Advanced search

Message boards : Graphics cards (GPUs) : GPUGRID is a hog on current: 34 amps

Author Message
Profile JStateson
Avatar
Send message
Joined: 31 Oct 08
Posts: 186
Credit: 3,331,546,800
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 6322 - Posted: 2 Feb 2009 | 19:22:20 UTC

Latest GPU-Z calculates current, at least for the gtx280. I am showing 27 amps for SETI CUDA -vs- 34 amps for GRID CUDA

Searching for little green men seems greener on the environment.

Profile Paul D. Buck
Send message
Joined: 9 Jun 08
Posts: 1050
Credit: 37,321,185
RAC: 0
Level
Val
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 6328 - Posted: 2 Feb 2009 | 21:43:00 UTC - in response to Message 6322.

Latest GPU-Z calculates current, at least for the gtx280. I am showing 27 amps for SETI CUDA -vs- 34 amps for GRID CUDA

Searching for little green men seems greener on the environment.


Except if you have CUDA it will also use the GPU at the same speed and draw ...

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2705
Credit: 1,311,122,549
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 6331 - Posted: 2 Feb 2009 | 22:43:41 UTC

How's the voltage of your GT200 chip? 1.12V? That's the highest for my 9800GTX+, so yous should be along these lines too. Assuming 1.12 V and 34 A means your card would be drawing 38 W. A figure suspiciously low, so I'm not sure if it actually tells us anything.

Due to the low voltage CPUs have been in the range of 70 - 100 amps since some time.. so amps alone are only relevant for the power circuit designer ;)

MrS
____________
Scanning for our furry friends since Jan 2002

Profile JStateson
Avatar
Send message
Joined: 31 Oct 08
Posts: 186
Credit: 3,331,546,800
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 6334 - Posted: 3 Feb 2009 | 1:50:33 UTC - in response to Message 6331.

How's the voltage of your GT200 chip? 1.12V? That's the highest for my 9800GTX+, so yous should be along these lines too. Assuming 1.12 V and 34 A means your card would be drawing 38 W. A figure suspiciously low, so I'm not sure if it actually tells us anything.

Due to the low voltage CPUs have been in the range of 70 - 100 amps since some time.. so amps alone are only relevant for the power circuit designer ;)

MrS


Yea - you are correct. I didnt even see that 1.12 volt as I failed to scroll the display up. According to this thread at the gpuz forum, that is indeed the gpu core voltage and the 38 watts is what the gpu core pulls. The rest of the power (about 70 watts or so) is consumed by the memory. I have a 9800gtx+ on another system but it seems not to have a current sensor.

Profile Paul D. Buck
Send message
Joined: 9 Jun 08
Posts: 1050
Credit: 37,321,185
RAC: 0
Level
Val
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 6336 - Posted: 3 Feb 2009 | 4:41:47 UTC - in response to Message 6322.

Latest GPU-Z calculates current, at least for the gtx280. I am showing 27 amps for SETI CUDA -vs- 34 amps for GRID CUDA

Searching for little green men seems greener on the environment.


I see I had earlier mis-read what you said...

But this is also a little bit of apples and oranges comparison too ... the tasks for SaH are fairly small data wise and so would not necessarily place the same demands on the memory or the bandwidth of the memory system while computations are in progress which could account for the differences in power draw.

To my mind it is also a question of what is of more importance to us ... finding little green men, or finding cures for diseases? Which would prove to be more useful to mankind in the near term?

If the point is to keep the GPU draw low we should be turning the machines off when we are not doing work and forget this BOINC stuff ... :)

I guess that might be an interesting side experiment this week... if I move my watt meter to the Q9300 I can watch it while I rip off some SaH tasks and then later recheck it when I put it back on GPU Grid ... and see what the difference is, if any, between the draw for the two projects ...

Thamir Ghaslan
Send message
Joined: 26 Aug 08
Posts: 55
Credit: 1,475,857
RAC: 0
Level
Ala
Scientific publications
watwatwat
Message 6337 - Posted: 3 Feb 2009 | 8:33:29 UTC - in response to Message 6322.
Last modified: 3 Feb 2009 | 8:35:31 UTC

Latest GPU-Z calculates current, at least for the gtx280. I am showing 27 amps for SETI CUDA -vs- 34 amps for GRID CUDA

Searching for little green men seems greener on the environment.


Consider your self lucky, did'nt know about the new GPU-Z release until I've read your post, upgraded GPU-Z and now I'm reading between 50 to 55 amps on GPU grid.

This is a stock GTX-280.

Voltage: 1.1875 V
Seti shows between 40 - 45 amps.

Idle is 17-18 amps.

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2705
Credit: 1,311,122,549
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 6340 - Posted: 3 Feb 2009 | 8:54:55 UTC

To put things into perspective.. would any of you mind to log the current while running 3D Mark 2006 or Vantage (DX10)? I'd expect a surprisingly high number ;)

And I guess the lower current draw under seti just means that it's using the GPu less efficiently.

MrS
____________
Scanning for our furry friends since Jan 2002

Profile Edboard
Avatar
Send message
Joined: 24 Sep 08
Posts: 72
Credit: 12,410,275
RAC: 0
Level
Pro
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 6456 - Posted: 5 Feb 2009 | 20:44:22 UTC
Last modified: 5 Feb 2009 | 20:45:22 UTC

I get:

gtx280 (16% OC, stock mem):

-seti 36 amp.
-gpugrid 41 amp.
-folding 70 amp. (with the harder units)

vdd = 1.1875V

Post to thread

Message boards : Graphics cards (GPUs) : GPUGRID is a hog on current: 34 amps

//