Been running stable tested on a few games all day. Great drivers so far. Looking forward to trying Reflex out in Warzone once it's patched in tomorrow.

Wonder if the stability thing is related to the caps issue on the 3080s

I'd like to see how the new drivers affect performance. Can't help but feel if you have a hardware problem (poor quality caps) how can you resolve it with software other than lowering the power. It's not the first time a graphics card series has been released with "fast" drivers for the reviews, and then slowed down when they realise they cause too many crashes.

IgWannA wrote on Sep 29, 2020, 09:53:
I'd like to see how the new drivers affect performance. Can't help but feel if you have a hardware problem (poor quality caps) how can you resolve it with software other than lowering the power. It's not the first time a graphics card series has been released with "fast" drivers for the reviews, and then slowed down when they realise they cause too many crashes.

I've seen claims they over-volted the shit out of these new cards to improve yields, and an actually good chip will perform just as well using a full 100W less power.

Source: "I've seen claims"
Claim: "they over-volted the shit out of these new cards to improve yields"
Statistical Evidence: "an actually good chip will perform just as well using a full 100W less power"

30% less power.
:-facepalm-:

The claims I've seen say that a really really good 3080 will actually give power back to the system and basically runs for free.
sigh

-50 MHz boost - this is what the new drivers seem to do to stabilize the RTX3000 cards. A bit of underclocking - who would have guessed?

There's also the possibility that there were driver bugs as well as component power issues.

Sorry, I guess I was unreasonable in making the assumption that everyone here knows how to use Google...

Reddit is full of average Joes experimenting with reducing voltage to save power with minimal to no performance loss, as well as discussion of reviewers and YouTubers demonstrating the same. Some stay in the range of 30-50W reduction to keep performance on par with the stock config, while others push it as far as they can, often ~100W.

WCCFTech also posted an article about it this morning, showing step-by-step with a video tutorial how they could cut power by 25% in exchange for 3% less performance (up to 35% less for 6% loss), though it looks like they only tested in Forza. Makes for an impressive comparison to the 2080 Ti, nonetheless, where the 3080 uses less power while still being at least 25% faster.

Like with overclocking results, YMMV, but it's becoming pretty clear NVIDIA made these things unnecessarily power-hungry in order for them to measure up to whatever arbitrary metric(s) they were targeting.

I still run a 1060 card, so I assume there is no point at all in updating to these drivers?