Question RTX 3090 Suprim X Pad change help pls!

[HelpPls514]

Due to Artifacting while watching YTB videos (and somehow only when watching YTB videos not while gaming etc., no matter the browser. I tried several) I checked VRAM temps and noticed that it idles at 60+C. I then proceeded to under-volt and it went down to 40-45C

I then saw a Guide on what the official thermal pad measurements are but didn't quite understand how to read the measurements and what to buy (and how much of what exactly). Where I'm from I have these options for purchase available, but I couldn't quite grasp how many of what exactly I need:

1.) Artic TP-3 0,5 - 1,5 mm thickness pads. (it doesn't say how much W/mk)
2.) EC360® Gold 14,5W/mK (50x50 --> 0,5mm soft - 2,0 mm soft // 50x50 --> 0,5-3,0mm // 100x100 --> 0,5- 1,5mm soft // 100x100 --> 0,5-1,5)
3.) EC360® PLATINUM 16,6W/mK (same as above)
4.) Gelid Solutions GP-Extreme ( 80x40x 0,5-3,0)
5.) THERMALRIGHT Thermal Pad 14.8 W/mk (85x45x 4)
6.) Gelid Solutions GP-Ultimate (90x50x 0,5-3,0)
7.) One enjoy 12,8 W/mK (85x45x 0,5-3mm)


Could any one please give me some advice on what to buy (and how many pads) please ?

P.S: I have a NZXT H7 Flow, Card's temp's (other than VRAM) were quite cool (but I never checked VRAM Temps because my card never went above 70C under full load --> prior to undervolting)

In Addition, my LED strip on the side died (and I have already read that that's sadly a rather common occurrence) but I found a replacement shell and if I'm to change the pads anyways, I can also just swap the entire shell. Just need to understand what pads I have to buy.

I'd highly appreciate help!!!

 

[Lutfij]

Welcome to the forums, newcomer!

This is a good read for you and anyone in the same boat as you're in now;

Reddit - Dive into anything

www.reddit.com

with this being a highlight out of that reddit;

https://imgur.com/SImF509

View: https://imgur.com/SImF509


Where are you located?

 

[HelpPls514]

Welcome to the forums, newcomer!

This is a good read for you and anyone in the same boat as you're in now;

Reddit - Dive into anything

www.reddit.com

with this being a highlight out of that reddit;

https://imgur.com/SImF509

View: https://imgur.com/SImF509


Where are you located?

Thank you for your answer. That is actually exactly what I came across upon researching, but I couldn't quite decipher how to read the measurements. I do not understand how many Pads I have to buy and what for example BS88 or BS75K etc. mean. This might be stupidity on my part, or rather I'm sure that it is but yeah.

 

[Phaaze88]

Only in Youtube? Have you tested different drivers, or your browser hardware acceleration settings?

GDDR6X is good up to at least 95C(sustained, a spike is mostly irrelevant). Are you even getting close to that?

 

[HelpPls514]

Only in Youtube? Have you tested different drivers, or your browser hardware acceleration settings?

GDDR6X is good up to at least 95C(sustained, a spike is mostly irrelevant). Are you even getting close to that?

Yeah only while watching youtube. I have updated drivers several times since I had this issue and HWA is activated but if I disable it watching 4k becomes impossible (but even if I disabled it, would that tell me something specific?)

Before undervolting the memory was getting very hot (don't remember exactly but somewhere around 90 under load I think) and the hotspot was also very hot (above 100) but the card itself was at max 70C which is why I never noticed it

 

[Phaaze88]

Yeah only while watching youtube. I have updated drivers several times since I had this issue and HWA is activated but if I disable it watching 4k becomes impossible (but even if I disabled it, would that tell me something specific?)

Even if 4K becomes unwatchable, would you still see artifacts? The monitor cable may be another factor to consider.

Before undervolting the memory was getting very hot (don't remember exactly but somewhere around 90 under load I think) and the hotspot was also very hot (above 100) but the card itself was at max 70C which is why I never noticed it

Applications akin to Msi Afterburner don't display gpu hot spot nor memory junction, so it's understandable that they'd go under the radar. Gpu-Z and hwinfo show them though.

You indirectly cooled down the memory because you cut down on how much power the gpu can use. The actual voltage going to the memory is not accessible to the user and is locked.

 

[HelpPls514]

Even if 4K becomes unwatchable, would you still see artifacts? The monitor cable may be another factor to consider.


Applications akin to Msi Afterburner don't display gpu hot spot nor memory junction, so it's understandable that they'd go under the radar. Gpu-Z and hwinfo show them though.

You indirectly cooled down the memory because you cut down on how much power the gpu can use. The actual voltage going to the memory is not accessible to the user and is locked.

I have turned it off now and will observe whether it appears again.
Yeah after seeing the artifacts I monitored the temps with GPU-Z which is why I found out about the temps.

And yes by under-volting the memory temps as well as the hotspot etc. went down. Now I have 27-28C on idle for the card itself and around 40C for the VRAMS which is quite impressive

 

[HelpPls514]

Even if 4K becomes unwatchable, would you still see artifacts? The monitor cable may be another factor to consider.


Applications akin to Msi Afterburner don't display gpu hot spot nor memory junction, so it's understandable that they'd go under the radar. Gpu-Z and hwinfo show them though.

You indirectly cooled down the memory because you cut down on how much power the gpu can use. The actual voltage going to the memory is not accessible to the user and is locked.

Update: Disabling Hardware Acceleration seems to have fixed the problem... I tried another fix where I changed the video encoding in the browser from DirectX to OpenGL since that supposedly works better with NVIDIA but sadly that didn't make a difference.