The HD 5970 Clubhouse

[erocker]

http://www.shrani.si/f/2I/Ld/2rZG7sWq/zzt.png

VDDC @ 130c.is this normal or should i look into this?

Too hot, your card is throttling.

 

[t_ski]

OK, I rolled back to 10.5 and now I can get all four cores to load under FurMark. Here are the VDDC temps. Left is water, right is air:

920 core, 1.2v
1300 memory, default v
Ambiant temps around 19C

 

[mtosev]

what should I do about my temp in furmark? no OC temps are only 10C lower

 

[t_ski]

What are your ambiant temps?

 

[mtosev]

22.5c

 

[t_ski]

Definitely seems high. How much higher than ambient are all the other temps in your case?

 

[mtosev]

idle temps

 

[Volkszorn88]

Everything seems good. Nice temps.

 

[t_ski]

Agreed, all the other temps look fine. The next thing I would suggest is to open GPUz with the sensors, then play some games on the system for a hour or so. When you're done, go back to the GPUz screen and see what the VDDC temps are then.

You have to remember: FurMark loads a video card in a way that no game ever will. While that's good for checking stability, etc., it's not good for establishing what real-world loads would be like.

 

[Black Panther]

3dmark11 scores

They are feeble for the 5970 - I attribute it being low cpu score since I'm on an E8400.

If you run this test please quote your detailed scores here (mine appear lower left of screenshot) - no need to put a screenshot of them, it's just so I can see where I got low whether it was in cpu or gpu

Thanks

 

[Vellari]

If you run this test please quote your detailed scores here (mine appear lower left of screenshot) - no need to put a screenshot of them, it's just so I can see where I got low whether it was in cpu or gpu

Most likely the cpu (being only dual-core). My 3DM11 score is (atm) P6084. Our graphics scores are very close (you: 6783, me: 6825), but physics and combined are significantly higher on my setup (you: phys 3407, comb 2986; me phys 5119, comb 3973).

However, my setup being quad-gpu this could also be the result of me having double the amount of Directcompute gpu:s. This shouldn't be a factor in physics score though, since that test to my knowledge is pure cpu.

Interestingly enough, having quad-gpu doesn't seem to have much of a benefit in comparison with single a 5970, atleast as far as my setup goes. Anyone else notice the same? Crossfire doesn't appear to be kicking in properly even though the 3DM11 sysinfo says 4 cards. Driver issue, perhaps...

 

[t_ski]

I think one quad core is powerful enough to push one card, but to really push two cards you needs some serious muscle (more cores + more speed).

 

[mtosev]

I should consider this normal?

HL2 gameplay

 

[Vellari]

I should consider this normal? http://gpuz.techpowerup.com/10/12/09/6nw.png
HL2 gameplay

Looks about right, if using stock cooling.

EDIT:
Also, here are my new scores after a moderate overclock (stock cooling, stock voltages)
CPU: 3.2GHz -> 3.5GHz
GPU: 725/1000 -> 780/1050

Score improvements:
3DM06: 20709 -> 21381
3DMV: 23530 -> 25821
3DM11: N/A -> P6084
Unigine Heaven: 1943 -> 2068
Stone Giant (average FPS): 119 -> 133

 

[brandonwh64]

Ok i was wondering how much a used 5970 would go for nowadays?

 

[fullinfusion]

this is what I got last night.

 

[Black Panther]

this is what I got last night.

http://img.techpowerup.org/101209/Captureaaaaa.png

http://img.techpowerup.org/101209/Captureaaa850.png

Comparing your scores to mine you got:

Graphics test 1 : +3fps
Graphics test 2 : +3fps
Graphics test 3 : +3fps
Graphics test 4 : +1fps
Physics test : +13fps
Combined test : +15fps

Meaning I'm failing mostly in the physics and combined tests. Am I right thinking it's my CPU which is lacking in that aspect? Probably the slightly lower fps in the other tests are cpu-related as well?

Edit: or perhaps it's the 8800gt which you're using for physics which is boosing those scores?

 

[2DividedbyZero]

Edit: or perhaps it's the 8800gt which you're using for physics which is boosing those scores?

3DM11 does NOT use nvidia card for physics 3DM11 does NOT use physx either

 

[Black Panther]

3DM11 does NOT use nvidia card for physics 3DM11 does NOT use physx either

Okay don't hurt your head so infusion's better scores for physics and combined are not because he's got an 8800gt card for physics but because he's got a much better cpu than mine, correct?

Edit: why does 3dmark11 run a physics test if it doesn't use physics

 

[2DividedbyZero]

3DM11 uses the CPU for physics yes

 

[2DividedbyZero]

Okay don't hurt your head so infusion's better scores for physics and combined are not because he's got an 8800gt card for physics but because he's got a much better cpu than mine, correct?

Edit: why does 3dmark11 run a physics test if it doesn't use physics

i think you're confusing physx and physics


http://3dmark.com/3dm11/61632

 

[Black Panther]

i think you're confusing physx and physics

Yup

 

[fullinfusion]

Comparing your scores to mine you got:

Graphics test 1 : +3fps
Graphics test 2 : +3fps
Graphics test 3 : +3fps
Graphics test 4 : +1fps
Physics test : +13fps
Combined test : +15fps

Meaning I'm failing mostly in the physics and combined tests. Am I right thinking it's my CPU which is lacking in that aspect? Probably the slightly lower fps in the other tests are cpu-related as well?

Edit: or perhaps it's the 8800gt which you're using for physics which is boosing those scores?

yes the Physics is strictly ran by the cpu cores and not my pissX 8800GT card.
Your only running 2 cores, so that's whats hurting the scores.

I'm thinking that the fps im getting are slightly higher then yours because I clocked the card up a bit. Have you tried 800-1100MHz clocks yet?
You should be able to run that at stock volts.

 

[Black Panther]

yes the Physics is strictly ran by the cpu cores and not my pissX 8800GT card.
Your only running 2 cores, so that's whats hurting the scores.

I'm thinking that the fps im getting are slightly higher then yours because I clocked the card up a bit. Have you tried 800-1100MHz clocks yet?
You should be able to run that at stock volts.

Lol overclocking my 5970 is just a waste on my processor. Card ran stable at 800/1100 but 3dmark11 increased by just 13 points

 

[fullinfusion]

Lol overclocking my 5970 is just a waste on my processor. Card ran stable at 800/1100 but 3dmark11 increased by just 13 points

http://img.techpowerup.org/101209/3dmark 800-1100.png

I'm going to see if I can mirror those marks. I'll down core to 2 with the same clocks cpu-gpu