NVIDIA 5000 SERIES

[Minstadave]

Because they want a small form factor card. And the FE is the only one out there.

Small form factor using a 2 grand and 600w GPU is a pretty niche market and noisy prospect. It's not going to massively increase their sales vs a 3 slot quieter card.

It's impressive but it feels almost a step too far.

 

[Nitefly]

The original 4080 and the 4080 SUPER are totally different kettles of fish.

In the original launch line-up, where the 4080FE was £1199 (actually £1269 for the first couple of months), the 4090 was easily the best bang-for-buck card.

With the launch of the 4080 SUPER which, whilst only a bit faster, was significantly cheaper at £959, that card became the best value.

Yes - I mentioned about this in my post; the 4080 was not great value but the 4080 super was.

 

[sx_turbo]

So 5090 review is up, as predicted it's pretty impressive, but the power draw...

NVIDIA GeForce RTX 5090 Founders Edition Review - The New Flagship

NVIDIA's GeForce RTX 5090 is the fastest graphics card ever released. It comes with a whopping 32 GB VRAM, and support for multi-frame generation, which achieves hundreds of FPS easily. We also managed to disassemble the card, and our review includes performance testing without upscaling or...

www.techpowerup.com

 

[Hostile_18]

I'm leaning to a 5080 TBH now. Insane power increase and noise is not what I want.

 

[Halliphax]

I wish reviewers would drop 1440p testing as it's mostly pointless, with 4090 class and faster GPU's, it's really time to stop benching these cards at less than 4K as they just aren't made for such a low resolution and the CPU's can't keep up.

???

If that were true you wouldn’t see such deviation between framerates across generations at 1440p, especially for raytracing.

Even 1080p shows deviations in some games. At any rate the absolute majority of gamers are in 1080p so these graphs are relevant to, well, almost everyone.

 

[Nitefly]

 

[scubes]

Can't wait for the 6090 to be 100% faster than the 5090, and with Nviidia's logic, cost 100% more, so around £4000.

I'd better start saving now then!

It will more like be 30% ..mote like ..

 

[Grim5]

Digital Shillery are at it again right from the opening statements..

I just skimmed the video with no sound, and noticed three things

1) The subtitle for the first section of the video is called "the phenomenal founder edition card" like wtf, the thing is the new Fermi, its hot and power hungry. How can you call this card great when the GTX480 got bagged for the same thing

2) The first benchmarks in the video is with MFG 4x, 400% more fake frames, yay!

3) When they get to the real results, the 5090 looks like a 4090ti lol

 

[ICDP]

I wish reviewers would drop 1440p testing as it's mostly pointless, with 4090 class and faster GPU's, it's really time to stop benching these cards at less than 4K as they just aren't made for such a low resolution and the CPU's can't keep up.

The lower resolution results help people get some idea on how the various upscaling settings will work. Don’t like them don’t read them.

 

[Vertigo]

Yes - I mentioned about this in my post; the 4080 was not great value but the 4080 super was.

The 4080S FE was one of the best cards Nvidia ever released. Best bang-for-buck in the 40-series range, won't be significantly bested by the 5080 and the all-black finish looks amazing.

 

[Raiden85]

???

If that were true you wouldn’t see such deviation between framerates across generations at 1440p, especially for raytracing.

Even 1080p shows deviations in some games. At any rate the absolute majority of gamers are in 1080p so these graphs are relevant to, well, almost everyone.

Hardware unboxed average frame rate increase with 17 games over the 4090 at 1440p was 12%, at 4K this more than doubled to 27%. So yes, anything below 4K for these class of cards is pointless, it may as well be a CPU benchmark.

17 game average, 5090 over a 4090

At 1080p - a 2fps improvement, so identical.
At 1440p - 12% faster
At 2160P - 27% faster

 

[Grim5]

And for anyone wondering about PSU requirements

I know a few people here have said they plan on using a RTX5090 with their 850w PSU, yeah dont do that.

A couple reviewers tested PSU's and if you use this card with a 850w PSU your PC will crash

 

[Jay85]

Fermi 2.0

And underwhelming as expected

 

[Tinek]

Someone here earlier mentioned some reviewer noticed big drop in FPS with 4090 in CP2077 between old and new DLSS model - I just checked it myself in 1440p UW, drop is negligible. DLSS Quality (no FG) old model 95FPS, new model 93FPS. With FG it's 153FPS (old model) vs 151FPS (new model). Quality is better in many places, but not everywhere - on some (on the road and not only) textures moiré is horrible both with old and new DLSS models in movement, though on new one it looks bit better when up close, yet from the distance just as bad on both. In 4k likely much better.

 

[Halliphax]

Hardware unboxed average frame rate increase with 17 games over the 4090 at 1440p was 12%, at 4K this more than doubled to 27%. So yes, anything below 4K for these class of cards is pointless, it may as well be a CPU benchmark.

17 game average, 5090 over a 4090

At 1080p - a 2fps improvement, so identical.
At 1440p - 12% faster
At 2160P - 27% faster

That 12% difference is tangible because well, it’s extra frames. This is even more relevant if you’re trying to decrease latency to make FrameGen feel nicer.

 

[Vtec9k]

The 5090 FE cooler feels like a backwards step, cost cutting, what happened to bigger = better !? Slightly more efficient design but not offset enough to warrant louder fans.

+5db! https://www.techpowerup.com/review/nvidia-geforce-rtx-5090-founders-edition/41.html

And so ends my review of the card I will never own

That’s loud. Although the GN review said 35db so maybe a bit lower in some situations s.

Still compared to Some of the AIB 40xx cards with quiet bios coming in at 26-30db it’s pretty bad.

 

[BS Dave]

Hardware unboxed average frame rate increase with 17 games over the 4090 at 1440p was 12%, at 4K this more than doubled to 27%. So yes, anything below 4K for these class of cards is pointless, it may as well be a CPU benchmark.

17 game average, 5090 over a 4090

At 1080p - a 2fps improvement, so identical.
At 1440p - 12% faster
At 2160P - 27% faster

1080p=4k DLSS P
1440p=4k DLSS Q

Pretty handy to know now right?

Edit - quoted the wrong person. Sorry Raiden

 

[1988-DD-1988]

As i was stating even before reviews dropped today if anyone on here was planning on upgrading to 5090 to game at resolutions lower than 4k it will be a less worthwhile upgrade coming from a 4090, but if you are gaming at 4k and even those that have 8k monitors and 8k tvs which gonna become a lot more common place now since the new hdmi and display port specifications allow for these resolutions at higher hz now then the 5090 is gonna make a lot of sense for future proofing,also have to take account the newer game engines to come and big hitter games like gta6 and new witcher on the unreal engine will only benefit even more from the newer card, yes it may not seem like a massive upgrade right now on current titles but when those new games drop and on older cards you are having to drop settings and turn features off just to have a playable experience then you will see the true benefit of a more powerful card like the 5090

 

[Capitano]

And for anyone wondering about PSU requirements

I know a few people here have said they plan on using a RTX5090 with their 850w PSU, yeah dont do that.

A couple reviewers tested PSU's and if you use this card with a 850w PSU your PC will crash

makes sense as the 5090 alone is pulling over 500w, only leaving 350 for the rest of the system

 

[Vexr]

Hardware unboxed average frame rate increase with 17 games over the 4090 at 1440p was 12%, at 4K this more than doubled to 27%. So yes, anything below 4K for these class of cards is pointless, it may as well be a CPU benchmark.

17 game average, 5090 over a 4090

At 1080p - a 2fps improvement, so identical.
At 1440p - 12% faster
At 2160P - 27% faster

Us ultrawide users needs the 1440p and 4K benchmarks to try and extrapolate what the performance might be for our resolutions, as reviewers won't benchmark them directly.