GeForce RTX 3080 Rips and Tears Through DOOM Eternal at 4K, Over 100 FPS

[Deleted member 185088]

It depends. Note the dive 2080 is taking at 4k (also tells you why those nice DF guys ran it that way):

View attachment 167815

"speed up":

View attachment 167818

That's huge I missed this, it reinforces my doubts on getting the 3080, and rumour has it that RDNA2 would have 16gb in GDDR6 and HBM2 (the real fastest graphic memory).

 

[VolutedJoker]

As someone who was / is totally looking forward to getting the 3080 I just wanted to say Doom Eternal is not THAT demanding. I'm currently running a Ryzen 3900X with a Sapphire 5700XT (PCIE 4 MOBO) and I'm getting 4K / 120, on Ultra Nightmare with 2G vram left to spare. Oddly I'm avg 85-100FPS on Fortnite. Oh HDR off for both comparisons.

Also concerned about the 3080 only have 10G vram, but we shall soon see if it's enough. Perhaps with PCIE 4 this is not as much as a bottle neck as with 3.0 ? It's funny that AMD is potentially releasing a card that falls between the 3070 / 3080 BUT with 16G vram at a lower price. So either vram won't matter much is something not right in the kool aid.

Let me know your thoughts, please keep comments polite and positive

 

[ThrashZone]

Hi,
If 3080ti has double the vmem as 3070 8gb to 3070ti 16gb shows 3080ti would be best with maybe 20gb vmem.

 

[efikkan]

… and I'm getting 4K / 120, on Ultra Nightmare with 2G vram left to spare.

Just FYI; allocated memory doesn't necessarily mean needed memory. Some games and even the driver can allocate a bit more than is strictly needed.

They way you check what's actually needed is through benchmarks, especially frame time consistency. Once you have too little VRAM, you'll get stutter, as the driver has to swap data between VRAM and RAM. When you are approaching the limit you will start to see occasional stutter, and when you push beyond that it will get pretty unplayable very quickly. It's not like you will get a 5-10% drop in performance, it will be far more severe than that if it's actually out of VRAM, in some cases you can even get glitching, texture popping even missing textures.

Perhaps with PCIE 4 this is not as much as a bottle neck as with 3.0 ?

PCIe 4,5,6… will not help with too little VRAM, the problem with swapping is latency, not bandwidth.

 

[VolutedJoker]

Just FYI; allocated memory doesn't necessarily mean needed memory. Some games and even the driver can allocate a bit more than is strictly needed.

They way you check what's actually needed is through benchmarks, especially frame time consistency. Once you have too little VRAM, you'll get stutter, as the driver has to swap data between VRAM and RAM. When you are approaching the limit you will start to see occasional stutter, and when you push beyond that it will get pretty unplayable very quickly. It's not like you will get a 5-10% drop in performance, it will be far more severe than that if it's actually out of VRAM, in some cases you can even get glitching, texture popping even missing textures.


PCIe 4,5,6… will not help with too little VRAM, the problem with swapping is latency, not bandwidth.

Ah ok, good info thanks. I did use the metric tool built into Doom, gave some useful info, some was over my head lol. Thanks for the reply just the same.

 

[HugsNotDrugs]

No. Unoptimized textures take up a lot of space. There are games with beautiful textures that use 4-5 GiB of VRAM in 4K. The games that fill up the VRAM completely usually have ugly textures and lowering their quality does not make much difference anyway.

No modern engines use un-optimized textures. I think where things vary is how the engine stores textures in VRAM that aren't immediately in use.

The fast storage systems in the next gen consoles guarantees fast access and as such only textures immediately in use need to be loaded, vastly reducing the VRAM usage. Without guaranteed fast access more texture data needs to be sitting ready in VRAM so that gameplay isn't interrupted due to slow storage speeds.

 

[Deleted member 185088]

No modern engines use un-optimized textures. I think where things vary is how the engine stores textures in VRAM that aren't immediately in use.

The fast storage systems in the next gen consoles guarantees fast access and as such only textures immediately in use need to be loaded, vastly reducing the VRAM usage. Without guaranteed fast access more texture data needs to be sitting ready in VRAM so that gameplay isn't interrupted due to slow storage speeds.

While SSD storage is fast, but ram is another level in both speed and latency.

 

[d0x360]

More VRAM doesn't give you more performance.

How on earth would a 290X run Doom (2016) in 4K at ultra at 100 FPS?

RTX 3070/3080 is carefully tested, and has the appropriate amount of VRAM for current games and games in development.

More vram can absolutely improve performance... Also you can believe me or not but the 290x can absolutely run doom 2016 at 4k max settings over 100fps. You realize vulkan is based on amd's mantle right?

In the game thief running dx11 I would get just under 60fps in dx11. In Mantle I would get around 140 fps.

If I still had my 290x I'd make a video and show you but it's pretty common knowledge that amd has better performance in vulkan and dx12 than nvidia. Even turing wasn't very good at dx12.

In dx11 mode shadow of the tomb raider would run at 60fps but with the same settings in dx12 it would run 6-10% slower.