PrimeGrid
Please visit donation page to help the project cover running costs for this month

Toggle Menu

Join PrimeGrid

Returning Participants

Community

Leader Boards

Results

Other

drummers-lowrise

Advanced search

Message boards : General discussion : Thermal Stress?, Easy fix

Author Message
Profile Tyler Yarger
Send message
Joined: 29 Apr 21
Posts: 2
ID: 1365301
Credit: 19,421,233
RAC: 0
ESP LLR Silver: Earned 100,000 credits (130,780)PPS LLR Amethyst: Earned 1,000,000 credits (1,297,242)PPS Sieve Jade: Earned 10,000,000 credits (13,837,955)AP 26/27 Gold: Earned 500,000 credits (517,504)GFN Amethyst: Earned 1,000,000 credits (1,253,960)WW (retired) Ruby: Earned 2,000,000 credits (2,376,000)
Message 150445 - Posted: 17 May 2021 | 21:38:11 UTC
Last modified: 17 May 2021 | 21:39:00 UTC

Hello World,

I am new here and wanted to share my experience with thermal issues that might help other novices like myself down the road.

For the record I am running
Win10/AMD Ryzen 5 3600/GTX 1660 on BONIC.

I had a lot of thermal issues trying to run both CPU and GPU on Ryzen balance power plan. I found that by going to Control Panel -> Power Plan advance options you can set maximum processor power to 99% (I did 95%). In doing so you prevent the factory CPU boost from activating, which is my case is jumping from 3.6GHz to 4 GHz. Since my CPU always runs under 3.6GHz now its voltage stays at 1V and my CPU rarely goes above 73C. I personally feel that AMD boost isnt worth all the added thermal stress. I'm trying to run a marathon not a sprint. Before it had hit 92C and I had to kill the task, now it caps at about 73C (Running 12 PPS-Mega tasks and GFN-17 running 100% CPU time).

I figure with AMD Boost I would average around 3.7 GHz since boost isnt always running (maybe someone has the real answer, just a guess) and at 95% power I run 3.4 GHz. So thats about a 10% performance cut for a 20C thermal drop.

https://imgur.com/a/MVs2MzI

Profile j.sheridanProject donor
Volunteer tester
Send message
Joined: 21 Mar 11
Posts: 817
ID: 91622
Credit: 2,686,055,085
RAC: 0
Discovered 2 mega primesFound 4 primes in the 2021 Tour de PrimesFound 2 mega primes in the 2021 Tour de Primes321 LLR Jade: Earned 10,000,000 credits (10,011,712)Cullen LLR Jade: Earned 10,000,000 credits (10,025,728)ESP LLR Jade: Earned 10,000,000 credits (10,176,490)Generalized Cullen/Woodall LLR Jade: Earned 10,000,000 credits (10,020,580)PPS LLR Sapphire: Earned 20,000,000 credits (36,093,877)PSP LLR Jade: Earned 10,000,000 credits (10,055,994)SoB LLR Emerald: Earned 50,000,000 credits (67,758,583)SR5 LLR Jade: Earned 10,000,000 credits (10,885,368)SGS LLR Jade: Earned 10,000,000 credits (10,007,143)TRP LLR Jade: Earned 10,000,000 credits (10,009,531)Woodall LLR Jade: Earned 10,000,000 credits (10,039,527)321 Sieve (suspended) Sapphire: Earned 20,000,000 credits (20,019,388)Cullen/Woodall Sieve Double Silver: Earned 200,000,000 credits (265,102,350)PPS Sieve Double Gold: Earned 500,000,000 credits (547,992,812)TRP Sieve (suspended) Turquoise: Earned 5,000,000 credits (5,801,812)AP 26/27 Double Bronze: Earned 100,000,000 credits (101,256,935)GFN Double Gold: Earned 500,000,000 credits (547,439,679)WW (retired) Double Amethyst: Earned 1,000,000,000 credits (1,003,356,000)
Message 150446 - Posted: 18 May 2021 | 7:49:43 UTC - in response to Message 150445.

Far better to switch the more aggressive boosting off in bios. You'll find it will still boost to nearly 4GHz but without all the added heat.

Profile Tyler Yarger
Send message
Joined: 29 Apr 21
Posts: 2
ID: 1365301
Credit: 19,421,233
RAC: 0
ESP LLR Silver: Earned 100,000 credits (130,780)PPS LLR Amethyst: Earned 1,000,000 credits (1,297,242)PPS Sieve Jade: Earned 10,000,000 credits (13,837,955)AP 26/27 Gold: Earned 500,000 credits (517,504)GFN Amethyst: Earned 1,000,000 credits (1,253,960)WW (retired) Ruby: Earned 2,000,000 credits (2,376,000)
Message 150455 - Posted: 19 May 2021 | 0:28:55 UTC - in response to Message 150446.

Thanks I was kind of hoping for a better solution. I'll give it a try over the weekend with my buddy and see how it goes.



I also tried different types of multi-threading on PPS-LLR Mega and found that running 1 tasks per thread and leaving 1 thread opened yielded the best results so far (100 min per task, 11/12 threads used), compared to running 12 tasks on 12 threads (120 minutes per task). Putting more than 1 thread to a task hasn't helped for that project. If anyone has any other suggestions of optimizing PPS-LLR I'd love to here them. I want to try leaving 2 threads/1core unassigned and see if that helps even more.

I dont have much experience in this field and idk how an AMD processor differs in performance to a lot of the information I read in the forums which I assume is mostly Intel based. If anyone has some general pointers for AMD Ryzen, please share them here.


____________

Profile GrebulonerProject donor
Volunteer tester
Avatar
Send message
Joined: 2 Nov 09
Posts: 495
ID: 49572
Credit: 3,166,203,939
RAC: 3,772,976
Discovered 5 mega primesFound 2 primes in the 2018 Tour de PrimesFound 4 primes in the 2019 Tour de PrimesFound 3 primes in the 2020 Tour de PrimesFound 1 mega prime in the 2020 Tour de PrimesFound 4 primes in the 2021 Tour de PrimesFound 7 primes in the 2022 Tour de PrimesFound 1 mega prime in the 2022 Tour de PrimesFound 10 primes in the 2023 Tour de PrimesFound 2 mega primes in the 2023 Tour de Primes321 LLR Emerald: Earned 50,000,000 credits (50,304,710)Cullen LLR Sapphire: Earned 20,000,000 credits (22,096,858)ESP LLR Sapphire: Earned 20,000,000 credits (20,453,074)Generalized Cullen/Woodall LLR Emerald: Earned 50,000,000 credits (50,327,156)PPS LLR Emerald: Earned 50,000,000 credits (75,487,921)PSP LLR Sapphire: Earned 20,000,000 credits (23,670,769)SoB LLR Emerald: Earned 50,000,000 credits (50,873,282)SR5 LLR Sapphire: Earned 20,000,000 credits (28,681,842)SGS LLR Sapphire: Earned 20,000,000 credits (31,592,082)TRP LLR Sapphire: Earned 20,000,000 credits (24,211,899)Woodall LLR Sapphire: Earned 20,000,000 credits (24,238,438)321 Sieve (suspended) Emerald: Earned 50,000,000 credits (55,630,889)Cullen/Woodall Sieve Double Bronze: Earned 100,000,000 credits (100,934,315)Generalized Cullen/Woodall Sieve (suspended) Emerald: Earned 50,000,000 credits (56,046,594)PPS Sieve Double Gold: Earned 500,000,000 credits (521,014,891)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Turquoise: Earned 5,000,000 credits (9,468,384)TRP Sieve (suspended) Jade: Earned 10,000,000 credits (10,076,645)AP 26/27 Double Gold: Earned 500,000,000 credits (584,220,621)GFN Double Silver: Earned 200,000,000 credits (470,765,709)WW (retired) Double Gold: Earned 500,000,000 credits (830,928,000)PSA Double Bronze: Earned 100,000,000 credits (126,200,096)
Message 150456 - Posted: 19 May 2021 | 4:26:49 UTC - in response to Message 150455.

Thanks I was kind of hoping for a better solution. I'll give it a try over the weekend with my buddy and see how it goes.



I also tried different types of multi-threading on PPS-LLR Mega and found that running 1 tasks per thread and leaving 1 thread opened yielded the best results so far (100 min per task, 11/12 threads used), compared to running 12 tasks on 12 threads (120 minutes per task). Putting more than 1 thread to a task hasn't helped for that project. If anyone has any other suggestions of optimizing PPS-LLR I'd love to here them. I want to try leaving 2 threads/1core unassigned and see if that helps even more.

I dont have much experience in this field and idk how an AMD processor differs in performance to a lot of the information I read in the forums which I assume is mostly Intel based. If anyone has some general pointers for AMD Ryzen, please share them here.



For best optimization of LLR projects in general, you want to run no more than 6 threads or 6 tasks of 1 thread each (50% CPU utilization in the Boinc preferences). You will also find this reduces the power consumption and you can increase the clocks, again, if you'd like. The 3600X is a 6 core processor, and for just about all CPUs, LLR runs best when only using the physical cores.

Intel CPUs are discussed most around here because for a long time, they held a significant advantage over AMD on Primegrid work and consequently, that's what most of us have. Now that, aside from a very small case (AVX-512 in certain high-end desktop/server CPUs), AMD has battled back and taken the crown and LLR2 kind of makes being fastest irrelevant, CPU brand choice is more flexible and you'll see more AMD in CPU optimizing discussion (like this thread).

Larger projects than PPS (GCW, 321, SoB, etc) are where you will see benefits running 1 task of 6 cores or 2 tasks of 3 cores each.
____________
Eating more cheese on Thursdays.

Profile j.sheridanProject donor
Volunteer tester
Send message
Joined: 21 Mar 11
Posts: 817
ID: 91622
Credit: 2,686,055,085
RAC: 0
Discovered 2 mega primesFound 4 primes in the 2021 Tour de PrimesFound 2 mega primes in the 2021 Tour de Primes321 LLR Jade: Earned 10,000,000 credits (10,011,712)Cullen LLR Jade: Earned 10,000,000 credits (10,025,728)ESP LLR Jade: Earned 10,000,000 credits (10,176,490)Generalized Cullen/Woodall LLR Jade: Earned 10,000,000 credits (10,020,580)PPS LLR Sapphire: Earned 20,000,000 credits (36,093,877)PSP LLR Jade: Earned 10,000,000 credits (10,055,994)SoB LLR Emerald: Earned 50,000,000 credits (67,758,583)SR5 LLR Jade: Earned 10,000,000 credits (10,885,368)SGS LLR Jade: Earned 10,000,000 credits (10,007,143)TRP LLR Jade: Earned 10,000,000 credits (10,009,531)Woodall LLR Jade: Earned 10,000,000 credits (10,039,527)321 Sieve (suspended) Sapphire: Earned 20,000,000 credits (20,019,388)Cullen/Woodall Sieve Double Silver: Earned 200,000,000 credits (265,102,350)PPS Sieve Double Gold: Earned 500,000,000 credits (547,992,812)TRP Sieve (suspended) Turquoise: Earned 5,000,000 credits (5,801,812)AP 26/27 Double Bronze: Earned 100,000,000 credits (101,256,935)GFN Double Gold: Earned 500,000,000 credits (547,439,679)WW (retired) Double Amethyst: Earned 1,000,000,000 credits (1,003,356,000)
Message 150465 - Posted: 21 May 2021 | 19:20:48 UTC - in response to Message 150455.

Thanks I was kind of hoping for a better solution. I'll give it a try over the weekend with my buddy and see how it goes.



I also tried different types of multi-threading on PPS-LLR Mega and found that running 1 tasks per thread and leaving 1 thread opened yielded the best results so far (100 min per task, 11/12 threads used), compared to running 12 tasks on 12 threads (120 minutes per task). Putting more than 1 thread to a task hasn't helped for that project. If anyone has any other suggestions of optimizing PPS-LLR I'd love to here them. I want to try leaving 2 threads/1core unassigned and see if that helps even more.

I dont have much experience in this field and idk how an AMD processor differs in performance to a lot of the information I read in the forums which I assume is mostly Intel based. If anyone has some general pointers for AMD Ryzen, please share them here.


you should run around 45mins/task if you run 6 at a time. As Grebuloner says, you have to think of the 3600x as 2 3-core processors (each group of 3 is called a CCX) for tasks up to around 2M fft size and then a 6-core over that.

Because of the processor architecure it helps to lock tasks to the same cores as windows likes to move them around which is relatively slow when it moves something from one ccx to the other.

Message boards : General discussion : Thermal Stress?, Easy fix

[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2023 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 2.13, 1.50, 1.39
Generated 5 Jun 2023 | 10:11:20 UTC