5700XT is sliding down from its 2070 Super comparisons to rtx 2060
It was always at 2060 level
https://www.techpowerup.com/review/amd-radeon-rx-5700/28.html
Not counting DLSS or RT even.
Wait till HUB finds out Mesh Shaders reduces vram usage.
Does anybody know why Polaris and VEGA perform so miserably compared to RDNA1?
Because they are even older then?
The newer athitecture of the RDNA1 still pulling its weight here even without Mesh Shaders.
Pascal GPU’s that are faster on paper but older like the 1080ti are DOA in this game though.
It helps it not running the same settings as the 2060s
https://www.tomshardware.com/features/alan-wake-2-will-punish-your-gpu
While we previously reported that Alan Wake 2 requires mesh shader support, it turns out Remedy implemented a fallback path for older GPUs that seems to work okay. Performance isn’t great, and one of the graphics options (post-processing quality) gets grayed out, which means certain computations aren’t done. We noted the game looked blurrier and perhaps grainier as well on the RX 5700 XT. But it was still technically playable, at 1080p with FSR2 Quality upscaling, if you want to give it a shot. Average fps was tolerable at medium settings, but 1% lows were only quite poor and there was noticeable stuttering and hitching. Perhaps that’s due to the drivers, or more likely just the workarounds for the lack of mesh shader hardware.
I mean thats fair.
I just don’t get why my comment got downvoted into oblivioun haha. I was comparing RDNA1 and Pascal.
5700XT doesn’t have DX12U, but it embraced aspects of DX12 much more than Pascal. Then again, pascal is much older than 5700XT and I remember how back then, a game being native DX12 was an event, much like a game being fully path traced
l.e.t .m.e. g.u.e.s.s .m.r. .a.m.d .s.a.y.s. a.m.d .wi.n.s
30 fps with medium settings at 1080p isn’t too far away from what consoles do, which target 1260p instead. Not sure why thinks it’s “unplayable”.
And mesh shaders do help, seeing how the pascal titan delivers a truly unplayable framerate.
It’s not playable with mouse input.
Use a controller then?
Peasantry inaccurate mess of an input device on a desktop computer? lol
especially in a “shooter” game XD
I remember I dropped first Alan Wake only because it was insane to aim with 15 fps on my old laptop that times - and it’s not cod with infinite ammo/healing. At my final try I decided to shoot “as is” to fight back disregarding misses, eventually(and obviously) ran out of ammo and deleted the game. Had I have gamepad instead of mouse it’d be easier to delete the game right away not even starting.
Having my whole life with PC(literally - from 5y.o. or so) why in the hell would I abandon perfect-aiming with vast controls for something “middle” of both which doesn’t do good any of those tasks? XD
It’s unplayable because that’s his bias against nvidia
I thought Turing and newer supported Mesh Shaders?
turing and newer does (including 16-series).
5700XT doesn’t. Technically 5500XT does tho, it’s a later release with some additional tweaks.
Hard to view 5700XT as anything other than RDNA 0.9 given that it also had constant stability problems that are resolved in the 5500XT too.
I ran one for a year and had no stability problems. Hot spot temperature issues heh, but no crashes. Fewer even than your normal nvdkm.dll crashes Nvidia has always had.
Do RX 5500XT support mesh shaders? I can’t find any benchmarks with this graphics card.
I like HU’s trivial tests like this squeezing to production.
“2060 super can’t deliver a playable experience”
well, if you turn on DLSS Balanced or Performance it probably is real close to 60fps. If you are willing to take the image quality hit from FSR2 Quality mode, the DLSS Balanced or Performance mode probably is about the same, but it gets you over the top on framerate.
Dunno why there’s a lecture about “how NVIDIA fans shouldn’t want DLSS in benchmarks”, the relevant comparisons at iso-image-quality are still FSR2 Native vs DLSS quality and FSR2 Quality vs DLSS Performance these days. HUB studiously ignores the requests for normalizing visual quality and just delivers lectures instead lol.
“5700XT wouldn’t deliver a playable framerate even if it had mesh shaders”
umm considering 5700XT normally performs around 2070 Super performance levels, and the 2070 Super does fine, I don’t see why the 5700XT wouldn’t be delivering playable framerates if it had mesh shader support. Take those 2060S numbers and add another 20-30% (for the step to 2070S performance) and that’s where the 5700XT-MESH would land. You really think that wouldn’t be over 60fps?
I dunno why reviewers still have a stick up their butt about early cards NEVER being able to do relevant things with DX12U or RTX, even the 2060 and 3050 can very much do useful things, it just seems like either editorial bias or a personal inability to admit that the data doesn’t support your thesis.
Like yeah if you bought the lowest-end RTX cards you’re gonna have a low-end experience in graphics showcase titles, but we’re still talking about the 2060S most likely hitting over 60fps with DLSS Performance, at no greater visual quality hit than people are willing to accept with FSR2 quality, in a title where even “low” settings is pretty fantastic. And the enthusiast cards, the 2070S and so on, certainly the 2080 Ti, definitely do deliver relevant experiences even today.
I dunno why reviewers still have a stick up their butt about early cards NEVER being able to do relevant things with DX12U or RTX,
I think it was more about the discourse at the time. They spoke at length about it more on their podcast. When the super series came out, NVIDIA said it was a requirement for RT and you’ll get left behind etc (and DLSS 1 was awful). People waited ages for RT to come out, and when it did that series gets absolutely smoked. People who brought 5700XT’s were told they brought an obsolete product because it wouldn’t be able to do RT. The truth is neither of those cards can in a meaningful way. It’s more a jab at NVIDIA at the time selling a bit of a porkie to the gamers.