mersenneforum.org  

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

Reply
 
Thread Tools
Old 2021-08-17, 20:05   #1
pontiff
 
Aug 2021
Europe

5 Posts
Default GTX 1660 or Core i9-10900K

Which one would be better for prime numbers hunt? I already have the GTX 1660 video card that I mentioned and I wonder if it's worth building a new rig with the latest and fastest CPU? Or the GPU computing with the hardware that I already have is superior to any new CPU? Thank you.
pontiff is offline   Reply With Quote
Old 2021-08-17, 20:31   #2
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

10110100010112 Posts
Default

Quote:
Originally Posted by pontiff View Post
Which one would be better for prime numbers hunt? I already have the GTX 1660 video card that I mentioned and I wonder if it's worth building a new rig with the latest and fastest CPU? Or the GPU computing with the hardware that I already have is superior to any new CPU? Thank you.
Welcome. GTX 1660 is excellent at TF, far faster than CPUs for the dollar invested or per kilowatt-hour.
CPUs and AMD GPUs are better at other forms of GIMPS computation typically. Radeon VII are CPU beaters on both price/performance and results/watt-hour. Radeon 6xxx are supposedly pretty good too.
There's also cloud computing, some of which is free, so hard to beat on price/performance.
So, it depends what you want to run, and how much. Try some out and see what you enjoy most or least.
See the reference info for lots more background info.
kriesel is offline   Reply With Quote
Old 2021-08-18, 00:11   #3
moebius
 
moebius's Avatar
 
Jul 2009
Germany

22×5×31 Posts
Default

Welcome to the forum,
I have values for the GTX 1650 Super and the GTX 1660 TI for a specific exponent in the gpuOwl-Comparison List, give me a benchmark for the GTX 1660 if you want, to make the sheet more useful.

Last fiddled with by moebius on 2021-08-18 at 00:18
moebius is offline   Reply With Quote
Old 2021-08-19, 17:27   #4
pontiff
 
Aug 2021
Europe

5 Posts
Default

Quote:
Originally Posted by moebius View Post
Welcome to the forum,
I have values for the GTX 1650 Super and the GTX 1660 TI for a specific exponent in the gpuOwl-Comparison List, give me a benchmark for the GTX 1660 if you want, to make the sheet more useful.

I'll gladly do that if you'd tell me how :)
pontiff is offline   Reply With Quote
Old 2021-08-19, 17:29   #5
pontiff
 
Aug 2021
Europe

510 Posts
Default

Quote:
Originally Posted by kriesel View Post
Welcome. GTX 1660 is excellent at TF, far faster than CPUs for the dollar invested or per kilowatt-hour.
CPUs and AMD GPUs are better at other forms of GIMPS computation typically. Radeon VII are CPU beaters on both price/performance and results/watt-hour. Radeon 6xxx are supposedly pretty good too.
There's also cloud computing, some of which is free, so hard to beat on price/performance.
So, it depends what you want to run, and how much. Try some out and see what you enjoy most or least.
See the reference info for lots more background info.

Thank you! I'll stick to my current video card then, but will probably invest in something better along the road. I tried cloud computing too and enjoying a free month, not bad, price-wise :)
pontiff is offline   Reply With Quote
Old 2021-08-20, 19:59   #6
moebius
 
moebius's Avatar
 
Jul 2009
Germany

22×5×31 Posts
Default

Quote:
Originally Posted by pontiff View Post
I'll gladly do that if you'd tell me how :)
Download the Windows version

Fastest Windows Version (with Proofs)
https://mersenneforum.org/showpost.p...94&postcount=8

Then run the first 300000 itterations with gpuOwl for the exponent 77936867 at the Windows command Line.

Type cmd in the windows search field and switch to the gpuOwl directory with cd path f.e. cd C:\Users\username\gpuowl-v6.11-364-g36f4e2a.
Execute the binary at the command line with gpuowl-win -iters 300000 -prp 77936867


or you use the Linux binary version (needs further installation effort, read: How to set up for running gpuOwl under Ubuntu (and other Linux) with OpenCL

Linux binary version Ubuntu 18.04
https://mersenneforum.org/showpost.p...1&postcount=40

gpuowl -iters 300000 -prp 77936867

post the result from the gpuowl.log file in this thread https://mersenneforum.org/showthread...779#post585779

Last fiddled with by moebius on 2021-08-20 at 20:22
moebius is offline   Reply With Quote
Old 2021-08-21, 07:31   #7
pontiff
 
Aug 2021
Europe

5 Posts
Default

Done, log posted!


You're saying that gpuowl-win v6.11-364-g36f4e2a is faster than the latest gpuowl-v7.2-69-g23c14a1 ?

Last fiddled with by pontiff on 2021-08-21 at 07:41 Reason: typo
pontiff is offline   Reply With Quote
Old 2021-08-21, 12:24   #8
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

29×199 Posts
Default

Quote:
Originally Posted by pontiff View Post
You're saying that gpuowl-win v6.11-364-g36f4e2a is faster than the latest gpuowl-v7.2-69-g23c14a1 ?
Yes for most fft lengths. See the last couple benchmark attachments here in the reference info.
kriesel is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
GTX 1660 Ti axn GPU Computing 14 2019-05-20 04:45
Core i5 2500K vs Core i7 2600K (Linear algebra phase) em99010pepe Hardware 0 2011-11-11 15:18
Dual Core to Quad Core Upgrade Rodrigo Hardware 6 2010-11-29 18:48
exclude single core from quad core cpu for gimps jippie Information & Answers 7 2009-12-14 22:04
Optimising work for Intel Core 2 Duo or Quad Core S485122 Software 0 2007-05-13 09:15

All times are UTC. The time now is 17:25.


Sat Oct 16 17:25:40 UTC 2021 up 85 days, 11:54, 1 user, load averages: 1.35, 1.32, 1.41

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, 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.