[H] users 7970 Overclock Results thread

Another newbie question... I have an Asus with a 86.% ASIC score that runs at 1.025v. I've read that the max ovp is .125v - which means my card can only handle 1.150v right? if this is the case has anyone had any luck with gpu tweak or trixx applying less than 1.170v? I can't get the overclock stable above 1075/1525 with CCC and any aftermarket software does not seem to let me apply less than 1.170v as a minimum, which crashes the card. Any ideas?
 
Just got a Diamond reference card, already had Afterburner installed, what's a safe o/c and do I use CCC or Afterburner to set the o/c settings? Won't they conflict with each other?
 
So, just got my EK full coverage block. Probably won't have a chance to install this along with the rest of my new WC stuff until the weekend.

This might seem like a silly question, but are the thermal pads for the RAM and such not supposed to be sticky at all?
 
YOUR OLD RIG GOT HIT BY LIGHTNING?!

Yeah, it was weird. It didn't come through the normal 'power' side, it decided to hit the coax going into the cable modem and scatter across the network. It was one of the most amazing things I've seen (and so, so costly).

Of course, it was the excuse needed to upgrade from a Phenom X4/4870 512 rig to the new one.
 
Just got a Diamond reference card, already had Afterburner installed, what's a safe o/c and do I use CCC or Afterburner to set the o/c settings? Won't they conflict with each other?
Running any overclock with stock voltages is safe. See if you can run the CCC max stably. MSI Afterburner is still kind of buggy, so if you aren't going for insane clocks, sticking to CCC limits will net you a ton of performance for little effort.
So, just got my EK full coverage block. Probably won't have a chance to install this along with the rest of my new WC stuff until the weekend.

This might seem like a silly question, but are the thermal pads for the RAM and such not supposed to be sticky at all?
It depends on the matrix used. As long as they stay between the RAM and the block, you're good to go.
Yeah, it was weird. It didn't come through the normal 'power' side, it decided to hit the coax going into the cable modem and scatter across the network. It was one of the most amazing things I've seen (and so, so costly).

Of course, it was the excuse needed to upgrade from a Phenom X4/4870 512 rig to the new one.
Totally unrelated to the thread, but do you have pics? :D
 
It depends on the matrix used. As long as they stay between the RAM and the block, you're good to go.
Don't want to take my rig apart until I have all the parts to assemble everything, so won't be able to find out for a few days. It just seems like it would be tough getting them to stay in one place while I work with it.

The instructions did recommend a small dab of MX-4 on the RAM chips in addition to the pads. Thought that was interesting, didn't realize you could mix them...
 
Running any overclock with stock voltages is safe. See if you can run the CCC max stably. MSI Afterburner is still kind of buggy, so if you aren't going for insane clocks, sticking to CCC limits will net you a ton of performance for little effort.

is it ok to still keep MSI Afterburner running after I use CCC to o/c? I use Afterburner to control/monitor my fan and usage levels..
 
Don't want to take my rig apart until I have all the parts to assemble everything, so won't be able to find out for a few days. It just seems like it would be tough getting them to stay in one place while I work with it.

The instructions did recommend a small dab of MX-4 on the RAM chips in addition to the pads. Thought that was interesting, didn't realize you could mix them...
TIM is TIM, mixing most shouldn't matter. It seems the pads are just there to fill the gap more than anything. Therefore, add a dab of MX-4, it won't hurt and should keep them in place until you get the block compressing it.
is it ok to still keep MSI Afterburner running after I use CCC to o/c? I use Afterburner to control/monitor my fan and usage levels..
That's what I do for my daily gaming; I use OSD to check out framerates and temps. The only problem I've had is using both applications to apply clock speeds over one another. Sometimes the card gets funky and stays locked in a low 3D mode (500MHz core, 1375MHz vRAM)
 
XFX 7970 Core Edition here.

1125/1575 overclock, via the CCC slider.

No problems, ever. Moving the extra power slider seems to have no impact on anything.

Max temp under load ~74 degrees
 
So, just got my EK full coverage block. Probably won't have a chance to install this along with the rest of my new WC stuff until the weekend.

This might seem like a silly question, but are the thermal pads for the RAM and such not supposed to be sticky at all?

I wondered that too when I installed my block. Instruction says peel the protective film. THERE'S NO FILM FFFFFFFUUUUU. Took me a while to just say fuck it and used a tiny dab of paste to hold the pads in place.

As to my OC:
Gigabyte 7970
1275/1700 @ 1.275v
Running at about 40C or so.

1300 was a no go at 1.3v :(
 
I wondered that too when I installed my block. Instruction says peel the protective film. THERE'S NO FILM FFFFFFFUUUUU. Took me a while to just say fuck it and used a tiny dab of paste to hold the pads in place.
Oh, mine arrived with plenty of protective film. Most of it just wasn't attached to the pads.

I'm wondering if I use a dab of thermal paste as well as the pads, if that'll make it a real pain to clean up if I ever need to re-install the stock HSF (like if I need to RMA it)?
 
I wondered that too when I installed my block. Instruction says peel the protective film. THERE'S NO FILM FFFFFFFUUUUU. Took me a while to just say fuck it and used a tiny dab of paste to hold the pads in place.

lol same here. was a film on one side - not both like the instructions insinuated and neither side was sticky. I just re-used the stock tim for memory.
 
I'm wondering if I use a dab of thermal paste as well as the pads, if that'll make it a real pain to clean up if I ever need to re-install the stock HSF (like if I need to RMA it)?
Not in the least, just give it a good wipe with some 91% isopropyl alcohol and you'll be good to go (you can pick that up at your local CVS/Walgreens/etc.).
lol same here. was a film on one side - not both like the instructions insinuated and neither side was sticky. I just re-used the stock tim for memory.
I'd be wary of using just TIM on RAM chips. They're very rough and you really have to gob it on to fill in the gaps. It might not make a difference, but generally the thermal pads perform the best since they're designed to fill the space. TIM performance suffers exponentially the thicker the space between the two mated surfaces.
 
I'd be wary of using just TIM on RAM chips. They're very rough and you really have to gob it on to fill in the gaps. It might not make a difference, but generally the thermal pads perform the best since they're designed to fill the space. TIM performance suffers exponentially the thicker the space between the two mated surfaces.

in this case the tim I'm refering to is the stock thermal pads. The ram chips had no grease.
 
I wonder if there's any reason not to? Are the ones that come with the waterblock better?


The stock ones were sorta rubbery and sticky. Teh ones that came with the EK block seems sorta like really thin paper/foam kinda stuff - not sticky at all.
 
The stock ones were sorta rubbery and sticky. Teh ones that came with the EK block seems sorta like really thin paper/foam kinda stuff - not sticky at all.
Right. I was just wondering if there would be any performance difference or something. Though I guess it would probably be marginal even if there was.
 
1225/1790 @ 1.225v 72°c full load

I started getting a black screen under full load after running for about 15 minutes at 1250 core with no artifacts during the test - increasing the voltage above 1225 didn't seem to help.

As for the memory, if I go above 1790 I get an all white screen and then a reboot when under full load after a few minutes. Oh well.

Edit: Looks like the AMD Leo demo made a liar out of me. I can do only hit 1150 on the core while running it. If I go any higher, the video driver fails and my screen goes black. I was able to run the Heaven Benchmark for hours with no issues. I guess Leo is quicker at telling you if you've got a stable clock or not?
 
Last edited:
im looking at the zalman reserator
http://www.zalman.com/ENG/product/Product_Read.asp?Idx=183
how much cooling power does this have in comparison to traditional 3x120 watercooling setups?
i am going to run a long water hose outside the house, the passive zalman dissipating all that juicy heat outside my house, saving electricity billz, i wont need the aircon in the room anymore....

going to get the zalman in conjunction with the EK-FC7970, any caveats? your thoughts?
 
im looking at the zalman reserator
http://www.zalman.com/ENG/product/Product_Read.asp?Idx=183
how much cooling power does this have in comparison to traditional 3x120 watercooling setups?


I guess that would depend on the outside temperature. If it's cold enough outside (like in the winter) that it could beat a 360 by quite a bit you might have to worry about condensation inside your case.

If it were inside the house the 320 would be better.

edit:

looking at the specs I see this:
Material : Anodized Pure Aluminum
could be bad. very bad.
 
Edit: Looks like the AMD Leo demo made a liar out of me. I can do only hit 1150 on the core while running it. If I go any higher, the video driver fails and my screen goes black. I was able to run the Heaven Benchmark for hours with no issues. I guess Leo is quicker at telling you if you've got a stable clock or not?
I had the same problem with Leo demo... 1150 seems to run fine. If you go higher than that, it's driver crash or black screen. I think it's the demo cannot handle the speed of it or it is too stressful compared to other software?

Maybe driver? :confused:
 
The Leo demo fully utilizes the capabilities of the 7970, much more so than Heaven.
 
PowerColor reference 7970 master XFX reference 7970 slave in Xfire

1125/1575 via CCC with 20% power saw a few slight artifacts on Crysis 2 ultra so backed them off to 1100/1500 and so far so good.

I love how much quieter these are than my old trifire 6990+6950 setup.. and the 7970s definitely seem a bit faster. Wish I would have done some direct FPS benchmark comparisons but alas was too impatient.
 
For the life of me I can not see why with MSI Afterburner earlier I was 1125/1725 24/7 @ 1.2 v and now ever since I fucking opened up Trixx the shit crashes with a black screen driver issue.

I uninstall Trixx, reinstall AB put the exact same settings in and nothing, same thing.

Originally Afterburner said I had a VID of 1.175 now beta 12 says 1.05. For some reason I simply can not, can not fuck with anything dealing the voltages or I get the black screen of death driver failure issue.

I want to say when I set it to constant voltages @ 1.20 originally it seemed to work fine but I can't remember how I got it to work right because that does as much as a monkey pissing in the wind atm.

GPUZ says I have an 83% quality GPU (Sapphire brand)
 
Last edited:
So, what is a good overclock on these cards? Also, what is safe voltage? Trixx says I am at 1.175 or something.
 
So, what is a good overclock on these cards? Also, what is safe voltage? Trixx says I am at 1.175 or something.

A good overclock depends on at what voltage you're talking about. At stock voltages, it happens to be around the CCC limits, but if you're willing to increase voltage in the 1.2-1.25V range and cool properly, a good OC seems to be upwards of 1200 core/1600 RAM.

As far as safe, I have no idea, so don't take my word for it, but I've seen others in this thread using values between 1.2 and 1.3 V, but some (many) of those people are water cooling too, so I would do some research on that one. I would assume (i.e. don't blame me if it fries your hardware) 1.2V is pretty safe though if you set your fan profile manually and toss the stock one (it doesn't ramp up quickly enough)
 
Max at ~ 1.3V is what I've seen repeatedly. I don't think I'd go over 1.25 tbh.
 
Thanks guys, I think I will stick with 1125/1575 for now. That is on stock voltages as well.
 
I had the same problem with Leo demo... 1150 seems to run fine. If you go higher than that, it's driver crash or black screen. I think it's the demo cannot handle the speed of it or it is too stressful compared to other software?

Maybe driver? :confused:

Yeah, it seems pretty weird that we both get stuck at the same core clock. Changing the voltage didn't seem to make a difference.

The Leo demo fully utilizes the capabilities of the 7970, much more so than Heaven.

I guess everyone should be running the AMD Leo demo instead of Heaven on loop to test the stability of their clock?
 
Leo runs fine on mine at 1150/1700 at 1.175v. Temps got up to 83C'ish at 60% fan after maybe 15 minute run.
 
I guess everyone should be running the AMD Leo demo instead of Heaven on loop to test the stability of their clock?
I want to see everyone to try Leo with their OC settings higher than 1150 core. If they can then this mean Leo is a great way to test your OC stability isn't it? :p
 
I'm running my Asus 7970 | 1200 core/1700 mem totally fine @1175. I am running a custom auto fan setting in Afterburner to keep the temps down. But its great.
 
I guess that would depend on the outside temperature. If it's cold enough outside (like in the winter) that it could beat a 360 by quite a bit you might have to worry about condensation inside your case.

If it were inside the house the 320 would be better.

edit:

looking at the specs I see this:
Material : Anodized Pure Aluminum
could be bad. very bad.


In Au, temps minimum are 10 o celcius minimum, 45 in summer, high humidity
what is bad about the anodize pure aluminum, will it corrode in the rain?
 
Back
Top