NVIDIA ‘Ampere’ 8nm Graphics Cards

[gpuerrilla]

Just out of curiosity when I retire from work, ideally if playing on 1440p screens should we upscale to 4k to avoid any possible bottleneck at 1440p, I assume my 3700x would bottleneck at 1440p?

I would try it out, be interesting to know.

 

[Jimmy Weirdarms]

Just out of curiosity when I retire from work, ideally if playing on 1440p screens should we upscale to 4k to avoid any possible bottleneck at 1440p, I assume my 3700x would bottleneck at 1440p?

This make no sense!
While your cpu *may* cause you to lose one or two FPS, it won't be much and it will be from a higher starting point.

Think of it this way - your CPU can probably cope with a certain number of FPS (game dependent). Let's say 120. Your GPU is capable of rendering at 1440p frames at 144 FPS. But oh no! You're down to 120 as the CPU is maxed out.

If you persuade your GPU to render at 4k and then down-res to 1440p, your GPU is doing a ton of extra work on the 4K image, and can only put out 80 FPS. Now your CPU is not a bottleneck! But you're still getting a lower framerate.

 

[Grim5]

CHONK

 

[gpuerrilla]

CHONK

OMG its already attacking your evga, greedy monster!!

 

[V F]

So the 3080 is the same depth as an EVGA 1080Ti FTW3 but thicker? They're still getting longer.

 

[ALXAndy]

https://videocardz.com/newz/msi-qui...-gaming-x-trio-design-amid-stability-concerns

MSI redesign their PCBs.

https://videocardz.com/newz/asus-al...eforce-rtx-3080-tuf-and-rog-strix-pcb-designs

And so do Asus.

So much for the TUF being a good card.

 

[Iron Man]

The rog strix was redesigned too so we will probably see very small quantities of this being released for a while.

 

[opethdisciple]

Apparently the new drivers fix the crashing at the expensive of GPU boost.

 

[Twinz]

Apparently the new drivers fix the crashing at the expensive of GPU boost.

https://www.pcworld.com/article/358...-3080-crashes-by-sacrificing-clock-speed.html

To put this in proper context, these maximum GPU Boost clock speeds remain well above the rated boost speeds for these cards. The “fix” results in such a minor speed difference that you won’t practically notice it in the game itself. After repeatedly hammering the benchmark, I managed to get HZD to complete a 1440p run once with the original, borked drivers. It averaged 129 frames per second. With the new, slightly lower-clocked drivers installed, it averaged either 127 or 128 frames per second across three runs. That’s essentially the same performance. When I tested the Nvidia RTX 3080 Founders Edition with the new drivers, there was no performance or power draw change whatsoever, though it didn’t come close to breaking the 2GHz barrier.

 

[Grim5]

Apparently the new drivers fix the crashing at the expensive of GPU boost.

depends on the card, some cards have better boost now - low end cards have lower boost now and high end cards either maintained or improved boost

 

[DFroN]

Apparently the new drivers fix the crashing at the expensive of GPU boost.

Depends, I’ve gained 100MHz because I had to downclock to avoid crashing before but now I can overclock and it’s stable.

At stock the clocks are the same for me with the new driver, except they are more stable instead of jumping about erratically.

 

[Vertigo]

https://videocardz.com/newz/msi-qui...-gaming-x-trio-design-amid-stability-concerns

MSI redesign their PCBs.

https://videocardz.com/newz/asus-al...eforce-rtx-3080-tuf-and-rog-strix-pcb-designs

And so do Asus.

No they didn't, they only changed some out-of-date pictures on their site to reflect the revised cards, which were revised in pre-production.

Claiming Asus "redesigned their PCBs" in the same way as MSI is disingenuous in the extreme as they never actually shipped any of the earlier revision.

By contrast, MSI initially blamed a potential driver issue and have now quietly altered their cards when a lot of customers already have the older versions with no word on what they'll do for those customers, if anything. This doesn't surprise me in the least as MSI have shown themselves to have a complete lack of ethics recently.

So much for the TUF being a good card.

The TUF is fine as they're all the new revision which appears to be one of the best, not only in it's PCB design but also price, cooling and performance.

 

[Vertigo]

Apparently the new drivers fix the crashing at the expensive of GPU boost.

Yep, so they've basically nerfed the cards, rendering all existing benchmarks invalid.

This total **** show just goes on and on.

 

[DFroN]

Yep, so they've basically nerfed the cards, rendering all existing benchmarks invalid.

This total **** show just goes on and on.

AIB cards can boost over 2GHz stable now, I’ve gained at least 100MHz, how’s that a nerf. There’s that one article above where a guy tries a single game and the new driver boosted 15MHz less for him, his room was probably 3 degrees warmer or something.

 

[Grim5]

Yep, so they've basically nerfed the cards, rendering all existing benchmarks invalid.

This total **** show just goes on and on.

whatbthe hell are you on about. Nothing was nerfed, those cards were incorrectly boosting too high

 

[XeNoNF50]

Yep, so they've basically nerfed the cards, rendering all existing benchmarks invalid.

This total **** show just goes on and on.

Nerfed is not my experience of the new driver. Perhaps the max boost that you used to get is slightly lower but sustained clock speed, performance over a longer period and stability is much higher. On my card I am getting 75mhz higher sustained clock speed and running at an even slightly lower voltage and it seems even more stable than my original undervolt.

 

[Iron Man]

So hard to figure out what’s going on with these new cards as there is too much conflicting information being reported now on these new cards.

 

[Darujhistan]

Nvidia 3000 launch - the comedy gift that keeps on giving!

Anyone willing to pre-order one of these at this stage is a lunatic.

 

[robfosters]

CHONK

Blimey!Its like a brick.

 

[Vertigo]

whatbthe hell are you on about. Nothing was nerfed, those cards were incorrectly boosting too high

Says who? NVidia? Sounds like marketing guff to me.

Were the cards really "incorrectly boosting too high" or was that how they were designed to boost and they've nerfed the speeds a bit due to the crashing problems?

Sounds very fishy to me to suggest that NVidia just happened to make a mistake with their drivers that caused the cards to boost too high at the exact same time that there was this variation in capacitor implementation by the AIBs which exacerbated the problem.