Advanced search

Message boards : Number crunching : e179s298_e7s25p1f282-PABLO_p27_wild_0_sj403_IDP-1-2-RND8896_1

Author Message
Zalster
Avatar
Send message
Joined: 26 Feb 14
Posts: 77
Credit: 2,218,612,423
RAC: 6,258,188
Level
Phe
Scientific publications
watwatwat
Message 49461 - Posted: 11 May 2018 | 22:35:22 UTC
Last modified: 11 May 2018 | 22:52:54 UTC

So I noticed that this type of work unit raised the average temperature of my card 11-12C higher than most other work units. I also notice that the power usage of the card increased by about 30% and my fan rate jump accordingly due to my fan curve I have.

My question is, are other seeing this increase on their cards as well? Since then, the card has worked on other work units and is back to normal range, power, fan etc compared to other. I've attempted to look at others that have processed similar work units. Only a few are Windows based(Linux machines don't report GPU temps) and I can't tell if the temps they are seeing is within their normal range.

Anyone else seeing the marked increase in Temps and Power for these work units?

Edit...

So I've gone back and looked at other results on my machine and it seems to repeat on different cards crunching similar work units. All of the Pablo...wild....sj4
____________

kksplace
Send message
Joined: 4 Mar 18
Posts: 12
Credit: 5,192,825
RAC: 0
Level
Ser
Scientific publications
wat
Message 49462 - Posted: 11 May 2018 | 23:22:35 UTC - in response to Message 49461.

Yes, I have seen something similar. Being a noobie, I thought I was a little crazy. It is not as dramatic a change as you note, but my GPU temp is 3 to 4 degrees higher and the related CPU usage is higher as well when crunching those work units. It is because of those WUs I reduced by one my cores for crunching. However, with all the new WUs over the last week, I can easily go back to 5 total cores crunching and my GPU temp stays 'reasonable' when not running the WUs you note.

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 110
Credit: 84,864,713
RAC: 626,043
Level
Thr
Scientific publications
wat
Message 49463 - Posted: 12 May 2018 | 0:29:17 UTC - in response to Message 49461.

I haven't had any luck with the latest work on my Windows10 system. One had an error last reported in 2014 that only had one comment. The last one had a segfault. The system previously had no issues with the work.

The Linux system reported a couple of valid tasks last week but hasn't got to the recent ones from today.

Betting Slip
Send message
Joined: 5 Jan 09
Posts: 666
Credit: 2,498,095,550
RAC: 294
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 49464 - Posted: 12 May 2018 | 10:10:09 UTC - in response to Message 49461.
Last modified: 12 May 2018 | 10:13:04 UTC

So I noticed that this type of work unit raised the average temperature of my card 11-12C higher than most other work units. I also notice that the power usage of the card increased by about 30% and my fan rate jump accordingly due to my fan curve I have.

My question is, are other seeing this increase on their cards as well? Since then, the card has worked on other work units and is back to normal range, power, fan etc compared to other. I've attempted to look at others that have processed similar work units. Only a few are Windows based(Linux machines don't report GPU temps) and I can't tell if the temps they are seeing is within their normal range.

Anyone else seeing the marked increase in Temps and Power for these work units?

Edit...

So I've gone back and looked at other results on my machine and it seems to repeat on different cards crunching similar work units. All of the Pablo...wild....sj4


No great mystery, these WU's are just more computationally intensive than others and you also get more credit for them.
____________
Radio Caroline, the world's most famous offshore pirate radio station.
Great music since April 1964. Support Radio Caroline Team -
Radio Caroline

Erich56
Send message
Joined: 1 Jan 15
Posts: 476
Credit: 2,375,476,677
RAC: 1,938,643
Level
Phe
Scientific publications
watwatwatwat
Message 49467 - Posted: 12 May 2018 | 13:46:10 UTC - in response to Message 49464.

No great mystery, these WU's are just more computationally intensive than others and you also get more credit for them.

that's exactly it.
Same thing happens here (except that here the effect is the other way round: due to my fixed GPU temperature setting in the Nvidia Inspector, the GPU clock drops by some extent with these type of tasks).

Zalster
Avatar
Send message
Joined: 26 Feb 14
Posts: 77
Credit: 2,218,612,423
RAC: 6,258,188
Level
Phe
Scientific publications
watwatwat
Message 49476 - Posted: 15 May 2018 | 3:31:03 UTC - in response to Message 49467.

Just looked at some of the results listed in the performance section. These work units definitely test the limits of the cards. Single card systems see a increase in temp as do multi-card systems but since they have spare room for air flow, the temp increase isn't a significant (they do go up but don't hit the Tmax before throttling occurs)

The few multi card systems I've seen are edging right up to the thermal max (usually 84C)


____________

Post to thread

Message boards : Number crunching : e179s298_e7s25p1f282-PABLO_p27_wild_0_sj403_IDP-1-2-RND8896_1