Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Gaming on M395X: I'm satisfied
- Thread starter MandiMac
- Start date
- Sort by reaction score
The source was in my post for the GPU numbers. If you mean numbers showing a 4Ghz i7-6700K Skylake is more than 16% faster than a 4Ghz i7-4790K Haswell, they don't exist because it's not. If Intel could improve IPC by more than 16% in a single architectural generation it would be revolutionary. Alas, they cannot. The point is the GPU on the top-spec iMac improved more from 2014 to 2015 than the CPU did, yet few are complaining about the CPU being a minimal improvement.
No, I mean a source that actually shows numbers, rather than an aggregate of things from who knows where. The only test I've seen that I'd consider a reliable (as far as that goes) source is barefeats.
http://barefeats.com/imac5k13.html
And for CPU benches, it does look like the new CPU is faster (as it should be).
http://barefeats.com/imac5k16.html
Bottom line is the GPU "upgrade" still sucks.
And yes, as someone else pointed out above, the M395X is not in the same league as a 980M, despite it showing as being exceptionally close on the link you posted. I'd take that site with a rather sizable grain of salt.
Last edited:
Just played more, on the outside if there are really many polygons to be rendered, the FPS drop down to 30. Never below, even in heated firefights. Inside buildings and sometimes on the highway you're getting 60 FPS. Mostly between 40-50, though.
Sorry, good point, but, sadly, completely misses the issue.
The i7-6700K is the most powerful consumer-grade CPU on the market right now - I expect nothing less in a high-price-AIO Apple advertises as "the ultimate desktop experience". So, Apple scored.
However, the GPU, even if it improved a bit more in comparison, still runs too hot (--> noise issues, longevity concerns) and seems to be inferior in performance both to Nvidia's devices, at least in some features, AND in comparison to the next generation which seems to hit the market very soon. That's disappointing and, frankly spoken, far away from 'ultimate'.
Besides, CPU performance isn't that important anymore. I guess even my old i3 Clarkdale goes to sleep mode between keystrokes - and I'm a fast typist using 10 fingers. ;-O
Why? Because you said so?
PassMark's rankings are an aggregate of user-submitted benchmarks run through their suite, PerformanceTest. They have no interest that I can discern in propping up the M395X nor in denigrating the 980M.
A possible avenue of critique is that some specific aspect of the suite of tests they use unreasonably favour ATI’s cards in general or the M395X in particular, but at least in the broad case that doesn’t seem true, since it’s mostly nVidia cards on the top of the stack.
Alternately, and more reasonably, you might attack the fact that the M395X is represented by a sample of exactly 1, compared to 466 samples of the 980M’s performance. It’s possible – maybe not probable – that the one guy who’s benched his iMac under Windows with the PassMark suite took the thing apart and overclocked the GPU while spraying liquid nitrogen on it, or maybe just got his hands on a really high-quality sample. We can certainly see that more testing needs to be done, because one sample doesn’t a conclusion make; but it is instructive.
You might also opine that ATI’s cards consistently over perform in benchmarks relative to real-world conditions, either because of some design flaw or bad drivers or whatever. Evidence to back up such an assertion could be reasonably compelling.
It’s long been an Internet Suspicion that ATI is binning these things for Apple, largely to ensure they don’t all die ATI HEAT DEATH, so I wouldn’t expect a significant amount of variance in subsequent sample benchmarks (I would assume that the binned chips are likely to all perform similarly if they’re being selected for similar characteristics.) But that’s my own WAG.
What we do know is that in the specific set of tests performed by PassMark’s PerformanceTest suite, the single M395X sample tested to date performed within 4% of the aggregate performance of 466 samples of the 980M
For the record the M295X only has 23 samples.
I thought I would throw my 2 cents worth in here.
I just received my first iMac, (i7, 16gb RAM, 512 flash, M395X). I've bootcamped it to play several games. I don't really run benchmarks or anything like that, as it doesn't really interest me. However, these are my "findings".
Final Fantasy XIV - cannot play the DX11 client as it crashes after several minutes. The DX9 client runs fine. Playing in Full Screen on Laptop High settings, 1440p, frames limited to 60 fps, the game runs very smooth, almost always sitting at 59-60 fps. I've also bumped settings up, and found that the fps start to dip from there.
SWTOR - Full screen windowed - 1440p, high settings, V-Sync, buttery smooth at 60 fps, never shifts down from that, not a single twitch.
Guild Wars 2 - Full screen - 1440p, high settings, V-Sync, very smooth as well, sits pretty much at 60 fps even with a lot of characters on screen during large battles.
The GPU temp hits about 85oC playing these games, the fan kicks in (which I cannot hear over the game) and the temp goes down to around 75oC.
Overall I am very happy with the system for my casual MMO gaming.
I just received my first iMac, (i7, 16gb RAM, 512 flash, M395X). I've bootcamped it to play several games. I don't really run benchmarks or anything like that, as it doesn't really interest me. However, these are my "findings".
Final Fantasy XIV - cannot play the DX11 client as it crashes after several minutes. The DX9 client runs fine. Playing in Full Screen on Laptop High settings, 1440p, frames limited to 60 fps, the game runs very smooth, almost always sitting at 59-60 fps. I've also bumped settings up, and found that the fps start to dip from there.
SWTOR - Full screen windowed - 1440p, high settings, V-Sync, buttery smooth at 60 fps, never shifts down from that, not a single twitch.
Guild Wars 2 - Full screen - 1440p, high settings, V-Sync, very smooth as well, sits pretty much at 60 fps even with a lot of characters on screen during large battles.
The GPU temp hits about 85oC playing these games, the fan kicks in (which I cannot hear over the game) and the temp goes down to around 75oC.
Overall I am very happy with the system for my casual MMO gaming.
I suspect that may be true, if for no other reason than nVidia seems to do a way better job with optimizing their drivers for real-world results. I'm just saying that I don't find opinions presented as fact without at least some data particularly compelling. At this point it's awfully difficult to find any side-by-side, all-other-factors-accounted comparisons between the two (or really, any rigorous tests of the 395X at all beyond some forum posts, barefeats comparison, and some synthetic benchmarks), so I'm not willing to come to any conclusions.
Last edited:
Man, is that ever wrong. In Lightroom it takes 2 min 15 sec just to import and render 1:1 previews on 30 photos from my Sony A7RII. That is on a top-spec 2015 iMac. This process is highly CPU-bound, not I/O or GPU-bound. On some projects I take 1,000 photos per day. That is only stills, not 4k video which is even worse.
Civilization V runs beautifully at maximum settings on 5K on the 2014 iMacs, you'll be fine.
Does FFXIV crash both in full screen and windowed dx11?
How is the fps in 1080p ultra and high desktop?
Nope, only 1440p on almost ultra and it runs good enough. No crashes on GTA V so far.
Small update: It seems to depend on the heroes. Played two rounds without a drop, and when playing Singed with his Poisoned Trail the FPS vary between 55 and 60. Noticed a half-second-long drop to 40 FPS in a teamfight in that game, but nothing serious...
1440p, Very High, Fullscreen, OS X.
Here's why many of us go Hack.
See...."the red bar" on every single chart.
http://barefeats.com/imac5k17.html
See...."the red bar" on every single chart.
http://barefeats.com/imac5k17.html
Wasn't really commenting about gaming specifically at 5k (those particular benchmarks just nicely highlighted the next point...) - Merely that a Hack can always smoke the Apple offerings. Just sucks that Apple isn't even remotely competitive no matter what price you pay them, as many would love an Apple designed & offered screamer!
I see what you're getting at. Many would say the current iMac is a screamer - be it at photo manipulation, clipping videos, as a house computer and so on. It just so happens that it can't do one thing perfectly, which is gaming.
And remember, Apple is the company for the 90 %. And hardcore gaming is sure as hell nowhere these 90 %. And gaming on Mac is an even smaller subset: That's what the PC route or Hack route is for. If you're paying $5000 for an iMac with a desktop GTX 980, the chance someone buys that is immensely low. Hence I understand why the iMac is the compromise it is: Good enough for 90 % of us, but not for the hardcore gamer.
For light and medium gaming, nothing is to be held against the current iMac.
It's more than gaming - ANYthing that's GPU intensive will be way better on a machine that packs a 980ti (in this example)
The frustrating part is that Apple has *nothing* competitive with a Hack's performance on GPU intensive tasks, no matter how much money you pay them. That's my main takeaway point that is a bummer.
I'd *love* to have a legit Mac tower/pro that could take the latest and greatest GPUs!