Gadgetory

the original Mirror's Edge was one of the most graphics intensive games of its time especially with physics which if you ever broke glass in the game or a helicopter shot at you and broke the glass there was terrible lag for low-end cards and it was all in the physics processing now a Mirror's Edge catalyst is out and we've got the full GPU benchmark on the new game using our benchmark course that we published in yesterday's video and that includes everything up to the GTX 10 series cards 1080 1070 and Andy's fury XO type the high end and the 390 X has sort of our baseline for AMD performance before getting to that all this coverage is brought to you by enter max on their new lick max giant open loop liquid cooling solution so back to the benchmarks we ran the same course three times on each card for parity using a couple different settings 1080 1440 and 4k for the resolutions with high ultra and hyper which is the highest setting for Mirror's Edge catalyst the test system is shown on the screen now you can see all the parts views there but for more information on our testing methodology the article is linked in the description below and contains more depth note that mirrors edge already has game ready drivers from both AMD and NVIDIA the drivers we used for testing are these latest ones as of today and that's the 36 8 39 and video drivers and 16 dot six dots one and the hotfix drivers graphics settings in Mirror's Edge catalyst again go from low to high / and hyper is a step above Ultra the main options include an fov slider for your field of view adjustment and quick note here it goes from 60 to 90 or a hundred and at the very high end of that I'm like some FPS games Mirror's Edge does get a bit distorted on our display so we didn't play with the 90 to 100 highest fov settings we use the lower settings just for a more default setup since that's probably what most people be playing on the usual vsync settings are also present and there's a couple of other important options like GPU memory restrictions and that's one where you'll want to disable it if you're trying to test yourself because the game will dynamically scale graphics based upon active GP performance so disabling is critical for benchmarking which we did you might want this on if using a card with lower memory capacity like two to four gigabytes the rest of the settings are the usual mix of texture quality which is basically texture resolution and post-processing mesh and reflection qualities reflection quality plays a bigger role and mirrors edge than most games given that so many reflective surfaces are present ie mirrors edge post-processing quality is also big as Mirror's Edge catalyst seems to derive most of its visual fidelity from graphics processing happening at the end of the pipeline after rasterization and that's one of the settings you'll want to know for FPS games before getting to the results for 1080p 1440 and 4k testing here's a quick look at some of the research we did so I ran tests from high ultra and hyper at 1080p on a 3 90 X and this was strictly for my own use to figure out how i should benchmark the game how we should run the tests on the different graphics settings the cards but it was worth publishing just to show these scaling of performance across different graphics settings Ultra runs about fourteen point three six percent higher an average FPS than hyper 1080p during our quick research hi runs about thirty eight point four percent faster than ultra and then we have been unable to get hyper reliably executing on some cards like the fury acts and 970 and that instability seems to coincide with devices running four gigabytes of vram device is running six to eight didn't have this issue though the RAM consumption is nearly seven point five gigabytes on 1080p with hyper settings and is around six point six gigabytes at 4k with high settings and then we top five gigabyte utilization on 1080p with ultra settings so it looks like the industry is continuing its push into games requiring more than four gigabytes of vram here's the chart showing 4k benchmarks with high settings the GTX 1080 pushes the highest frame rates at 51 FPS average and this is generally playable and mirrors edge but you have to tweak some settings to sustain 50-plus FPS with better regularity particularly in combat intensive scenarios the fury ex and r9 390 ex both suffer from poor frame times performance and low frame rates at the bottom end of the performance spectrum and the GTX 1070 holds better one percent and point one percent low outputs than these two cards but still sits at just above 38 FPS average here's 1440p the r9 390 X now sits at 53.3 FPS average but still has poor 0.1% low performance the 390 X can generally handle 1440p what would require some tweaks to improve low-end FPS the fury X has these same issues but hasn't improved average at 63 FPS and that puts the fury x at sixteen point six eight percent faster than the 390 x 1.5 seven percent slower than the nine seventy and forty point two percent slower than the 1070 the gtx 1080 1070 and 980 TI are all performing at similar marks in this test due to some sort of bottleneck in the system we're still investigating this but it could be memory or some other limitation I'm not positive yet either way these cards perform effectively the same and performance there's some variants in the lowes but really that could all be wrapped up as some sort of bottleneck within the system and more or less the averages are pretty close to our tolerance for error at 1080p ultra thins loosen up a bit more the gtx 1080 pushes 137 fps and holds a noteworthy lead over the 980 TI hybrid with a 21-point eighty-six percent gap between the two the r9 390 X does much better but still has poor 1% in 0.1 percent low performance The Fury X shows nearly a ninety-five percent gap in its average and 0.1% load dips making for some visible and jarring frame output at times the r9 380 x holds strong at 56.3 FPS average again with gaps in frame times and the gtx 960 runs nine point eighty eight percent slower and it's averages than the 380 x but has 13.8 percent better one percent blows and 41.3 % better 0.1 percent blows the 970 and 1070 hold above a 60 FPS average with the 1070 in 1080 separated by about 18.5 percent in terms of their Delta at 1080p hyper the 1080 runs 29 point five percent slower than on 1080 ultra so that's the GTX 1080 just to be clear on 1080p hyper vs. 1080p ultra the r9 390 x holds at 54 FPS about 25 FPS behind its 17 9.7 average at 1080 altra and the fury ex and 970 were unable to complete this benchmark and exhibited instability that we believed to be a result of their lower vram capacities of 4 gigabytes so then as always a quick conclusion with sort of the best video cards depending on what settings you're looking for kind of your price range at 1080p with ultra graphics if that's your thin the gtx 960 and r9 380 x both do pretty well and their averages with the 960 generally pushing a better one percent and point one percent lower than the 380 x at a mix of medium and low settings if you're willing to go down to there at 1080p you can get away with the 950 or the even though r9 270x if you have that older card in your system but you will be running a mix of low and medium settings to achieve that especially in the combat intensive post of X and physics intensive scenarios at 1440p the GTX 970 is still a good card if that's what you're running I know a large part of the market does have a 970 it's about forty percent behind the GTX 1070 of this particular resolution but still proves pretty price friendly and is good with performance overall the r9 390 X hits about 53 FPS 53.3 FPS average and it's got a good frame rate overall but again has those lower one percent and point one percent low frame rate outputs which is not great because there's some sort of sluggishness in gameplay at times and hopefully that will be fixed with drivers or something else it may even be a setting of the game I haven't gone as far as individually checking every setting but it may be the case that Andy struggles with one of those some very specific settings in the game and that couldn't act the low performance as far as 4k goes even at high settings which is reduced from ultra of course the GTX 1080 does struggle to hit 50 FPS but manages to hit it on average you would want to do some further tweaking to ensure that you stay at that 50-plus FPS mark Mirror's Edge generally doesn't require quite as high frame rate as something like black ops might if you're a competitive player because it's mirrors edge is more of a single player game but all the motion move all the movement all the motion blur all these effects will sort of stack up and if you turn too quickly at times with the lower frame rate like 40 or below it really does impact gameplay so you want to stay in the higher mark 50-plus in this case overall Mirror's Edge catalyst it was pretty consistent in its performance it's one of the easier games to benchmark you can find our benchmark course in the previous video or probably scattered throughout this one but everything 963 idx are better if you're at the 1080p ultra resolution otherwise obviously a higher-end card as always links in the description below for the full benchmark if you want more detail thank you for watching patreon link postural video and subscribe for more information I'll see you all next time