DOOM Eternal Tested on Low-end Graphics Cards

[IHauntedl]

Hi I have a i5-2400 and 8gb ram with the gt 1030 and running at 45-60fps on low to med settings and resolutions I play at a 1280x720 and 1366x768 for doom eternal so its pretty playable im up to the part where you get the crucible blade

 

[Reehahs]

Ed, Geforce GTX 1650 [4GB] appears twice in the first graph (RS-off).

 

[PEnnn]

I made a mention of that in another bench article. It is well known that Nvidia stops optimizing the older products so they fall further and further down.

Another sneaky way for nGridia to sell more cards.

 

[PEnnn]

My 960 GTX was able to handle (high settings) all Crysis releases. I bet it can handle any Doom flavor as well!
But maybe it's time for the 2080 Super and be happy for the next 6-8 years.

 

[mileston3]

I'm really interested to see what is the level I can reach with my Sapphire R7 260X OC. ?

 

[GreenNova343]

Exactly, these kinds of comparisons are essentially meaningless without a realistic matching cpu.

If you pair a low-end GPU with a low-end CPU, then how do you know which one is limiting your performance, & what kind of inexpensive upgrade would give you the most boost?

Techspot, Tom's Hardware, & all of the reliable testing sites have had to reiterate this for years:
-- to eliminate GPU bottlenecks when testing CPU performance, you test all of the CPUs with the top-end GPU available at the time of testing.
-- to eliminate CPU bottlenecks when testing GPU performance, you test all of the GPUs with the top-end CPU available at the time of testing.

That way, the ONLY bottlenecks that show up in the test results will be due to the GPUs themselves -- the generation of card, the amount of VRAM, its core/processor count vs. other chips of its same generation & maker, etc.

 

[Linux7055]

To be honest I don't think its fair to say Nvidia cards age poorly in general. Why?

1. This is a game that GCN takes full advantage of Vulkan, so im not really surprised Maxwell and Pascal don't shine as bright as AMD, but they do not fall significantly behind AMD. Steve revisited Maxwell cards last year and he found the 970 to be very competitive with the 290 and the 980ti doing well against a 1070 in most modern titles. I expect to see more games implement Vulkan, but it won't be a mainstream api, at least not in the near future. Turing has much better Vulkan support than Pascal or Maxwell, so I expect it to age better than any Nvidia architectures we've seen in the past decade.

2. Kepler is the real outlier here. The 970 matches the 1060 3gb and RX 570 4gb, but is over 50% faster than the 780ti, a card that was just as fast back at launch. Maxwell launched about 6 years ago and is still proves to be relevant architecture today. While Kepler 700 series already began to fall after only 2 to 3 years after launch. I am convinced Kepler was just a poorly designed architecture that even Nvidia has trouble optimizing for. Unfortunately Kepler's missteps was what sparked the AMD finewine talk we hear about everyday.

What do you think @Steve ?

 

[EClyde]

I'm getting 60 or so on my R9 280 1920x 1080 and it's playable and looks good. I will see soon what I get on the GTX 1660. I am capped by my monitor at 75 with the 1660

 

[jpuroila]

Fascinating how AMD cards age significantly better than NVIDIA cards do. Either it's NVIDIA failing to support older hardware with optimized drivers while AMD does, or perhaps AMD's architecture is simply more future proof.

I'd say it's a bit of both(though as Linux7055 says, Kepler is something of an outlier). In general, you get more raw GFLOPS from AMD cards which translates into better performance when you don't get driver optimizations.

 

[Bas Keur]

That 'Gameplay' vid was made by grandma, should have used this one.