PDA

View Full Version : Calculating the heat output of a GPU


cV
03-12-04, 04:03 PM
The following is my original Sysopt posting. I decided to share it with all here because we know we're overclockers and could really benefit from knowing the heat output figures of our GPUs to see what thermal solutions need to be designed for them... as well as achieving that maximum overclock :D

I know many of us are wondering about how to calculate the wattage of our GPUs. With no solid numbers released to the public, the public is forced to guess. Well, I decided it'd be time for an EDUCATED guess so that we know what kind of heat these bad boys throw off and design our aftermarket cooling solutions around them.

I may be wrong here, but I think I've found a way to estimate the power consumption of a GPU. It's most likely a little bit off, perhaps far off. All I know is that the power consumption figures seem about right.

I decided it'd be time to figure out just how much heat my R350 puts out at 400MHz. This requires knowing the transistor count (~110 million) and process technology (.15-micron) as well.

The goal is to find the unknowns based on a knowns. To start, I used a reference CPU that I knew the wattage of, a Thoroughbred 2400+ (68.3W Max @ 2.0GHz). I then divided the wattage by the approximate number of transistors (37 million) in the CPU to get watts/transistor.

68.3/37000000
1.8459459459459459459459459459459e-6

This resulted in that number of watts per transistor AT 2GHz operational frequency. To get their frequency @ 400MHz, I had to set up the following equation:

(1.85x10-6 watts/tr)/(2GHz)=(x watts/tr)/(0.4GHz)

answer*.4
7.3837837837837837837837837837838e-7
answer/2
3.6918918918918918918918918918919e-7

this resulted in 3.69x10-7 watts/transistor at 400MHz.

Now, the radeon 9800 GPU does not have 37 million transistors; it has 110 million. So I simply multiplied that value by 110 million to get the total heat output of the GPU @ 130nm:

answer*110000000
40.59

which ends up being 40.59W.

But... the Radeon 9800 GPU is not .13-micron; it is .15-micron. Since the relationship is linear, I set up an equation to find a compensating factor:

.15-.13
_______ = .1333333...
.15

So with this equation we find that on average a .15-micron transistor runs 13.3% hotter than a .13-micron transistor.

Multiply the wattage by 113% and we get the extra heat output factored into the total wattage:

40.59*1.13333333333333333333333333333333333
46.002

which ends up being 46.002 watts.



Now, one final addition. The R350 core does not run at 1.65v, most are set at 1.75v. Again, another ratio must be established, but this is not as simple as the earlier one. Heat and voltage is not a linear relationship; it is an exponential one.

The equation for finding a fudge factor for the additional voltage is as follows:

overvolt^2
_____________
original volt^2

=

(1.75/1.65)^2 = 1.1248852157943067033976124885216

multiplying this value by the wattage at 1.65v and you get a wattage that factors in the additional .1v:

answer*46.002
51.74696969696969696969696969697

So the heat output of our R350 GPU running at 400MHz on a .15-micron process at 1.75v is, in theory, at most 51.75W. People with a sense of this heat output would immediately get alarmed by the anemic size of the R350 heatsink in comparison to the CPU-like thermal output. Undoubtedly, the R350 has been measured to reach nearly 80 degrees celsius under full load.

However, there is no way in hell a heatsink like that would be able to even remotely cool nearly 52W of thermal power. GPUs are deeply pipelined devices and the transistors are almost always under-utilized. And 52W is a theoretical value - the most thermal loading programs have been able to load CPUs is 88% or 7/8 of their maximum power.

Since there is no graphics program dedicated to filling the pipelines to their maximum with predictable values, GPUs can never see even 88% of their theoretical output. Even under "full load" rendering conditions, the number probably hovers between 66% and 75% utilization.

Now these are numbers I've pulled out of nether regions, but they are more close to reality IMHO.

51.74*(2/3)=34.49333333333333333333333333333...
51.74*(3/4)=38.805
arithmetic average:
(34.493+38.805)/2

= an average of 36.65W under "full load" rendering/gaming conditions.

which still does heat up the anemic heatsink quite a bit.


Now, based on those values, let's calculate the degC/W of the stock heatsink. C/W is a measure of how many degrees celsius a cooling apparatus allows a heat source to rise over the ambient (environment) temperature. The lower, the better.

Xbit labs says their R350 heated up to 79.6 degrees Celsius under load. They said this was on a lab bench in the open air, so I'll assume their room temperature was around 22 degrees celsius. This is a ballpark figure, whose basis on a 20MHz clock differential has almost no bearing on the results.

(79.6-22)=57.6 (this is how much the heatsink allowed the core to rise above ambient, ouch)
heat output of R350 @ 400MHz (from my possibly skewed calculations)=36.65 under "gaming" load

57.6 C / 36.65 W -> 1.57 degrees C/W

(Note that a small AMD stock cooler has about .35 degrees C/W in comparison)

No doubt about it, the stock cooler is very poor, but it does keep the GPU below the error-producing heat point. GPUs tend to be more tolerant than CPUs of heat because they run at much lower clock speeds on a given process and are loaded on average far less.

I hope this gives some folks some insight as to how to perform ballpark estimates on GPU thermal output (or output of any chips) when no solid wattage information is released to the public - as for this calculation, I know I'm a bit off, but I think I'm pretty close to reality. :burn: :beer:

TehGoober
03-12-04, 05:10 PM
:eek: :eek: :eek:

Wow. Thanks for all of the information. I definately know where to look if I want to find out what my GPU is running at!

JigPu
03-12-04, 06:42 PM
That's amazing! Though it does explain why my huge heatsink I added to my card was so hot without a fan :eek:

I think we need to get whoever does the HSF compairsons for the front page to test out GPU HSFs. If they test the HS on the die simulator, we should get a descent idea of just how accurate your wattages are.

JigPu

Steven4563
03-12-04, 06:48 PM
wow nice post :D maybe ths could be come a sticky ;)

cV
03-13-04, 10:25 AM
Wow, my first sticky! Thanks, guys.

I'm going to develop this further to see where it heads.

Gautam
03-13-04, 10:30 AM
Wow! Very nice analysis!

This also explains why my 9800np ran at nearly ambient with my MCW-50. ;)

Xplos!ve
03-15-04, 01:09 PM
OMG!! i made up 1 of these for calculating the FPS ur gfx can get, here goes :) (probobly wrong) FSB (the fsb before the cpu is quad pumped etc *very important u do this*) of cpu MULTIPLIED memory or clock speed of gfx (whichever is greatest) MINUS memory or clock speed of gfx (which ever is less) THEN u take the first two numbers and then times it by the lowest number in the rest of the answer ONLY if the first two numbers are less than 50 but if they are higher than 50 times it by the highest single number in the rest of the answer, thn wotever u get divide by two and u have ur FPS :) works for me...... i done 166mhz fsb (333mhz but it is double pumped) x 350mhz memory speed - 250mhz clock speed = 57850 so 57 x 5= 285 divided by 2= 142.5!! and OMFG i get 135 fps on couter-strike! (+/- 10fps or so from ur answer)!!!! can any1 try this out for me 2 see if it is an actual theory? ta wot u will need to do is find out the highest fps u can get then do the sum and see how close u get from ur actual highest fps u can get then post results. thx guys :)
What u will need to do experiment: Computer, Calculator, a Brain and a bit of common sense!!!!

johan851
03-27-04, 02:12 PM
Originally posted by Xplos!ve
OMG!! i made up 1 of these for calculating the FPS ur gfx can get, here goes :) (probobly wrong) FSB (the fsb before the cpu is quad pumped etc *very important u do this*) of cpu MULTIPLIED memory or clock speed of gfx (whichever is greatest) MINUS memory or clock speed of gfx (which ever is less) THEN u take the first two numbers and then times it by the lowest number in the rest of the answer ONLY if the first two numbers are less than 50 but if they are higher than 50 times it by the highest single number in the rest of the answer, thn wotever u get divide by two and u have ur FPS :) works for me...... i done 166mhz fsb (333mhz but it is double pumped) x 350mhz memory speed - 250mhz clock speed = 57850 so 57 x 5= 285 divided by 2= 142.5!! and OMFG i get 135 fps on couter-strike! (+/- 10fps or so from ur answer)!!!! can any1 try this out for me 2 see if it is an actual theory? ta wot u will need to do is find out the highest fps u can get then do the sum and see how close u get from ur actual highest fps u can get then post results. thx guys :)
What u will need to do experiment: Computer, Calculator, a Brain and a bit of common sense!!!!
You're joking...right?

sandman001
03-29-04, 07:13 PM
lol, now do that for morrowind.

Manbot
04-02-04, 04:42 PM
thank you so much! i dont know even how to do those kind of calculations but since i have a R350 running @400mhz i didnt have to!


btw expect to hear from me soon about alcohol, H2O, TR2M3 heatsink, acetylene+O2, copper and PVC! (guess what im doing to my video card... )




btw
awesome job!

twitch579
04-06-04, 07:49 PM
im gona try this, anybody no of any software that tells u the gpu temp? I no with the radeon XT models, it tells u in the ati control panel?

Cyrix_2k
04-16-04, 09:15 PM
Hmm, maybe I should WC my R350 when i get it...

2XS
04-22-04, 01:41 PM
I think we need to get whoever does the HSF compairsons for the front page to test out GPU HSFs. If they test the HS on the die simulator, we should get a descent idea of just how accurate your wattages are.


I'll second that.

DDR-PIII
05-04-04, 09:48 PM
k, now to find a way to figure out the heat output of memory :) such a way to do this haha ?

Dukemurmur
05-16-04, 09:47 PM
WOW this is great so really a GPU at 452 is putting out almost as much Wattage as the stock 2400!!! So when u think about it they deserve there own 2400 cooler.....They also cool much better becuase of the MASSIVE surface area to conduct heat with.

Moto7451
07-16-04, 12:43 PM
My 9100 puts out a measly 10W at stock settings. Knowing that helped me with choosing a fanless heatsink for it (a good ol' 486 sucker to be precise).

PLOBBY
10-09-04, 11:49 PM
wow. Uhh i didnt follow half of that :)....i mmight look at it a little more closely some other time since my 9800 pro doesnt have a temp probe

DDR-PIII
12-30-04, 05:26 PM
how about people complie a list of wattage at speeds ?

fAlCoNNiAn
07-06-05, 12:11 AM
how about people complie a list of wattage at speeds ?
i like that idea.

Awesome read, btw. :)

drunkn.bear
09-20-05, 05:41 PM
nice post, helps for knowing how much power you need and how much you can use

Careface
11-27-05, 08:08 AM
hmm i just worked out my 6200's heat output.

470MHz Core, 1.3VCore, 147mln transistors and 110nm process..

(1.8459459459459459459459459459459e-6*0.470)/2
=4.3379729729729729729729729729708e-7 so
4.33x10^-7/watt/transistor @ 470MHz

ANS*147,000,000
=63.76820270270270270270270270267

0.11-0.13
________ = -0.181818181818181818181
0.11

63.76W*0.82(close enough)
=52.2832W

1.3V^2
________ = 0.62075298438934802571166207529844
1.65V^2

ANS*52.2832W
=32.454952433425160697887970615243W

32.5*(2/3)=21.667
32.5*(3/4)=24.375

(21.667+24.375)/2=23.021

so my overclocked 6200 puts out on average 23.021W of heat? o.0; that doesnt seem right does it? meh, even if it isnt right, i enjoyed doing the math at 2 in the morning lol!

Careface*

swimmer_01
12-12-05, 12:15 PM
too much info for peanut sized brain.
why dont i drink some beer? good idea.
*goes drinks beer, destroys brain cells*

actually, thats very complicated for me. itd be nice to have a table with all the video cards.

largon
10-15-07, 10:05 AM
Just a FYI,
the methods for making the calculations in this thread are completely, utterly wrong.

dfgrghr
12-28-08, 08:02 AM
too much info for peanut sized brain.
why dont i drink some beer? good idea.
*goes drinks beer, destroys brain cells*

actually, thats very complicated for me. itd be nice to have a table with all the video cards.


oh , good ideas ,and i dream to have my own room , and play my own radio, when i have time , and i will play the game , and i think only the game and the music can give me much pleasure