the GPU was running at during the run as well, but a completed NF run is a good start.
https://www.mersenne.ca/mfaktc.php#benchmark
maybe you could using the result I submitted(~4600GHz-d/day) now since I could not provide anymore knowledge of 3080/3090
I'll try my best to buy a 3090 with water-cooling, before that, the post I posted yesterday is all I have now.
Neutron3529 is offline   Reply With Quote
Old 2020-09-27, 16:47   #3329
DrobinsonPE
 
Aug 2020

3916 Posts
Default

I just finished tuning my GTX 1650 Super running mfaktc 0.21 and thought I would share the results.

Code:
mfaktc tuning.
I3-4130T, 8GB RAM, SSD, GTX 1650 Super,  Linux Mint 20.1, mfaktc 0.21
GPU locked to base frequency using - sudo nvidia-smi -lgc 1530 
exp=101214193 bit 75 to 76 and 101214227 bit 74 to 75
Program was initially running with the following settings set by the person who compiled it.
GPUSieveProcessSize=32, GPUSieveSize=2047, GPUSievePrimes=82486, 1194.15GHz-day

Step 1, very GPUSieveProcessSize  with GPUSieveSize=2047
# Default: GPUSieveProcessSize=16 (previous running at 32)
GPUSieveProcessSize=8	1227.10GHz-day
GPUSieveProcessSize=16	1236.01GHz-day	*
GPUSieveProcessSize=24	1235.57GHz-day
GPUSieveProcessSize=32	1194.15GHz-day

Step 2: very GPUSieveSize with GPUSieveProcessSize=16
# Minimum: GPUSieveSize=4
# Maximum: GPUSieveSize=2047
# Default: GPUSieveSize=64 (previous running at 2047)
GPUSieveSize=32		1009.53GHz-day
GPUSieveSize=64		1111.63GHz-day
GPUSieveSize=96		1151.90GHz-day
GPUSieveSize=128	1174.16GHz-day
GPUSieveSize=192	1195.20GHz-day
GPUSieveSize=224	1202.32GHz-day
GPUSieveSize=256	1206.64GHz-day
GPUSieveSize=288	1209.00GHz-day
GPUSieveSize=320	1211.37GHz-day
GPUSieveSize=384	1216.13GHz-day
GPUSieveSize=512	1221.81GHz-day
GPUSieveSize=1024	1231.32GHz-day
GPUSieveSize=2047	1236.01GHz-day	*

Step 3: very GPUSievePrimes with GPUSieveSize=96, GPUSieveProcessSize=32
# Minimum: GPUSievePrimes=54
# Maximum: GPUSievePrimes=1075766
# Default: GPUSievePrimes=82486
GPUSievePrimes=21814	1183.14GHz-day
GPUSievePrimes=67894	1234.00GHz-day	exponent change 101214227 bit 74 to 75
GPUSievePrimes=82486	1236.01GHz-day	*
GPUSievePrimes=99894	1233.10GHz-day
GPUSievePrimes=120374	1227.32GHz-day
DrobinsonPE is offline   Reply With Quote
Old 2020-09-27, 22:47   #3330
storm5510
Random Account
 
storm5510's Avatar
 
Aug 2009
U.S.A.

2×811 Posts
Default

Quote:
Originally Posted by DrobinsonPE View Post
I just finished tuning my GTX 1650 Super running mfaktc 0.21 and thought I would share the results.

Code:
mfaktc tuning.
I3-4130T, 8GB RAM, SSD, GTX 1650 Super,  Linux Mint 20.1, mfaktc 0.21
GPU locked to base frequency using - sudo nvidia-smi -lgc 1530 
exp=101214193 bit 75 to 76 and 101214227 bit 74 to 75
Program was initially running with the following settings set by the person who compiled it.
GPUSieveProcessSize=32, GPUSieveSize=2047, GPUSievePrimes=82486, 1194.15GHz-day

Step 1, very GPUSieveProcessSize  with GPUSieveSize=2047
# Default: GPUSieveProcessSize=16 (previous running at 32)
GPUSieveProcessSize=8    1227.10GHz-day
GPUSieveProcessSize=16    1236.01GHz-day    *
GPUSieveProcessSize=24    1235.57GHz-day
GPUSieveProcessSize=32    1194.15GHz-day

Step 2: very GPUSieveSize with GPUSieveProcessSize=16
# Minimum: GPUSieveSize=4
# Maximum: GPUSieveSize=2047
# Default: GPUSieveSize=64 (previous running at 2047)
GPUSieveSize=32        1009.53GHz-day
GPUSieveSize=64        1111.63GHz-day
GPUSieveSize=96        1151.90GHz-day
GPUSieveSize=128    1174.16GHz-day
GPUSieveSize=192    1195.20GHz-day
GPUSieveSize=224    1202.32GHz-day
GPUSieveSize=256    1206.64GHz-day
GPUSieveSize=288    1209.00GHz-day
GPUSieveSize=320    1211.37GHz-day
GPUSieveSize=384    1216.13GHz-day
GPUSieveSize=512    1221.81GHz-day
GPUSieveSize=1024    1231.32GHz-day
GPUSieveSize=2047    1236.01GHz-day    *

Step 3: very GPUSievePrimes with GPUSieveSize=96, GPUSieveProcessSize=32
# Minimum: GPUSievePrimes=54
# Maximum: GPUSievePrimes=1075766
# Default: GPUSievePrimes=82486
GPUSievePrimes=21814    1183.14GHz-day
GPUSievePrimes=67894    1234.00GHz-day    exponent change 101214227 bit 74 to 75
GPUSievePrimes=82486    1236.01GHz-day    *
GPUSievePrimes=99894    1233.10GHz-day
GPUSievePrimes=120374    1227.32GHz-day

I must be using a config file from a previous version. It says maximum GPUSieveSize is 128? I tried 256 and it accepted it. I wonder what else I may be missing...
storm5510 is offline   Reply With Quote
Old 2020-09-28, 01:04   #3331
Xyzzy
 
Xyzzy's Avatar
 
"Mike"
Aug 2002

22×41×47 Posts
Default

WARNING: Read GPUSieveSize=1024 from mfaktc.ini, using max value (128)
Xyzzy is offline   Reply With Quote
Old 2020-09-28, 03:11   #3332
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

41×113 Posts
Default

Quote:
Originally Posted by Xyzzy View Post
WARNING: Read GPUSieveSize=1024 from mfaktc.ini, using max value (128)
Maybe post 3189's attachment will help (also linked to at 3208).

Last fiddled with by kriesel on 2020-09-28 at 03:44
kriesel is offline   Reply With Quote
Old 2020-09-28, 03:13   #3333
DrobinsonPE
 
Aug 2020

3·19 Posts
Default

Quote:
Originally Posted by storm5510 View Post
I must be using a config file from a previous version. It says maximum GPUSieveSize is 128? I tried 256 and it accepted it. I wonder what else I may be missing...
The version of mfaktc I am using is the linux compiled version I found at this post https://mersenneforum.org/showpost.p...postcount=3208

The mfaktc.ini file was already configured with the settings GPUSieveProcessSize=32, GPUSieveSize=2047, GPUSievePrimes=82486.

Other than gaining about 3% by changing the GPUSieveProcessSize from 32 to 16, the rest of the settings were already at the optimal configuration.

Quote:
Originally Posted by Xyzzy View Post
WARNING: Read GPUSieveSize=1024 from mfaktc.ini, using max value (128)
I did not see any warning but I will go back and check mfaktc.ini and look for it.
DrobinsonPE is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
mfakto: an OpenCL program for Mersenne prefactoring Bdot GPU Computing 1657 2020-10-27 01:23
The P-1 factoring CUDA program firejuggler GPU Computing 752 2020-09-08 16:15
"CUDA runtime version 0.0" when running mfaktc.exe froderik GPU Computing 4 2016-10-30 15:29
World's second-dumbest CUDA program fivemack Programming 112 2015-02-12 22:51
World's dumbest CUDA program? xilman Programming 1 2009-11-16 10:26

All times are UTC. The time now is 10:40.

Thu Oct 29 10:40:21 UTC 2020 up 49 days, 7:51, 1 user, load averages: 1.49, 1.59, 1.60

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.