The HD 5970 Clubhouse

[Black Panther]

woooooohoooooo i think i've don it. but gpu-z showes two differt versions of gpu1 and gpu2.

I just realised mine do as well

So either we're both right, or we both made the same mistake.....

I doubt it's a mistake though since my card has functioned impreccably since the bios flash..

I'll pm fullinfusion to give us his screenies...

 

[t_ski]

All four of mine have the same number:

 

[Black Panther]

crap...

 

[jbkappirossi]

is it bad that gpu 1 2 dont have the same version numbers?

because the you repley't earlier is the same as the one in the tech power data base.

 

[Black Panther]

Well, I dunno. I'm a noob at flashing, it's the first time I did it and since my card functioned well after the flashing (it corrected the 2D instability I had before) I was convinced I had done it right.

Now I ran furmark and monitored two instances of gpu-z and here's what I found.....

Hope there's nothing terribly wrong here...

 

[jbkappirossi]

me to yikessss, so there is really no headroom to overvolt, shittt!!!!!

 

[t_ski]

Interested in some aftermarket cooling?

http://www.arctic-cooling.com/catalog/product_info.php?cPath=2_&mID=545

 

[jbkappirossi]

Interested in some aftermarket cooling?

http://www.arctic-cooling.com/catalog/product_info.php?cPath=2_&mID=545

yes thats what i was thinking two, but it seem that only the core temps are getting better.
Because the VRM's still whit Accelero are to high. http://www.overclock.net/ati-cooling/754712-accelero-xtreme-5970-review.html

So i was thinking to reduce the VRM bottleneck, is to replace the refference cooler for the Accelero and combine it whit this VRM cooler, but i'm not sure it wil fit propperly, or even to make it fit. http://www.thermalright.com/products/index.php?act=data&cat_id=29&id=105

 

[Black Panther]

Well, considering I'm not overclocking, in my case at least I expected everything to function fine out of the box, without needing to void warranty by installing a 3rd party cooler... Then on the other hand I only realised this by running a benchmark and gpu-z. Card always performed flawlessly, even with the old bios which used to give trouble when I wasn't gaming

I wonder if any other 5970 owners could perhaps post a screenshot of gpu-z temps while running furmark so we could compare?

 

[TheMailMan78]

Well, considering I'm not overclocking, in my case at least I expected everything to function fine out of the box, without needing to void warranty by installing a 3rd party cooler... Then on the other hand I only realised this by running a benchmark and gpu-z. Card always performed flawlessly, even with the old bios which used to give trouble when I wasn't gaming

I wonder if any other 5970 owners could perhaps post a screenshot of gpu-z temps while running furmark so we could compare?

I would PM some people. This is an old thread and some of the guys may not come by here much anymore.

 

[fullinfusion]

Sorry Im just getting back to you now BP, but work takes me from home most of the week

My numbers are also the same. BUuuuuuuuuT WAIT!
Something is goofy here!
One time I check the bios number is the same but the next time it's not.

I wouldnt worrie about it. It seems like GPUZ is buggy atm with these new bios es.
Id open gpu-z up and select the bottom drop down for gpu#2 and the numbers would be the same..
Next time Id select the top box then thats when the number changed.

Im running a PissX card and even dissabled it and sitll the same thing.

I'm sure Wizard will fix it soon enough if he reads this post.

 

[Black Panther]

Ummm well form me it shows 012.020.000.019.000000 for gpu 1 and 012.020.014.000000 for gpu 2.....

But that isn't worrying me, since everything seems fine and I'm no longer getting lockups/artifacting/driver shutdowns @ 157/300.

What's worrying me now is that darn vddc temperature, screenshots of which me and jbkappirossi posted above...

When you have the time would you mind taking a screenshot of furmark running alongside 2 instances of gpu-z one showing gpu1 and the other gpu2? Pleeasee?
(Make sure they're not both showing the same gpu and that you select from the drop down menu)

 

[Black Panther]

Take the screenshot during the furmark with gpu loads near 100% and the clocks maxed out (your memory shows at 300Mhz in the screenshot)

 

[fullinfusion]

Here you go BP. I wouldn't worrie about the mem temps
I been seeing those temps sense I've owned this gpu.

This run is at stock setting with fan at 50%

I run 1000MHz - 1300MHz with no worries even @ 120c

 

[fullinfusion]

800-1100MHz

 

[Black Panther]

So I guess I shouldn't worry, your vddc temperature in the first screenie shows 103 degrees...
Though I honestly hope it's a flaw with gpu-z, I don't think any part of any pcb would be enjoying such temperatures

 

[cadaveca]

VRMs should be fine to 125c.

If they get too hot, they WILL cause throttle.

As the VRMs get hotter, thier ability to provide stable clean power is diminished, however, the overcurrent protection mechanisms within the card should prevent and possible damage.


The heatsinks themselves are quite restrictive, and as such, cooling capacity can be greatly reduced over time with dust buildup. I am not sure if your warranty covers cooler replacement, but I'd basically suggest giving it a good blast of air or something, as this may improve things a bit.

Keep an eye on clocks, and watch for artifacting...if clocks are stable, and there is no artifacting, there is no cause for concern, but if either is present, I;d suggest you RAM that card ASAP.

I haven't been talking about the 5-series and temp issues for months for no reason. It wasn't just complaining...it also serves basis to serve as a precedent for such situations.

 

[t_ski]

Or go to water

 

[Black Panther]

Or go to water

Fullinfusion's on water and he's getting vddc temps over 100 degrees centigrade too..

 

[fullinfusion]

Fullinfusion's on water and he's getting vddc temps over 100 degrees centigrade too..

Yes Fullinfusion is on water but not on my 5970
I have never cooled any of my gpu's using water.
I wouldn't wanna screw with warranty
just water cooling this 1090T

Can you say 2c

 

[fullinfusion]

VRMs should be fine to 125c.

If they get too hot, they WILL cause throttle.

As the VRMs get hotter, thier ability to provide stable clean power is diminished, however, the overcurrent protection mechanisms within the card should prevent and possible damage.


The heatsinks themselves are quite restrictive, and as such, cooling capacity can be greatly reduced over time with dust buildup. I am not sure if your warranty covers cooler replacement, but I'd basically suggest giving it a good blast of air or something, as this may improve things a bit.

Keep an eye on clocks, and watch for artifacting...if clocks are stable, and there is no artifacting, there is no cause for concern, but if either is present, I;d suggest you RAM that card ASAP.

I haven't been talking about the 5-series and temp issues for months for no reason. It wasn't just complaining...it also serves basis to serve as a precedent for such situations.

Thanks Cad for making this issue clear to all of us.

I clock the crap outta this gpu and still no problems with artifacts.

I think you pointed it out to me a long time ago that 100c was no problem for concern.

Thanks again bro

 

[t_ski]

Well, I was going to try and post up some VDDC temps under water for you, but I noticed that I am only running two cores under FurMark MultiGPU. Is there a trick to get them all (4) to run?

 

[fullinfusion]

Well, I was going to try and post up some VDDC temps under water for you, but I noticed that I am only running two cores under FurMark MultiGPU. Is there a trick to get them all (4) to run?

Not that i know of..... just run 5min of the stress and lets see what your water can keep the Vregs cool at.

 

[t_ski]

Want to make sure I am posting the right card.

 

[mtosev]

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