Good replacement thermal paste for 980 Ti?

[alcoholbob]

Thinking of replacing the thermal paste on the pair of EVGA 980 TI ACX cards I have now. A few months ago I had the Titan Xs with ACX cards and they ran way cooler. The ACX fans have to run 15% fan speed higher on these 980 Tis to maintain the same temps (~85C) versus the Titan Xs w/ ACX Coolers, which really increases the noise. Good lord the stock EVGA thermal paste must be crap because I have the side case open, it's practically an open bench.

I have the 980 Tis running at stock clock speeds (1190MHz) and they can't even keep up with my old Titan X's in temps (was running them at 1425MHz). Since the cooler is identical, and the ASICs of the cards are all about the same (70% range for both 980 Tis and Titan Xs) I'm suspecting this is a thermal paste issue since I used Noctua NT-H1 I had sitting around on the Titan Xs when I had originally installed the ACX coolers on them.

So should I just re-paste the cards using regular Noctua paste or should I try something more exotic or like liquid metal?

 

[TaintedSquirrel]

The best thermal paste is what you have on-hand. NT-H1 is well reviewed so I wouldn't hesitate to use it.

 

[Hail The Brain Slug]

I used Gelid GC Extreme to repaste my 980ti Classified. I had a custom fan profile in the bios I flashed onto it, but the repaste saw a 3C drop along with a fan speed drop of around 6%.

If I had thought about it I would have set a constant fan speed to test it, but I'm happy with those results. I have the fan profile set to to target 75C as that's one of the voltage steps. Before repasting it would sit at 72-74 at 64% fan, now it's 69-71 at 56-58% fan.

My first 980ti had a bump on the face of the heatsink that pressed into the GPU leaving a mark, so I had to send it back. But, I had to repaste it (And for kicks, I did testing and removed the heatsink again to see how the paste spread).

I was extremely pleased with how well it spread after applying it with the included spatula in a thin layer (As thin as it would go with the spatula while still covering the whole GPU).

I have extensively used AS5 and AS Ceramique and I even have a tube of Antec Nano-Diamond paste, but the GC extreme won me over just by how well it looked after I took the heatsink back off.

 

[Sabrewings]

I used the thermal paste that came with my EK block for it. Not sure if they sell it separately, but between it and the back plate my 980 Ti never gets over 40C. It definitely can move some heat.

 

[thetuna]

If you're not overclocking, I'd say just use whatever.

If you want the absolute best paste, I believe that current consensus gives the award to 'Thermal Grizzly Kryonaut'.
(I'm using that on my 290x and it hasn't caught fire yet)

If you want liquid metal... be very careful since it is conductive!
I would never use it on a 290x because the die is surrounded by tons of components.
It's probably just not worth the hassle for a few degrees.

 

[guskline]

Agree with Sabrewing. I also bought a separate tube of Gelid GC Extreme paste. I've also used the Noctua NT-H1 with great success.

 

[tweakboy]

Antec will do you good.

 

[RussianSensation]

The best thermal paste is what you have on-hand. NT-H1 is well reviewed so I wouldn't hesitate to use it.

I second this. If the OP already has it, I don't think buying another one is worth it as long as it's closer to the top pastes.

Source

Source

http://www.xbitlabs.com/articles/coolers/display/thermal-interface-roundup-2_7.html#sect0
http://www.xbitlabs.com/articles/coolers/display/thermal-interface-roundup-1_12.html#sect0

If you want the absolute best paste, I believe that current consensus gives the award to 'Thermal Grizzly Kryonaut'.

Thanks for the tip. I haven't kept up with TIMs fo a couple of years and never heard of that one. It does look great.

It is on the expensive side though - almost $6 for 1 gram. Where did you purchase yours?

 

[alcoholbob]

Thanks for the advice guys I'll just use the Noctua paste I currently have. Thanks for the graphs RS, very informative.

 

[RussianSensation]

Thanks for the advice guys I'll just use the Noctua paste I currently have. Thanks for the graphs RS, very informative.

:thumbsup:

Can you please let us know what the end results were with a thermal paste swap? That would be helpful for other 980Ti users as well as for any of us upgrading to 250W+ GPUs in the future and considering a TIM swap.

Thanks!

 

[Z15CAM]

What's the matter RussianSensation - Can't give up your old tube of Artic Silver 5. Neither can I but I`ll use MX3 from time to time whether it lasts as long - Who`s to know - LOL

I don't have to contend with 980Ti's but CF 290X's and they haven't change the initial temps sense I timmed them with Silver 5 almost 3 years ago under XSPC RAZOR Water Blocks.

I look for a long term constant solution rather than for a short term cooler solution.

 

[RussianSensation]

What's the matter RussianSensation - Can't give up your old tube of Artic Silver 5. Neither can I but I`ll use MX3 from time to time whether it lasts as long - Who`s to know - LOL.

What makes you think that? I haven't used AS5 in years and I switched to far superior options last decade. I've gone through Prolimatech PK-1, Tuniq TX-2, MX-2, MX-4 and now I use Thermalright Chill Factor 3. In fact, I've already switched from using AS5 when I moved from my C2D Q6600 to i7 860. :awe:

You probably haven't read a lot of my posts on the cooling sub-forum and CPU overclocking when that sort of thing was fun, but I stopped recommending AS5 ever since Tuniq TX-2 hit the scene and that was ages ago.

 

[alcoholbob]

Actually my results are pretty much in line with Linus' testing. My temps have dropped by about 3C. I ended up going with the spread method while applying the NH-T1 thermal paste. I would say it's okay but not spectacular, but took basically no effort (like 5 minutes for each card). I'm able to lower the fan speeds by about 5% and while maintaining slightly lower temps. I'm really starting to think ASIC scores aren't directly comparable between 980 Tis and Titan Xs though and maybe that's why I was able to run the ACX coolers with such a significantly slower fan speed on them.

 

[RussianSensation]

Actually my results are pretty much in line with Linus' testing. My temps have dropped by about 3C. I ended up going with the spread method while applying the NH-T1 thermal paste. I would say it's okay but not spectacular, but took basically no effort (like 5 minutes for each card). I'm able to lower the fan speeds by about 5% and while maintaining slightly lower temps. I'm really starting to think ASIC scores aren't directly comparable between 980 Tis and Titan Xs though and maybe that's why I was able to run the ACX coolers with such a significantly slower fan speed on them.

Thanks for that info. 3C is still a lot if you think about it since you used on-hand TIM and it's just TIM, not a new heatsink/waterblock or anything. I am pretty sure if a top-of-the-line waterblock was winning by 3C against its 2nd best competitor, that would be a big deal in that market segment. Also, 3C might allow your card to keep its maximum boost longer but that requires more extensive testing to prove.

How did you end up with a FreeSync monitor and 980Ti SLI? :awe:

 

[alcoholbob]

Thanks for that info. 3C is still a lot if you think about it since you used on-hand TIM and it's just TIM, not a new heatsink/waterblock or anything. I am pretty sure if a top-of-the-line waterblock was winning by 3C against its 2nd best competitor, that would be a big deal in that market segment. Also, 3C might allow your card to keep its maximum boost longer but that requires more extensive testing to prove.

How did you end up with a FreeSync monitor and 980Ti SLI? :awe:

I needed a new monitor and briefly switched to Fury Crossfire but the had really poor minimum frame rate (literally hitching) in a lot of games so I'll probably switch to AMD in the next gen lineup but I wasn't really feeling the benefit of Freesync with the high current AMD DX11 overhead.

 

[RussianSensation]

I needed a new monitor and briefly switched to Fury Crossfire but the had really poor minimum frame rate (literally hitching) in a lot of games so I'll probably switch to AMD in the next gen lineup but I wasn't really feeling the benefit of Freesync with the high current AMD DX11 overhead.

I think you made the right decision anyway. This generation above $550, NV has it in the bank for most use cases. $1100 Fury CF vs. $1300 980Ti SLI but he latter is much faster once OCed, 50% more VRAM. Sounds odd that you even considered the Fury CF with your budget / components. :biggrin: Honestly, from where the Fury sits vs. 390/390X, I bet in the real world, the Fury CF won't be that much better than 390/390X CF. At least the 980Ti offers 20-25% more overclocking headroom and it's a decent amount faster than the Fury.

 

[thetuna]

Thanks for the tip. I haven't kept up with TIMs fo a couple of years and never heard of that one. It does look great.

It is on the expensive side though - almost $6 for 1 gram. Where did you purchase yours?

It is relatively expensive and doesn't have very many suppliers, but I guess you can charge a lot when you're the best
I got it at performance-pcs.com, but only because I was ordering other stuff as well.
If you factor in shipping, amazon is probably a tad cheaper.