Choosing a Gaming CPU at 1440p: Adding in Haswell
by Ian Cutress on June 4, 2013 10:00 AM ESTMetro 2033
Our first analysis is with the perennial reviewers’ favorite, Metro 2033. It occurs in a lot of reviews for a couple of reasons – it has a very easy to use benchmark GUI that anyone can use, and it is often very GPU limited, at least in single GPU mode. Metro 2033 is a strenuous DX11 benchmark that can challenge most systems that try to run it at any high-end settings. Developed by 4A Games and released in March 2010, we use the inbuilt DirectX 11 Frontline benchmark to test the hardware at 1440p with full graphical settings. Results are given as the average frame rate from a second batch of 4 runs, as Metro has a tendency to inflate the scores for the first batch by up to 5%.
One 7970
With one 7970 at 1440p, every processor is in full x16 allocation and there seems to be no split between any processor with 4 threads or above. Processors with two threads fall behind, but not by much as the X2-555 BE still gets 30 FPS. There seems to be no split between PCIe 3.0 or PCIe 2.0, or with respect to memory.
Two 7970s
When we start using two GPUs in the setup, the Intel processors have an advantage, with those running PCIe 2.0 a few FPS ahead of the FX-8350. Both cores and single thread speed seem to have some effect (i3-3225 is quite low, FX-8350 > X6-1100T).
Three 7970s
More results in favour of Intel processors and PCIe 3.0, the i7-3770K in an x8/x4/x4 surpassing the FX-8350 in an x16/x16/x8 by almost 10 frames per second. There seems to be no advantage to having a Sandy Bridge-E setup over an Ivy Bridge one so far.
Four 7970s
While we have limited results, PCIe 3.0 wins against PCIe 2.0 by 5%.
One 580
From dual core AMD all the way up to the latest Ivy Bridge, results for a single GTX 580 are all roughly the same, indicating a GPU throughput limited scenario.
Two 580s
Similar to one GTX580, we are still GPU limited here.
Metro 2033 conclusion
A few points are readily apparent from Metro 2033 tests – the more powerful the GPU, the more important the CPU choice is, and that CPU choice does not matter until you get to at least three 7970s. In that case, you want a PCIe 3.0 setup more than anything else.
Facebook
Twitter
RSS
116 Comments
View All Comments
TheJian - Thursday, June 6, 2013 - link
http://www.tomshardware.com/reviews/a10-6700-a10-6...Check the toms A10-6800 review. With only a 6670D card i3-3220 STOMPS the A10-6800k with the same 6670 radeon card in 1080p in F1 2012. 68fps to 40fps is a lot right? Both chips are roughly $145. Skyrim shows 6800k well, but you need 2133memory to do it. But faster Intel cpu's will leave this in the dust with a better gpu anyway.
http://www.guru3d.com/articles_pages/amd_a10_6800k...
You can say 100fps is a lot in far cry2 (it is) but you can see how a faster cpu is NOT limiting the 580 GTX here as all resolutions run faster. The i7-4770 allows GTX 580 to really stretch it's legs to 183fps, and drops to 132fps at 1920x1200. The FX 8350 however is pegged at 104 for all 4 resolutions. Even a GTX 580 is held back, never mind what you'd be doing to a 7970ghz etc. All AMD cpu's here are limiting the 580GTX while the Intel's run up the fps. Sure there are gpu limited games, but I'd rather be using the chip that runs away from slower models when this isn't the case. From what all the data shows amongst various sites, you'll be caught with your pants down a lot more than anandtech is suggesting here. Hopefully that's enough games for everyone to see it's far more than Civ5 and even with different cards affecting things. If both gpu sides double their gpu cores, we could have a real cpu shootout in many things at 1440p (and of course below this they will all spread widely even more than I've shown with many links/games).
roedtogsvart - Tuesday, June 4, 2013 - link
Hey Ian, how come no Nehalem or Lynnfield data points? There are a lot of us on these platforms who are looking at this data to weigh vs. the cost of a Haswell upgrade. With the ol' 775 geezers represented it was disappointing not to see 1366 or 1156. Superb work overall however!roedtogsvart - Tuesday, June 4, 2013 - link
Other than 6 core Xeon, I mean...A5 - Tuesday, June 4, 2013 - link
Hasn't had time to test it yet, and hardware availability. He covers that point pretty well in this article and the first one.chizow - Tuesday, June 4, 2013 - link
Yeah I understand and agree, would definitely like to see some X58 and Kepler results.ThomasS31 - Tuesday, June 4, 2013 - link
Seems 1440p is too demanding on the GPU side to show the real gaming difference between these CPUs.Is 1440p that common in gaming these days?
I have the impression (from my CPU change experiences) that we would see different differences at 1080p for example.
A5 - Tuesday, June 4, 2013 - link
Read the first page?ThomasS31 - Tuesday, June 4, 2013 - link
Sure. Though still for single GPU, it would be a wiser choice to be "realistic" and do 1080p that is more common (on single monitor average Joe gamer type of scenario).And go 1440p (or higher) for multi GPUs and enthusiast.
The purpose of the article is choosing a CPU and that needs to show some sort of scaling in near real life scenarios, but if the GPU kicks in from start it will not be possible to evaluate the CPU part of the performance equation in games.
Or maybe it would be good to show some sort of combined score from all the test, so the Civ V and other games show some differentation at last in the recommendation as well, sort of.
core4kansan - Tuesday, June 4, 2013 - link
The G2020 and G860 might well be the best bang-for-buck cpus, especially if you tested at 1080p, where most budget-conscious gamers would be anyway.Termie - Tuesday, June 4, 2013 - link
Ian,A couple of thoughts for you on methodology:
(1) While I understand the issue of MCT is a tricky one, I think you'd be better off just shutting it off, or if you test with it, noting the actually core speeds that your CPUs are operating at, which should be 200MHz above nominal Turbo.
(2) I don't understand the reference to an i3-3225+, as MCT should not have any effect on a dual-core chip, since it has no Turbo mode.
(3) I understand the benefit of using time demos for large-scale testing like what you're doing, but I do think you should use at least one modern game. I'd suggest replacing Metro2033, which has incredibly low fps results due to a lack of engine optimization, with Tomb Raider, which has a very simple, quick, and consistent built-in benchmark.
Thanks for all your hard work to add to the body of knowledge on CPUs and gaming.
Termie