|
Author |
Message |
RhynoCZ |
This post is not being displayed .
|
RhynoCZ Super Spammer
Joined: 09 Mar 2012 Karma :
|
Posted: 20:28 - 17 Jun 2019 Post subject: Thermal paste replacement (GPU) |
|
|
Did anyone here ever do the thermal paste replacement on a GPU? I've got a MSI GTX 1070 Ti Gaming X here and looking at all the benchmarks, how people get anything between 66°C and 73°C at full load, while the fans are running at rather lower rpm and the GPU boost over 1800Mhz, I can't seem to get such values. Not even when I remove the side panel and run the gfx card fans at 100% and blow a fresh air on it using a 140x140mm Noctua fan at 100% speed (roughly 1200rpm). My card runs above 80°C and throttles even down to the base clock of roughly 1600 MHz.
The internet is not full of answers, in this case. Some people even run a generous overclock while still keeping the temperatures between 66°C and 73°C.
So, replacing the paste? Could it help? What kind of paste? The process of replacing it is fairly straightforward. ____________________ '87 Honda XBR 500, '96 Kawasaki ZX7R P1, '90 Honda CB-1, '88 Kawasaki GPz550, MZ 150 ETZ
'95 Mercedes-Benz w202 C200 CGI; MZ 150 ETZ, '98 Mercedes-Benz w210 E200 Kompressor |
|
Back to top |
|
You must be logged in to rate posts |
|
|
Pjay |
This post is not being displayed .
|
Pjay World Chat Champion
Joined: 18 Jan 2016 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
RhynoCZ |
This post is not being displayed .
|
RhynoCZ Super Spammer
Joined: 09 Mar 2012 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
Pjay |
This post is not being displayed .
|
Pjay World Chat Champion
Joined: 18 Jan 2016 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
P. |
This post is not being displayed .
|
P. Red Rocket
Joined: 14 Feb 2008 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
Easy-X |
This post is not being displayed .
|
Easy-X Super Spammer
Joined: 08 Mar 2019 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
Hetzer |
This post is not being displayed .
|
Hetzer Super Spammer
Joined: 19 Feb 2007 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
Pjay |
This post is not being displayed .
|
Pjay World Chat Champion
Joined: 18 Jan 2016 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
RhynoCZ |
This post is not being displayed .
|
RhynoCZ Super Spammer
Joined: 09 Mar 2012 Karma :
|
Posted: 09:56 - 18 Jun 2019 Post subject: |
|
|
A lot to think about, thank you, lads.
My experience with ''stock'' paste was always it wasn't the best you could use, but it worked more than adequately and was very long lasting. With this MSI GTX 1070 Ti Gaming X however I can see already it doesn't perform that well, thermally, compared to any benchmark data out there.
And the issue is not only the noise, but also the card is slowing down when it gets near its maximum temperature target (83°C by default), which means losing up to 10 frames per second at 60 fps, which is annoying.
I'm baffled, as I have mentioned before, all the reviewers don't go above 70°C while testing their samples, with the fans running at roughly 1200 to 1400 rpm, which is what, 50% of the maximum speed, at best?
The chip in my card otherwise is very capable and with the max. thermal target moved to 92°C, I can do 2.0GHz on the GPU and put extra 500 Mhz on the v-ram with no issues at all. I don't run the card OC'd because the performance benefits aren't that great in the games, that I play, at 60Hz.
So, the only culprit I can see is either the paste, or the heatsink or maybe MSI did something wrong and didn't share this bit of info with its customers. I do know there was a slight controversy when MSI switched to Micron vram and many people had to flash the bios to get the cards working properly, but again, not my issue, I've got a later revision of the card, obviously.
Thinking of it, I could have just bought the blower cooler style card for less money and with only one 8 pin connector, to get this kind of performance (temp/clocks/noise), that I get from my ''premium'' gaming X card.
Link: https://youtu.be/kU0wZZ8QjWw?t=578 2GHz at 68°C. I can't get 68°C even with the stock clock speed, with the side panel of the case off. Even his idle temp. is way bellow what I get.
With the side panel on and two intake, one exhaust fan, I get 79°C while the card throttles from 1820Mhz (boost) to about 1770Mhz and the fan is running at 85% (roughly 2200rpm). With the panel off, we're talking 76 to 78°C, still throttling. ____________________ '87 Honda XBR 500, '96 Kawasaki ZX7R P1, '90 Honda CB-1, '88 Kawasaki GPz550, MZ 150 ETZ
'95 Mercedes-Benz w202 C200 CGI; MZ 150 ETZ, '98 Mercedes-Benz w210 E200 Kompressor |
|
Back to top |
|
You must be logged in to rate posts |
|
|
P. |
This post is not being displayed .
|
P. Red Rocket
Joined: 14 Feb 2008 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
RhynoCZ |
This post is not being displayed .
|
RhynoCZ Super Spammer
Joined: 09 Mar 2012 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
P. |
This post is not being displayed .
|
P. Red Rocket
Joined: 14 Feb 2008 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
Hetzer |
This post is not being displayed .
|
Hetzer Super Spammer
Joined: 19 Feb 2007 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
Pjay |
This post is not being displayed .
|
Pjay World Chat Champion
Joined: 18 Jan 2016 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
Pjay |
This post is not being displayed .
|
Pjay World Chat Champion
Joined: 18 Jan 2016 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
RhynoCZ |
This post is not being displayed .
|
RhynoCZ Super Spammer
Joined: 09 Mar 2012 Karma :
|
Posted: 21:15 - 18 Jun 2019 Post subject: |
|
|
That's not really a fair comparison. He, just like you did, replaced a stock thermal compound, which was designed to perform well for the whole life time of the card and to have long shelf life, with a high performance paste that will however deteriorate very quickly, compared to the ''stock'' paste.
That still does not explain why my gfx card, having it about 1 and a half year now, is running hot, even though if still in spec, therefore no warranty claims would be acknowledged. There is no dust in the heatsink, the fans do work as they should, the air flow in the case is perfect, I even tried to put the card into my other Define S2 PC (a huge case with 3 140x140mm fans, pushing thought so much air, that I can run the CPU passively. The S2 is also much wider, so the card has more space the breath, TLDR, I could get it down by 1°C which means no difference at all.
I will change the paste, as the last resort, then it's liquid cooling time. ____________________ '87 Honda XBR 500, '96 Kawasaki ZX7R P1, '90 Honda CB-1, '88 Kawasaki GPz550, MZ 150 ETZ
'95 Mercedes-Benz w202 C200 CGI; MZ 150 ETZ, '98 Mercedes-Benz w210 E200 Kompressor |
|
Back to top |
|
You must be logged in to rate posts |
|
|
Hetzer |
This post is not being displayed .
|
Hetzer Super Spammer
Joined: 19 Feb 2007 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
Hetzer |
This post is not being displayed .
|
Hetzer Super Spammer
Joined: 19 Feb 2007 Karma :
|
Posted: 08:47 - 19 Jun 2019 Post subject: |
|
|
RhynoCZ wrote: | That's not really a fair comparison. He, just like you did, replaced a stock thermal compound, which was designed to perform well for the whole life time of the card and to have long shelf life, with a high performance paste that will however deteriorate very quickly, compared to the ''stock'' paste.
That still does not explain why my gfx card, having it about 1 and a half year now, is running hot, even though if still in spec, therefore no warranty claims would be acknowledged. There is no dust in the heatsink, the fans do work as they should, the air flow in the case is perfect, I even tried to put the card into my other Define S2 PC (a huge case with 3 140x140mm fans, pushing thought so much air, that I can run the CPU passively. The S2 is also much wider, so the card has more space the breath, TLDR, I could get it down by 1°C which means no difference at all.
I will change the paste, as the last resort, then it's liquid cooling time. |
I've come across blocks that have partially un-clamped, very hard to spot until disassembly. ____________________ "There's the horizon! Ride hard, ride fast and cut down all who stand in your way!" |
|
Back to top |
|
You must be logged in to rate posts |
|
|
Ste |
This post is not being displayed .
|
Ste Not Work Safe
Joined: 01 Sep 2002 Karma :
|
Posted: 09:05 - 19 Jun 2019 Post subject: |
|
|
Play with furmark for testing GPU temperatures. https://geeks3d.com/furmark/downloads/
I'm with Hetzer about replacing the thermal paste shouldn't make much difference. Then again, you can't tell if it's a poorly clamped chip, a crappy quality surface on the block, cheap-ass original paste or whatever without dismantling it and replacing the thermal paste.
It's an easy enough job and you'll also get to clean all the dust out from the heat sink(s) which will also help lower temperatures. |
|
Back to top |
|
You must be logged in to rate posts |
|
|
RhynoCZ |
This post is not being displayed .
|
RhynoCZ Super Spammer
Joined: 09 Mar 2012 Karma :
|
Posted: 17:28 - 19 Jun 2019 Post subject: |
|
|
So, lads, I have replaced the thermal compound. I used the Silentium Pactum PT-1, had one syringe in my drawer, thinking if the thermal paste/heatsink fitment was my issue, even a lower tier thermal compound should show some difference, and so far it does. Although, I'm not sure about its longevity. Its heat conductivity rating is only 4 W/mK, while the standard pastes on the market are double of that. That being said, I haven't got a slightels clue if 8 W/mK is really better than 4 W/mK, or just negligeably.
I might gonna invest in some ARCTIC MX-4, or whatever, if today's efforts fail.
Notice the boost clock and fan speed. The game runs in true 4K.
BEFORE
https://cdn.bcf.44bytes.net/files/img_20190617_175842.jpg
AFTER
https://cdn.bcf.44bytes.net/files/img_20190619_180134.jpg
AFTER with roughly the same fan speed as it would run before
https://cdn.bcf.44bytes.net/files/img_20190619_180354.jpg
Some deep grooves on the heatsink
https://cdn.bcf.44bytes.net/files/img_20190619_172622.jpg ____________________ '87 Honda XBR 500, '96 Kawasaki ZX7R P1, '90 Honda CB-1, '88 Kawasaki GPz550, MZ 150 ETZ
'95 Mercedes-Benz w202 C200 CGI; MZ 150 ETZ, '98 Mercedes-Benz w210 E200 Kompressor |
|
Back to top |
|
You must be logged in to rate posts |
|
|
Ste |
This post is not being displayed .
|
Ste Not Work Safe
Joined: 01 Sep 2002 Karma :
|
Posted: 19:29 - 19 Jun 2019 Post subject: |
|
|
Use bench marking / stress testing programs if you're going to get geeky about measuring differences in temperatures. |
|
Back to top |
|
You must be logged in to rate posts |
|
|
RhynoCZ |
This post is not being displayed .
|
RhynoCZ Super Spammer
Joined: 09 Mar 2012 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
Pjay |
This post is not being displayed .
|
Pjay World Chat Champion
Joined: 18 Jan 2016 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
Ayrton |
This post is not being displayed .
|
Ayrton World Chat Champion
Joined: 02 Sep 2010 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
Ste |
This post is not being displayed .
|
Ste Not Work Safe
Joined: 01 Sep 2002 Karma :
|
|
Back to top |
|
You must be logged in to rate posts |
|
|
Old Thread Alert!
The last post was made 4 years, 283 days ago. Instead of replying here, would creating a new thread be more useful? |
|
|
|