I installed the Kritical pads thermal pad set today and the results are just about as I was expecting.  My particular card already had fairly good memory temps with the stock EVGA thermal pads and compared to others I've seen posting online, I couldn't see how they could or would get dramatically better.   Long story short?  The Kritical pads performed noticeably better (4C drop) during the Nbminer test and almost identical to the EVGA stock pads during the TimeSpy Extreme stress test. 
 
I had replaced the EVGA thermal putty that covers the VRM areas with T-Global TG-PP-10 putty a couple months ago when I first got the card so that may or may not have some influence on temps when using the stock EVGA thermal pads.  I re applied the T-Global today when installing the Kritical pads and did not use the pads that Kritical supplies to replace the putty over the VRM areas. 
 
A few notes:

• GPU is an EVGA 3080 ti XC3 Ultra Gaming
• All tests conducted at a constant ambient temperature of 22 Celcius.  
• NBMiner EHT for 15 minutes was used as a pure VRAM max/avg temperature test.
• 20 loops of Timespy Extreme stress test was used a general gaming max/avg temp test.
• The NBMiner test used a static fan speed (details below)
• The Timespy Extreme stress test used a fan curve (details below)
• T-Global TP-PP-10 putty was used on VRM's in place of EVGA putty for all tests

 
NBMiner EHT - 15 minutes
------------------------------------------------------------
Static Fan Speed: 80% / 2,615 RPM
Power Limit: 85%
Mem OC (After burner): +500
Room Ambient: 22C
 
EVGA Stock Pads
--
Max VRAM temp: 84C
Avg VRAM temp: 73C
 
 
KRITICAL Pads
--
Max VRAM temp: 80C
Avg VRAM temp: 66C
  
 
Timespy Extreme Stress Test - 20 loops
---------------------------------------------------------------------------------------
Fan Curve Max: 72% / 2,376 RPM
Power Limit: 100%
Core OC (Afterburner): +150
Mem OC (Afterburner): +500
Room Ambient: 22C
 
EVGA Stock Pads
--
Max VRAM temp: 78C
Avg VRAM temp: 74C
 
 
KRITICAL Pads
--
Max VRAM temp: 78C
Avg VRAM temp: 72C
 
So definitely some worthwhile gains in the mining scenario and pretty much identical temps with the Timespy Extreme tests.  Real world gaming thermals differ dramatically from game to game, however and perhaps similar gains as seen with the mining test will be more apparent in different games. It was fun to test, though and I hope the new pads/putty last a good while.  
 

Kritical Pads: General Gaming Test  
 
I followed the "controlled tests" above with about 4 hours of non stop gaming at my cards max overclock and fans at +/- 73% ... Horizon Zero Dawn, Cyberpunk 2077, Metro Exodus, Rise of the Tomb Raider, Control and Hellblade: Senua's Sacrifice. All gamed running at 3840x1600 resolution and ultra settings. These are the games that I've personally found to most push my GPU the most. During that real world gaming scenario, VRAM temps with the Kritical pads never exceeded 78C while core temps never went past 68C ... those two numbers were eerily consistent throughout the entire 4 hour gaming session, with a near constant 10C delta between VRAM and core temps throughout. Here's the data as reported by HWINFO64 during the above mentioned gaming test...
 
 
 
I  honestly didn't want to install the Kritical pads cause things were already working very well with the card and thought I might jinx it. LOL ... I'm very much of the opinion that if it works, then don't try and fix it but my curiosity generally wins out.
 
Final Thoughts  
 
It's safe to say the Kritical pads are a sure bet and well worth the money ($29.95 for a complete 3080ti XC3 set).  My particular 3080Ti performed quite well already with the stock EVGA pads on the VRAM but some users are having a much different experience, seeing VRAM temps hitting upwards of  100C in many cases depending on the particular 30 series GPU and in those cases, replacing thermal pads is the obvious first step in attempting to bring down memory temps. 
 
The Kritical pads come very well packaged and the cut quality is precise and consistent. The owner has built a very clean and functional ordering system and the entire process was snag free from click to ship.  The pads themselves are also very easy to work with and do not deform too easily but also are highly compressible.  And Kritical pads come in the precise EVGA spec'd heights that are nearly impossible to find short of custom sourcing (ie: 2.25mm pads over VRAM / 2.75mm pads over other areas, etc.).  So again, absolutely no reason to buy whole sheets and cut your own pads unless you enjoy self mutilation or are otherwise slightly masochistic lol ... when you can simply order demonstrably high quality pre cut kits from Kritical. 
 
It should be said that one can also order complete pad sets direct from EVGA too if you request them and as the above tests have shown, the EVGA pads are really not that bad. Unlike Kritical sets, EVGA does not include pads to replace the thermal putty over the VRM areas and you'd need to buy the EVGA putty separately (or T-Global TG-PP-10 putty).  And also very important to note here ... if you ever need to request RMA/Warranty repair for your EVGA GPU, then you will need to send your card back in absolute factory condition.  Therefore, if you've changed either the stock EVGA thermal pads or putty, then you will need to re install those stock pads/putty before shipping your card back to avoid possible (and probable) denial of your warranty claim. 
 
Oh yeah, repasted the GPU core today (Noctua NH-1 same as before) and max core temps went from 70C to 68C in the Timespy Extreme test so I'll take that too, thank you.  
 
~s1rrah

ty_ger07
Note: if you ever send the card into EVGA for a warranty claim, you will need to buy that thermal putty and reapply it where EVGA originally decided that thermal putty waa appropriate.

For the time, expense, and hassle, your results seem like it isn't really worth it.

Your point about replacing the VRM putty with EVGA's own putty and also re applying stock EVGA thermal pads on the VRAM in an RMA/Warranty claim scenario is very important and absolutely correct as EVGA has and will deny warranty claims if any given card is not in am absolute factory configuration when returned.  I keep both stock EVGA pads and putty on hand for just such a worst case scenario but hope I never see that to be necessary.  Thanks for that and I edited the original review to reflect your point...
 
And your also correct in regards to general gaming and my particular test results, though serious miners will take every thermal performance gain they can get so a 4C drop in my test could be of interest to some. 
 
That being said, other users have had dramatically better results overall when replacing EVGA stock thermal pads, some seeing up to a 20C drop but my card simply worked really well right out of the box. 
 
Depending on the particular 30 series GPU, be it from EVGA or some other manufacturer?  ~100C VRAM temps are not uncommon in the least.  In such cases I think it certainly worthwhile to try various modifications to lower that number if possible and replacing VRAM pads is the obvious first step.  And having access to pre cut, complete sets such as Kritical provides can save folk a great deal of time/work during that process.  
 
 

They do supply the original pads and putty just for this reason.

ty_ger07
I just wonder if such small improvements of the mod done in this thread is worth the time, hassle, and expense.

For me it was simply a novel experiment and completely unnecessary as my memory temps were really great right out of the box with this XC3 card.  Other users, some with the very same card and some with 30 series cards from Asus/Gigabyte/Nvidia/etc. are commonly seeing memory temps hit 100C and higher and in those latter cases I think replacing the VRAM thermal pads would be highly recommended be it using pads from Kritical or otherwise. 
 
A user at overclock.net recently messaged me, saying his stock TUF 3080Ti sees memory temps spike at 104C when gaming and that he would be applying new thermal pads simply to try and reduce that number.  So I think it's a case by case situation whether it might be worth the effort or not.  But as far as tear downs go, the EVGA 30 series (and others) are really quite simple/easy; there is really nothing difficult about it.  The Nvidia FE cards are another story LOL ... I installed two EVGA hybrid coolers on dual 980 FE cards back in the day and man what a nightmare of tiny screws and strange alignments that was.    
 
But I enjoy taking things apart and trying to further improve them so results notwithstanding, I find it worthwhile and fun.