OS X Mac vs Windows Performance Benchmark Video

[TheGreatWumpus]

Yes, this is on my channel, but I didn't put it together, a contributor did. I think its a decent overview of some AAA games and benchmarks between the two operating systems:

https://www.youtube.com/watch?v=hRfqNuyyPvQ

 

[numberfour]

Great video mate, now I don't have to compare for myself. I'll just stick with Windows (bootcamp) for gaming. Thanks for sharing

 

[Dirtyharry50]

A 2011 MacBook Pro with a 6770M Radeon connected to a Cinema Display is the hardware you chose to use to study gaming performance in 2015? That GPU was inadequate to drive that display at 1440p (for gaming) when it was a new model which is why it was lowest one available for a 2011 27" iMac while the 6970M was the GPU to get if you planned to play games on a 27" display at native resolution. I would expect the 6770M performed very well at the MacBook Pro's native resolution on a 17" screen assuming you got the late 2011 model that offered that GPU.

No matter how you slice it, while this yields unsurprising results it is not representative of what is possible with a Mac that is more suitable for gaming from a newer MacBook Pro on up to a current iMac and Mac Pro.

It would have been far more interesting to see current models of those three Macs running the tests where I suspect the difference in performance would be less of an issue for the tested games and anything newer you could get your hands on with benchmarking. For example, does Tomb Raider include a benchmark? Even if it does I presume the test system could not run it acceptably and that was why it was left out?

Basically, this gives some idea to Mac users with older Mac portables and Minis what the differences are but does not show what newer Macs that are more suited to gaming perform like.

What I am talking about in more specific terms is, if say Hitman performs great on current MacBook Pros, iMacs and of course Mac Pros, the fact that it runs even faster with Windows is moot isn't it? A better example would be to see how that plays out with something like Tomb Raider or perhaps CoD 3 if it has a benchmark, etc.

The way you did and presented the study was nice but again, the results would have been more useful at the minimum and probably more telling if newer hardware had been tested as well for comparison. Leaving the test system you chose in the mix as the lowest tier for testing would have been fine but no test using just one system is adequate for this. Without testing others you do not have the data to see what difference newer hardware and drivers for it make.

Sorry to be tough on you but I led a QA group in software development where we did performance testing along with everything else we needed to test. I'd never be satisfied that a single test system would tell the whole story for testing like this. As I've said, there is just way too much data missing to form any truly trustworthy conclusion beyond the expectation that many titles running on Windows are going to run faster although not always. How much faster is the question and this study doesn't answer it for many Mac gamers using Macs more suited for gaming than the test system is.

Another thing to be gained from a more comprehensive study would be to educate Mac consumers on what to expect for performance at various product tiers, GPUs, etc. The line is small enough that you could pick the current model year and test across the board which would be very interesting info for Mac gamers.

I am not always so good at delivering the news so please accept my comments in the spirit in which they are intended, as constructive criticism. Please consider this a suggestion for a more in depth followup. That would be great and interesting to compare to the first round of testing as well that you just did.

 

[GoldfishRT]

I did the video so maybe I can help clear some things up because I just think you misinterpreted what this video is supposed to be.

So, to your first point. This video was not to try and demonstrate what modern Macs are capable of in terms of games. If that were the case I would have gone out of my way to find a modern new system. Tomb Raider did have a benchmark and was included in the video if you had watched it and it did run it just fine (within reasonable expectations).

This was simply to test whether Windows performance was generally better than Mac OS still and to what degree. And I absolutely agree. I would have loved to have more systems to test on. The bigger the sample size the better. As it stands though I'm only a college student with so much money and so many friends. Understand that all of us contributing to MacGameCast are just regular guys with regular jobs trying to do some writing and fun stuff regarding games as we can. I would absolutely love, funds ever permitting, to go through every refresh and bench the improvements in performance but that simply isn't possible at this time and that wasn't the point of the video.

It's also worth noting that yes, the highest end MacBook Pro and most of the higher end new machines will perform better than this 6770m equipped notebook from 2011. However, if you head on over to the Steam Hardware & Software survey you'll find that most people on Steam with Macs are running dual core 13" MacBook Pros that still perform worse than this aging machine. 20% of them are using systems with Intel HD Graphics 4000 and the Iris 5100 machines are at 10%. The first dedicated card is a 6750m at 3.85% of users surveyed.

I absolutely welcome the criticism and I hope to be able to do a follow up in a fashion similar to your description at some point.

 

[Dirtyharry50]

I did watch the video in its entirety. I guess I am getting older than I thought I was to somehow manage to miss the Tomb Raider results. Sorry about that.

Something I was trying to get at was the whole idea of what degree the performance is better. You measured that on some fairly old hardware in gaming terms and it isn't necessarily representative of whatever the current gap might be because of that. For example, other than the Mac Pro, the systems are Nvidia for dedicated GPUs and therefore a different driver as well. How would this play out? The video doesn't go there.

In any case, the demonstrated point that the difference in performance on some games is close enough to call into question whether spending for Windows and allocating disk space to it is really worthwhile is well taken. Personally, I do not think it is on my late-2013 27" iMac. It's close enough for me to not care.

 

[TheBunny]

.

So when we were testing ESO at work
side by side windows and mac.

Our OpenGL performance was about same on OS X
and Windows using the same GL 2.1 API on both.

Our DX 11 renderer was faster then OpenGL.

Our DX9 renderer is about the same speed as OpenGL.

I doubt any performance tests are really comparing hardware so much as tech stack differences and maybe drivers.

 

[Dirtyharry50]

So when we were testing ESO at work
side by side windows and mac.

Our OpenGL performance was about same on OS X
and Windows using the same GL 2.1 API on both.

Our DX 11 renderer was faster then OpenGL.

Our DX9 renderer is about the same speed as OpenGL.

I doubt any performance tests are really comparing hardware so much as tech stack differences and maybe drivers.

Within any given hardware range I'd agree but in this case, any newer Mac hardware with a dedicated GPU would be using drivers from Nvidia and they would be newer as well. That was certainly an important part of what I thought was problematic with the video's analysis.

Thanks for sharing that information by the way. That's interesting, particularly the lack of difference with DX9.

 

[VoR]

OpenGL, HFS (and for gaming, the mouse acceleration) are OSX' biggest downfalls.

Unfortunately, OSX really is quite slow 🙁

http://www.phoronix.com/scan.php?page=article&item=osx-fedora21-vivid&num=1

 

[JordanNZ]

OpenGL, HFS (and for gaming, the mouse acceleration) are OSX' biggest downfalls.

Unfortunately, OSX really is quite slow 🙁

http://www.phoronix.com/scan.php?page=article&item=osx-fedora21-vivid&num=1

Slow at WHAT though?... Phoronix doesn't exactly pick apps, or games for that matter to benchmark that the average person uses on a daily basis.

 

[Janichsan]

Slow at WHAT though?... Phoronix doesn't exactly pick apps, or games for that matter to benchmark that the average person uses on a daily basis.

The whole thing is further skewed by not using the same versions of the compilers they used for testing on all systems: for example, they used gcc 4.2.1 on OS X, but 4.9.2 under Linux.

 

[edddeduck]

OpenGL, HFS (and for gaming, the mouse acceleration) are OSX' biggest downfalls.

Unfortunately, OSX really is quite slow 🙁

http://www.phoronix.com/scan.php?page=article&item=osx-fedora21-vivid&num=1

Phoenix are pretty pro Linux, if you look closely the tests are more optimised for Linux than OS X. Linux is still quicker in many aspects of OpenGL however those tests don't look like they are completely equal footing on both platforms.

Janichsan for example mentioned compiler differences.

 

[VoR]

Phoenix are pretty pro Linux, if you look closely the tests are more optimised for Linux than OS X. Linux is still quicker in many aspects of OpenGL however those tests don't look like they are completely equal footing on both platforms.

Janichsan for example mentioned compiler differences.

Slightly confused by the last couple of posts.
PTS isn't optimised for anything and the configurations are clear, public and available for you to use yourself. Compiler versions (funny how neither of you mentioned llvm?) are both completely irrelevant and exactly the point in testing like this (I'll let you figure that one out...). Well done for spotting the .1 though, I must have missed the '100's of % generic application performance increase' changelog in 2007 🙂

This test is the epitome of 'equal footing' for the entire graphics stack, and cpu/disk throughput. If we're doing the weird justification stuff that some people love to do, just concentrate on the fact it's a benchmark.

 

[goMac]

Our OpenGL performance was about same on OS X
and Windows using the same GL 2.1 API on both.

Our DX 11 renderer was faster then OpenGL.

Our DX9 renderer is about the same speed as OpenGL.

GL2.1 is the same generation of renderer as DX9, so it's not terribly fair to pit it again DX11. We're up to OpenGL 4 now, which includes performance enhancements.

I'm not saying the Mac would win an OpenGL 4 vs. DX11 benchmark. But OpenGL 2.1 vs. DirectX 11 is just pitting something ancient against something brand new.

 

[Dirtyharry50]

Slightly confused by the last couple of posts.
PTS isn't optimised for anything and the configurations are clear, public and available for you to use yourself. Compiler versions (funny how neither of you mentioned llvm?) are both completely irrelevant and exactly the point in testing like this (I'll let you figure that one out...). Well done for spotting the .1 though, I must have missed the '100's of % generic application performance increase' changelog in 2007 🙂

This test is the epitome of 'equal footing' for the entire graphics stack, and cpu/disk throughput. If we're doing the weird justification stuff that some people love to do, just concentrate on the fact it's a benchmark.

You seem a little peeved and I don't want to make it worse but in all these sorts of conversations I often wonder to myself what for many games is really the problem? I'm nearly done playing a fairly recent title called, "The Darkness II" which is a "cider port" that has been running flawlessly for me and it's been great fun. Now, I don't consider anything done with cider to be ideal of course and yet this works well, so good enough. I suspect a native port would be better but in my own case not subjectively so. It already runs well at max.

I don't know what the fps is and I don't care. It seems very smooth and doesn't bother my eyes. The screen is not tearing. I have the settings on max. It looks great and the game is a good time.

So in all the performance and fps conversations I see here and elsewhere I often wonder why the obsession with performance unless it is terribly bad and if it is I expect the hardware is not up to the task of playing a particular game or the settings have been pushed higher than a given hardware setup can support.

I was playing on a late-2013 iMac 27" and so far everything I play on it has been a good experience.

Anyway, I just don't get all the angst over this stuff. I think people would do well to forget about benchmarks and fps counters personally and just adjust something to run well being reasonable about what a given hardware setup is capable of.

OS X is very slow seems pretty subjective in this discussion to me. I don't think so at all. I'm not having any performance issues here.

 

[VoR]

You seem a little peeved and I don't want to make it worse but in all these sorts of conversations I often wonder to myself what for many games is really the problem? I'm nearly done playing a fairly recent title called, "The Darkness II" which is a "cider port" that has been running flawlessly for me and it's been great fun. Now, I don't consider anything done with cider to be ideal of course and yet this works well, so good enough. I suspect a native port would be better but in my own case not subjectively so. It already runs well at max.

I don't know what the fps is and I don't care. It seems very smooth and doesn't bother my eyes. The screen is not tearing. I have the settings on max. It looks great and the game is a good time.

So in all the performance and fps conversations I see here and elsewhere I often wonder why the obsession with performance unless it is terribly bad and if it is I expect the hardware is not up to the task of playing a particular game or the settings have been pushed higher than a given hardware setup can support.

I was playing on a late-2013 iMac 27" and so far everything I play on it has been a good experience.

Anyway, I just don't get all the angst over this stuff. I think people would do well to forget about benchmarks and fps counters personally and just adjust something to run well being reasonable about what a given hardware setup is capable of.

OS X is very slow seems pretty subjective in this discussion to me. I don't think so at all. I'm not having any performance issues here.

I'm not really peeved, just frustrated at the misinformation from people trying to justify the poor performance (and for what?..). Ditch the ridiculous brand loyalty and demand more.
It's not a new thing and it's a little painful to swallow for us users, but OSX really is (and has been for a fair while) quite a poor performer in key areas when compared to Windows and Linux - Especially poor when OSX is tightly tied to hardware and the alternatives are massively generic.

As you rightly distinguished though, these are "just" benchmarks - There's no reason to say that you can't happily play a AAA game/churn through data and it's not saying anything about the arguably excellent user experience.

 

[JordanNZ]

So when we were testing ESO at work
side by side windows and mac.

Our OpenGL performance was about same on OS X
and Windows using the same GL 2.1 API on both.

Our DX 11 renderer was faster then OpenGL.

Our DX9 renderer is about the same speed as OpenGL.

I doubt any performance tests are really comparing hardware so much as tech stack differences and maybe drivers.

Hi Chris. How's the new 3.2 Renderer coming along? 😉

 

[Agent-J]

Apple doesn't really have any interest in gaming, or else they'd put something besides the feeble Intel graphics in their machines. Looking at PassMark's benchmark comparisons, I see that my 2012 MacBook Pro Retina uses Intel HD4000, with G3D benchmark value of 455, vs. the 6900 of the R9 290X I have in my gaming PC. Intel Iris Pro 5200 is only 1210.

 

[Irishman]

Apple doesn't really have any interest in gaming, or else they'd put something besides the feeble Intel graphics in their machines. Looking at PassMark's benchmark comparisons, I see that my 2012 MacBook Pro Retina uses Intel HD4000, with G3D benchmark value of 455, vs. the 6900 of the R9 290X I have in my gaming PC. Intel Iris Pro 5200 is only 1210.

Apple really has an interest in gaming, because they put something besides the feeble Intel graphics in their machines. I see that my late 2012 iMac 21.5" uses the nVidia GeForce 650M GT gpu.

See what I did, there??

 

[TheBunny]

GL2.1 is the same generation of renderer as DX9, so it's not terribly fair to pit it again DX11. We're up to OpenGL 4 now, which includes performance enhancements.

I'm not saying the Mac would win an OpenGL 4 vs. DX11 benchmark. But OpenGL 2.1 vs. DirectX 11 is just pitting something ancient against something brand new.

That was kind of the entire point of my comment.

A lot of ports to OS X are DX9 not DX11...

and many use GL 2.1 still...

You can't trust that a Mac Vs Windows version are using comparable tech.

Anyhow OpenGL on OS X has always seemed to be around 20% slower
then DX for whatever reason.

Another thing to note.
GPUs on mac vs pc.

The same GPU in an apple laptop may not be running at the same clock speed as in a windows machine. Apple likes to make machines that run quiet, require less cooling etc.
The same might be true for the newer iMacs.




----------

Hi Chris. How's the new 3.2 Renderer coming along? 😉

Sigh... on hold as the console ports got in the way...

They wrap up shortly here... after that we will see... 🙂

 

[Dirtyharry50]

I'm not really peeved, just frustrated at the misinformation from people trying to justify the poor performance (and for what?..). Ditch the ridiculous brand loyalty and demand more.
It's not a new thing and it's a little painful to swallow for us users, but OSX really is (and has been for a fair while) quite a poor performer in key areas when compared to Windows and Linux - Especially poor when OSX is tightly tied to hardware and the alternatives are massively generic.

As you rightly distinguished though, these are "just" benchmarks - There's no reason to say that you can't happily play a AAA game/churn through data and it's not saying anything about the arguably excellent user experience.

I can appreciate what you are saying there. I almost went back to edit something more onto my post and thought no, I'll just turn it into another of my walls of text but...

Reading this topic certainly has been of interest to me and I am hopeful that Apple will move forward with performance improvements over time. It's a must as software will become ever more demanding, particularly games as pertains to me personally. The other apps I use would all run beautifully on any older lower spec Mac. So I do care about performance going forward certainly.

I am aware too that we're behind the curve vs other operating systems. As I mentioned I haven't found it impactful on my own system but that doesn't mean for a moment that I won't down the road if Apple isn't proactive about investing in and implementing performance improvements.

I do think their involvement in this and all the improvements I've seen in recent years along with the ever increasing availability of quality titles to play represent great progress though. So, I am hopeful it will continue and I believe it will.

----------

Apple doesn't really have any interest in gaming, or else they'd put something besides the feeble Intel graphics in their machines. Looking at PassMark's benchmark comparisons, I see that my 2012 MacBook Pro Retina uses Intel HD4000, with G3D benchmark value of 455, vs. the 6900 of the R9 290X I have in my gaming PC. Intel Iris Pro 5200 is only 1210.

Is your gaming PC a portable computer? If not I think the comparison is apples to oranges, no pun intended.

I don't pay a lot of attention to portable hardware but doesn't a MacBook Pro include a dedicated GPU in addition to the onboard graphics with switching between the two as is needed? I was under the impression that you got a dedicated mobile GPU in a MacBook Pro but maybe I am wrong?

 

[Agent-J]

Apple really has an interest in gaming, because they put something besides the feeble Intel graphics in their machines. I see that my late 2012 iMac 21.5" uses the nVidia GeForce 650M GT gpu.

See what I did, there??

1255 on the benchmark I quoted. Thanks for proving my point.

Apple COULD support external GPUs via Thunderbolt--folks have actually made those work. Or they COULD bring out a Mini with discrete GPU (like a 290X) meant for gaming. Instead, they know the faithful will be happy with poor graphics and low frame rates, because Apple.

And I know the Retina iMac has a better GPU, but that's necessary just to drive all those pixels. Don't go looking to run graphics-intensive games in high detail mode at 60 FPS.

 

[Dirtyharry50]

1255 on the benchmark I quoted. Thanks for proving my point.

Apple COULD support external GPUs via Thunderbolt--folks have actually made those work. Or they COULD bring out a Mini with discrete GPU (like a 290X) meant for gaming. Instead, they know the faithful will be happy with poor graphics and low frame rates, because Apple.

And I know the Retina iMac has a better GPU, but that's necessary just to drive all those pixels. Don't go looking to run graphics-intensive games in high detail mode at 60 FPS.

Apple doesn't make gaming rigs but they do make models that play games fine. I own one of them and I'm happy with it. Will my iMac post benchmark numbers to compete with a PC gaming rig? Nope. It isn't a PC gaming rig so I'm fine with that. I don't want a PC with Windows. I want an iMac with OS X, etc. It works plenty well enough for me.

In a perfect world, Apple would introduce to their lineup a system design with a focus on gaming but until that happens, if it ever happens, all this back and forth about stuff is going nowhere. It is what it is.

So deal with it. Get whatever you like and be happy. Don't worry. Be happy. 😎

 

[Agent-J]

Is your gaming PC a portable computer? If not I think the comparison is apples to oranges, no pun intended.

I don't pay a lot of attention to portable hardware but doesn't a MacBook Pro include a dedicated GPU in addition to the onboard graphics with switching between the two as is needed? I was under the impression that you got a dedicated mobile GPU in a MacBook Pro but maybe I am wrong?

You can get Windows notebooks with darned fine GPUs. And, you can get a mobile GPU for extra price in rMBP, but we're not talking anything meant for serious gaming.

I like my Macs, and they enable me to be very productive at work. But to try to call them adequate for anything but low-end gaming is fairly silly IMO.

I've no doubt Apple could make good gaming computers, but have seen little sign of them having any interest. iOS is good for gaming, but many games cannot translate to touch input.

 

[Dirtyharry50]

You can get Windows notebooks with darned fine GPUs. And, you can get a mobile GPU for extra price in rMBP, but we're not talking anything meant for serious gaming.

I like my Macs, and they enable me to be very productive at work. But to try to call them adequate for anything but low-end gaming is fairly silly IMO.

I've no doubt Apple could make good gaming computers, but have seen little sign of them having any interest. iOS is good for gaming, but many games cannot translate to touch input.

Well, I'm not sure what your criteria for good gaming is. I can play Bethesda's stuff from Fallout 3 to Skyrim with Wineskin at high details and good performance on this system so how am I suffering? That's just a windows oriented example since there aren't OS X versions of those games. I was playing The Darkness II at 1440p and max detail with smooth performance yesterday on Steam for good times. That seems pretty good to me.

I don't like windows but I did reboot for a little Warhammer Dawn of War II at 1440p with high details and it ran the benchmark with excellent results to verify the turned up settings were alright. I've played Crysis in the past on 2011 iMac at 1440p with high settings.

How is it that the hardware is bad for gaming? I'm not seeing it here but then i don't measure my gaming fun by fps counters and benchmarking software. I measure it by smooth running, nice looking, fun to play games and my system delivers the goods there.

I know you can buy PC gaming laptops for a premium. Unfortunately, Apple doesn't make comparable gaming laptops so there is nothing to compare to. Fortunately though, some of the ones they do make can play games decently but no, they are not going to be as powerful for gaming as a competing specialized PC laptop computer.

 

[Ledgem]

How is it that the hardware is bad for gaming? I'm not seeing it here but then i don't measure my gaming fun by fps counters and benchmarking software. I measure it by smooth running, nice looking, fun to play games and my system delivers the goods there.

The frustration with the hardware has more to do with the lack of upgradeable options and the lack of a gamer-centric option. As TheBunny said, even if Apple uses a higher-end GPU in a system there's a chance that the clock rate will be underclocked compared to generic parts because Apple desires thin, quiet computers, and a hot graphics chip isn't conducive to that goal. External GPU options would be a nice compromise, but it strikes me as being more complex than Apple would want.

Otherwise, OS X is viewed as a poor gaming platform because of the software side. I don't know much about the graphics driver optimizations, but I do know that most gaming companies aren't targeting OS X for development. I think everyone would agree that performance-wise, a native application will always be preferable to a port. This is only furthered by the fact that not all games are fortunate enough to have Aspyr or Feral handling the ports, and we've seen some truly awful efforts. In those cases, running the game in Bootcamp results in massive performance improvements, even though it's all on the same hardware.

I game on my Mac, using Parallels where I have to (because I don't care to set up Bootcamp). I'm content, but I'm also not trying to play anything very intensive. For me, it's not about benchmarks or raw FPS numbers. I want the game to look good, and I want it to run smoothly. It would be a major disappointment if I had to turn the graphics settings all the way down, knowing that there was no upgrade (aside from buying a new Mac) that would allow me to bump up the settings.