mersenneforum.org  

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

Reply
 
Thread Tools
Old 2019-03-21, 13:30   #3125
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

4,583 Posts
Default

Quote:
Originally Posted by nomead View Post
The four I've seen are
VRel = Reliability. Indicating perf is limited by reliability voltage.(lower than the next one)
VOp = Operating. Indicating perf is limited by max operating voltage.
Pwr = Power. Indicating perf is limited by total power limit.
Thrm = Thermal. Indicating perf is limited by temperature limit.
You can adjust the power limit a bit (nvidia-smi or MSI Afterburner or something equivalent) and also the temperature target, but there is a hard thermal limit at 88C (for RTX20xx, used to be higher). Running the GPU that hot may affect the longevity of the card already, so I wouldn't recommend it. Just keep it cool and other improvements will follow.
Very good. For completeness, there's also
None
Idle (which can occur at least intermittently while a gpu application is running)
combinations of the 4 nomead listed, eg Pwr, VRel, or Pwr, Thrm, VRel
Gpu temperature limits vary considerably with model, and the trend seems to be downward over the years. (GTX480 is much higher than GTX1080) https://www.mersenneforum.org/showpo...11&postcount=2
kriesel is online now   Reply With Quote
Old 2019-03-21, 13:54   #3126
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

4,583 Posts
Default

Quote:
Originally Posted by LaurV View Post
TechPowerUp was not able in about 5 years since I opened a ticket with them to fix a couple of bugs, to keep up with some new cards, and to make the interface scalable. I think they don't know how to do it.
I've been reporting the AMD gpu and stable driver/Win7 host OS/remote desktop usage/certain gpu sensor readings go MIA upon onset of RDP use since 2.7.0 or earlier; it's now at 2.18.0, at least 11 releases have had the problem. Also reported it as an AMD driver issue to AMD. No fix, no response, beyond autoresponders, ever. TechPowerUp seems interested only in releasing support for new gpu models.
Code:
GPU Core Clock, GPU Memory Clock, GPU Temperature sensor values are not available if a Windows Remote Desktop Session is employed to view the desktop. A running instance of GPU-Z V2.8.0 displaying these will drop to zero frequency and no temperature value upon initiation of an RDP session. it is necessary to go to the local console and relaunch GPU-Z to get a look at these sensor parameters. This occurs on all 3 RX550s installed (2 @4GB, 1 @2GB)

reported again for GPU-Z 2.8.0 3/15/18, validation code zyydx, as bug report with gmail address
https://www.techpowerup.com/gpuz/details/zyydx
Apparently there is a long tradition of such. https://www.techpowerup.com/forums/t...-7650m.251680/
https://www.techpowerup.com/forums/t...mation.238288/
(From early 2009!) https://www.techpowerup.com/forums/t...-hd4870.80813/
(0.1.0, 2007!) https://www.techpowerup.com/forums/t...gs-only.43617/
kriesel is online now   Reply With Quote
Old 2019-03-21, 14:38   #3127
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
Jun 2011
Thailand

5·29·61 Posts
Default

Indeed. And the people on their forums jumped on my throat like "what do I expect from a free tool?", etc. But that is another story (the discussions are still there).
LaurV is offline   Reply With Quote
Old 2019-03-21, 16:47   #3128
petrw1
1976 Toyota Corona years forever!
 
petrw1's Avatar
 
"Wayne"
Nov 2006
Saskatchewan, Canada

23·7·79 Posts
Default

Quote:
Originally Posted by LaurV View Post
Hey Sid. Or Andy?

You can get more juice from that card, for a lower contribution to the global warming*, just install Nvidia Precision and play a bit with clocks and voltages.

Also, the RoG skin of GPU-Z is not the best choice

* I want a Tesla one day; does that count?

I upgraded to RoG thinking it was better for an ASUS MB; fixed.

This one??

https://www.evga.com/precisionx1/

Or???

https://www.evga.com/precisionxoc/

Last fiddled with by petrw1 on 2019-03-21 at 16:51
petrw1 is offline   Reply With Quote
Old 2019-03-22, 10:23   #3129
nomead
 
nomead's Avatar
 
"Sam Laur"
Dec 2018
Turku, Finland

23×41 Posts
Default

Quote:
Originally Posted by nomead View Post
Then temperatures. Ambient +21C, GPU is running at +68C while being power limited to 240W, CPU is +78C running stage 1, and +67C running stage 2.
I practice what I preach. This finally nudged me to change the case fans, 1 front and 1 rear, from 1200 rpm models to 1900 rpm. I already had them on the shelf for a couple weeks, but it is often difficult for me to get even simple tasks started. Some would call me lazy, my doctor calls it depression.

Same ambient temperature as before, all other things being equal except the case fans, CPU is running P-1 stage 1 now at +73C, and the GPU, running mfaktc, is down to +63C. Also the GPU clocked itself a couple notches higher to 1875 MHz, while power usage is still capped at the same 240W.
nomead is offline   Reply With Quote
Old 2019-03-22, 11:04   #3130
SELROC
 

E9C16 Posts
Default

Quote:
Originally Posted by nomead View Post
I practice what I preach. This finally nudged me to change the case fans, 1 front and 1 rear, from 1200 rpm models to 1900 rpm. I already had them on the shelf for a couple weeks, but it is often difficult for me to get even simple tasks started. Some would call me lazy, my doctor calls it depression.

Same ambient temperature as before, all other things being equal except the case fans, CPU is running P-1 stage 1 now at +73C, and the GPU, running mfaktc, is down to +63C. Also the GPU clocked itself a couple notches higher to 1875 MHz, while power usage is still capped at the same 240W.

In general Trial Factoring heats less than PRP because the GPU RAM runs at a lower frequency.
  Reply With Quote
Old 2019-03-22, 12:08   #3131
nomead
 
nomead's Avatar
 
"Sam Laur"
Dec 2018
Turku, Finland

23×41 Posts
Default

Quote:
Originally Posted by SELROC View Post
In general Trial Factoring heats less than PRP because the GPU RAM runs at a lower frequency.
It's quite the opposite on NVidia because of the lousy crippled 1:32 FP64 execution. And, okay, LL (CUDALucas) not PRP, since I think there is no CUDA PRP available, and the NVidia OpenCL implementation is semi-broken and slow. But CUDALucas only takes about 140W on RTX2080 when running. GPU usage says 100% but this is just because all available FP64 capability is in use. And the memory bus is not even close to saturation.
nomead is offline   Reply With Quote
Old 2019-03-22, 12:14   #3132
SELROC
 

7,691 Posts
Default

Quote:
Originally Posted by nomead View Post
It's quite the opposite on NVidia because of the lousy crippled 1:32 FP64 execution. And, okay, LL (CUDALucas) not PRP, since I think there is no CUDA PRP available, and the NVidia OpenCL implementation is semi-broken and slow. But CUDALucas only takes about 140W on RTX2080 when running. GPU usage says 100% but this is just because all available FP64 capability is in use. And the memory bus is not even close to saturation.

On AMD GPU Radeon RX580 it is:
- (mfakto) Trial factoring: gpu Cores: 2000MHz, gpu RAM: 300MHz
- (gpuowl) PRP: gpu Cores: 2000MHz, gpu RAM: 2000MHz


the power is around 145W.

Last fiddled with by SELROC on 2019-03-22 at 12:15
  Reply With Quote
Old 2019-03-28, 00:54   #3133
petrw1
1976 Toyota Corona years forever!
 
petrw1's Avatar
 
"Wayne"
Nov 2006
Saskatchewan, Canada

23·7·79 Posts
Default

I've discovered that when my wife's account is logged on (Windows 10); whether it is active or not the GPU TF performance drops by about 8%.

That is, if I switch user (and I might see no active applications) and logout of her session the performance immediately improves by about 8%.
petrw1 is offline   Reply With Quote
Old 2019-03-29, 03:54   #3134
Rodrigo
 
Rodrigo's Avatar
 
Jun 2010
Pennsylvania

39F16 Posts
Default

Quote:
Originally Posted by petrw1 View Post
I've discovered that when my wife's account is logged on (Windows 10); whether it is active or not the GPU TF performance drops by about 8%.

That is, if I switch user (and I might see no active applications) and logout of her session the performance immediately improves by about 8%.
Wonder what it is about being logged in that could account for that difference in GPU performance.
Rodrigo is offline   Reply With Quote
Old 2019-03-29, 20:30   #3135
Thecmaster
 
"Arvid Björklin"
Apr 2016
Pitea, Sweden

73 Posts
Default

It could be a different energy saving option. Your wife's profile are probably on balanced and your are on high performance.

My first guess. If this is not the case it could be that your wife's profile could run an application you don't run as auto start.
/Arvid
Thecmaster 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 1656 2020-10-13 14:21
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 23:47.

Thu Oct 22 23:47:40 UTC 2020 up 42 days, 20:58, 0 users, load averages: 1.80, 1.66, 1.65

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.