Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

w00tini

macrumors 6502a
Original poster
Feb 28, 2008
661
62
Hi all,
Thought you might be interested in some gaming and gaming-related benchmarks from the 2010 Hex-Core Mac Pro with the ATi HD5780 graphics processor. Copied from my post in the Mac Pro section of this forum:

Configuration
  • One 3.33GHz 6-Core Intel Xeon “Westmere”
  • 6GB (3x2GB)
  • 1TB 7200-rpm Serial ATA 3Gb/s hard drive
  • ATI Radeon HD 5870 1GB
  • One 18x SuperDrive
  • Apple Magic Mouse
  • Apple Keyboard with Numeric Keypad (English) & User's Guide
  • Apple Mini DisplayPort to DVI Adapter
Well, I just ordered my first Mac Pro. I'm a four year Apple user that is finally taking the plunge to a single computer solution after years of various MacBooks and a MacMini sitting next to my PC tower. It seems more efficient to have a powerful Mac like the Mac Pro which is capable of running both the OS X and Windows platforms instead of two large, power-hungry machines running at one time.

I was torn on the Quad vs Hex debate but after reading a lot about the differences between them and my needs the Hex was a better solution, albeit a more expensive one. I'm planning on acquiring the new Apple 27" Cinema display when they start shipping and I feel that the Hex will be more capable of running video games at the beautiful full 2560x1440 resolution given the CPU power needed to push the GPU to such higher resolutions. It also appears, given the current benchmarks of StarCraft 2, that Blizzard is utilizing multiple cores in their latest engines. I'm a huge Diablo fan and plan on spending way too many hours on Diablo 3 when it hits so I'm hoping the Hex will also be of benefit in that instance. Not that the Quad is bad, but the Hex is well... the Hex :D

Update after arrival:

Ok so it has been here for a few days now and I've had some time to dig into it. I've installed Windows 7 via Bootcamp and have Steam running on both the OS X and Windows side of things. I have also run some various benchmarking software where applicable given which OS platform they run on.

Here's what I have so far....

XBench
XBench.jpg

(XBench in OS X)

Some solid numbers here. Very impressive memory and CPU bandwidth. I'm running 2x 3GB memory sticks in triple channel mode which is accounting for the memory test numbers. Know that you will certainly slow down to a degree running 4 memory sticks, but also understand that not all programs require fast speeds, some require more memory overhead and not such a fast thoroughfare through the bus.

GeekBench
geekbenchWin.jpg

(GeekBench in Windows7)

I'm not sure why it is reporting 32-bit as I'm running Windows7 Professional in 64-bit mode. But nonetheless, great metrics were taken away from running this benchmark.

Cinebench OS X
CinebenchOSX.jpg

(OS X Cinebench)

As you can see, very poor 3D capabilities in Cinebench on the OS X side given the rather disappointing OpenGL support Apple has at this time. While the CPU score is solid, OpenGL is lacking.

Cinebench Windows7
CinebenchWin.jpg

(Cinebench Windows7)

Using Windows7 via Bootcamp is where the Mac Pro really shines for an all around good gaming experience. Note the near-double OpenGL score in Win7 as compared to the OS X benchmark. The Mac Pro is a flat out scorcher in Windows7.

3DMark 06
Note: Windows Only

3DMark06.jpg

(3DMark 06 in Windows7)

Need I say more? That's a very, very respectable score for any custom PC build rocking a single core/single graphics card configuration, and will surely satisfy any current PC game's requirements.

3D Mark Vantage
Note: Windows Only

3DMarkVantage.jpg

(3DMark Vantage in Windows7)

Game Benchmarks

Counter-Strike Source Stress Test

Note: Run with monitor resolution of 1920x1200 with all graphics and advanced graphics settings at maximum in Windows7.

While Counter-Strike Source is an old game, the Stress Test built into it is still a good indicator of CPU and GPU bandwidth due to the how the two need to work together to push FPS in PC games. Source is dependent on the CPU pushing a lot of information through the GPU since the CPU is where the bottleneck would most likely occur in an older game like this.

The FPS were pegged at 300 (the maximum Source records in this benchmark) for most of the run but did dip down to ~ 200 for the more taxing parts of the stress test. Still, slowed down to 200 fps?? After several runs the average stress test score was 225.04 FPS in Windows7.

Unfortunately the OS X version of CS:Source does not have the stress test so I was not able to run it for a good comparison.


Left4Dead and Left4Dead 2

Note: Run with monitor resolution of 1920x1200 with all graphics and advanced graphics settings at maximum. Windows Only for now. Rumors are October 4th for OS X versions of this game.

Given that timedemos are usually broken whenever a game is updated I was not able to run any specific in-game benchmark for either of the L4D titles. But I did take pretty detailed notes on FPS with hordes of zombies running around me, being boomed, and attacked by all angles to really push the Mac Pro in this game.

The lowest FPS I saw at any one time in the most massive of horde attacks was 150 FPS in Left4Dead and 90 FPS in Left4Dead 2. In most cases it was over 200 FPS in both games. The games played beautifully no matter what was going on on-screen. Just a great gaming experience and much, much better than the Intel i7/Nvidia Gforce GTX260 that this Mac Pro replaced.

Team-Fortress 2

Note: Run with monitor resolution of 1920x1200 with all graphics and advanced graphics settings at maximum. Tested in both Windows7 and OS X.

Much like the previous two titles here, I ran this game with a screen full of 24 bots in a server to get a feel for how well the game would run. I never saw TF2 game drop below 120 FPS in Windows. OS X fared a bit lower with an average FPS of around 80 FPS.
Considering the CPU time required to run 23 bots in expert mode, it is a testament to the great multi-core utilization we're seeing from Valve in their games.

Battlefield Bad Company 2

Note: Run with monitor resolution of 1920x1200 with all graphics and advanced graphics settings at maximum. Tested in Windows7.

BC2.jpg


BC2c.jpg


BC2a.jpg


Overall, BC2 is a smooth and lag/stutter-free gaming experience. Even with a lot of things happening at one time on the screen the game is a cinematic experience on the Mac Pro with every single video enhancement enabled to its highest setting.

That's it for now. I'm going to mess around with more games as time goes on and update this thread. Please be sure to post any of your own gaming or benchmark metrics here so we can all share our experiences with the Mac Pro.
 
wow, that is an incredibly thorough and solid post. i'm actually pretty impressed with the performance it has as far as games go. I was expecting good, but not great. i'd like to see some benchmarks for other games on windows as well. excellent post sir!
 
Thanks Zao. I'm loading Bad Company 2 right now. 5Gb download over Steam will take about an hour. I'll then get back to this thread with some numbers and comments.
 
Updated first post with Battlefield Bad Company 2 screenshots and comments.
 
That's quite impressive - thanks for posting this, now I have (more) ammo for those conversations about Macs being bad for games...

You've just got a great setup going...I run an open-source fork of Quake 3 (old old game engine) on my old old Macbook at 125 FPS only when I cut down all the graphical enhancements - if not, it's supposed to look kind of like those screenshots of yours.

No envy here, just satisfaction - now I know what I've got to invest in at some point!

So, benchmarking...I suppose that would be the best way to go about diagnosing a possible issue with a CPU/GPU unit, then.

I see two "scores" - how does a benchmarking program "score" your computer's performance? I mean, I understand the concept, but to me it would make more sense if it was scored out of an even 100 and said something like "You are getting 98% of what you paid for" or something like that. Still, quite fun to read. :p
 
oooo, thanks for the screenies. that looks amazing, that's really awesome to see a mac producing that kind of gaming performance. seeing you have this kind of performance with such a beautiful machine just makes me happy. now if only they could spread those cards out to cheaper macs :p

the 'ol 9600gt isn't cutting it for me anymore.
 
No wonder about the results. The MacPro has the best components for a mac and with a hex core CPU, the beastly ATI 5870 and 6 gigs of RAM, that is gaming bliss for a PC atm.

Would be interesting to know the benchies in OS X rather than Windows ( if there is an OS X version of the tests ).

btw, nice screenshots.
 
unfortunately the benchmarks I ran on Windows are not available for OS X.

I will be sure to keep reporting back here with feedback, games, and screenshots as I continue to try out new games. I have a few days of business travel but I'll get back to it by next weekend.
 
unfortunately the benchmarks I ran on Windows are not available for OS X.

I will be sure to keep reporting back here with feedback, games, and screenshots as I continue to try out new games. I have a few days of business travel but I'll get back to it by next weekend.

I have the same setup except with 12 gigs of ram, I am assuming when you say all settings at high, you still have v-sync off. I don't see how you can have that FPS on counter strike with v-sync on as it pretty much locks you to 60 fps.

My average FPS with counter strike source stress test with aliasing at 8x and filtering at 16x with everything on high was 200 fps. Do you have certain settings on ATI catalyst turned on or something? Or did you maybe have some settings turned down....

EDIT:

Just ran cinebench in windows 7 64 bit bootcamp...my opengl score is really low compared to yours, mine is 34 fps. I wonder what is going on here, I did just update my drivers, I wonder if the drivers just suck. Any idea's?

2nd Edit:

Wow, just uninstalled my ati drivers..... my score went from 34 fps to 65.24!!!! So I guess I am not installing those again.
 
crevice,
good point that I forgot to bring up. I'm using the stock bootcamp ATI drivers. they are actually quite good. I would not mess with anything else given how well the system performs right out of the box. I'm very impressed with Apple's attention to this detail. they and ATI have written a solid set of drivers for the Windows side of the equation.
 
crevice,
good point that I forgot to bring up. I'm using the stock bootcamp ATI drivers. they are actually quite good. I would not mess with anything else given how well the system performs right out of the box. I'm very impressed with Apple's attention to this detail. they and ATI have written a solid set of drivers for the Windows side of the equation.

Yeah, I am leaving these drivers until other people will guinea pig for me and let me know when its safe to update, because as of now it is not. Is there anyway you can re-run your Counter Strike source test and just make sure you have every single thing on maximum (except v-sycn), with aliasing on x8 filtering on x16 and let me know your score. I can't seem to get close to your score with everything on maximum, though we have identical machines besides the RAM. Thanks in advance if you can help me with this.
 
yes, the score I am getting is with everything set to max (8x, 16x) and v-sync disabled.
 
yes, the score I am getting is with everything set to max (8x, 16x) and v-sync disabled.

I can get about 195-200 fps average on the score, I max out at 300 for a good part of the tail end of the test, but my final score is usually 195-200. I still don't get how your final average score is almost 40-50 fps more then mine. Once again if you can do the test once more, and let me know what your score is that would be great.
 
I can get about 195-200 fps average on the score, I max out at 300 for a good part of the tail end of the test, but my final score is usually 195-200. I still don't get how your final average score is almost 40-50 fps more then mine. Once again if you can do the test once more, and let me know what your score is that would be great.

Just ran it again: 277 FPS

df3710af.jpg
 
I actually just got up to 254 fps. I disable aero theme, and disabled a few services. Did you tweak your windows 7 for better performance, and if so can you share some tips?

hi crevice
I left the aero desktop active and have not made any changes to Win7 at all from the stock install. have you run all of your software updates in Windows?

what is your RAM configuration? are you running 3 x 4Gb sticks in triple-channel?
 
hi crevice
I left the aero desktop active and have not made any changes to Win7 at all from the stock install. have you run all of your software updates in Windows?

what is your RAM configuration? are you running 3 x 4Gb sticks in triple-channel?

Yeah it is 3x4gb in triple channel. I can't really complain, especially since it seems that the test varies so much. Even your tests have jumped around, I actually kind of doubt disabling aero did anything, and it's the just the test itself.

Every game runs great, battlefield on high runs on 60 fps with v-sync on, etc. So I don't really think anything is wrong. Just curious if you had some settings that I wasnt aware of, thanks.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.