680MX Benchmarks please

here is a good guide on how to oc that 680mx

http://forum.notebookreview.com/gam...guide-workaround-max-clocks-oc-stability.html

Mr MM said:

here is a good guide on how to oc that 680mx

http://forum.notebookreview.com/gam...guide-workaround-max-clocks-oc-stability.html

Click to expand...

680MX is already the most powerful GPU with TDP of 122W. Do you really think, it needs any overclocking higher than standard +125(GPU)/+1000(vRAM) MSI Afterburner could offer?

Just talked to someone from NVIDIA, they said there are no drivers for 680MX at the moment. They said they will be providing some in the future. However, the only drivers available at the moment are the ones provided by Apple.

They said signup to their newsletter to be alerted about when new drivers are released.

Confuzzzed said:

I feel guilty entering the thread as a non-serious gamer. Do dabble on a bit of Civilization V but that's about it. Question I have is this, what sort of improvement does the 680MX have over the 675MX all else being equal?

I have to decide between i7/675MX or i5/680MX for my machine which is currently ordered for the former but the more I dig, the more I want to change the order to the latter

Click to expand...

No one knows for sure yet. Very little benchmarks at this point. From my research 680mx theoretically should be 20-30% faster than 675MX.

I was in the same boat with choosing between i7/675MX or i5/680MX. It was an easy choice for me to pick the i5/680MX option. Most users will see very little improvement from i5 to i7 unless they do a lot of hyper threading intensive tasks. While even if you do some gaming, you should see a good bit of boosted frame rates with the 680MX over the 675MX.

We won't know for sure until driver's are running smoothly and a larger pool of benchmarks can be analyzed.

Fenn said:

2012 iMac, 64-bit Geekbench 14155
http://browser.primatelabs.com/geekbench2/1398889

Click to expand...

I wonder about this bench. I ran a quick Geek 64 bit on the same unit except 32gb ram and an SSD and got 13,2xx something. I had just closed a session of VMware Fusion, but still. . . something seems off about the 14,155+ scores. Anyone else run this with a similar unit?

Looking at the spread of similar imacs, they are all over the place from in the 9,000's to the 15'000s. Why so much variation for such similar units? http://browser.primatelabs.com/geek...platform:"Mac+OS+X"+architecture:x86+bits:64+

3DMark 11 scores for GPUs used in 5 different iMacs

Looking at 3DMark 11 results from the FutureMark website,

Radeon HD 6970 = 3174 score (2011 top iMac CTO)

GeForce GT 650M = 2185 score (2012 21" iMac and 15" Retina MBP)

GeForce GT 660M = 2394 score (2012 27" iMac 2.9GHz i5)

GeForce GTX 675M = 3337 score (not the 675MX in the 2012 27" iMac 3.1GHz Core i5 but . . .)

GeForce GTX 680M = 6437 score (not the CTO 680MX available in the top 27" iMac but definitely faster than all above)

GeForce GTX 680 (desktop model) = 10590 (and that's not even overclocked like the one we have one in our Mac Pro)

I had to reinstall BC and forgot the second time to change energy profiles. I will adjust that and run vantage and OpenGL bench as well.

The only drivers available from nvidia atm that even install were the current beta ones released on the 4th. They showed a 500 point increase over the default apple ones.

Apple has a history of underclocking chips for thermal or power consumption reasons. Is this the case with the 680MX in the 27" iMac?

barefeats said:

Looking at 3DMark 11 results from the FutureMark website,

Radeon HD 6970 = 3174 score (2011 top iMac CTO)

GeForce GT 650M = 2185 score (2012 21" iMac and 15" Retina MBP)

GeForce GT 660M = 2394 score (2012 27" iMac 2.9GHz i5)

GeForce GTX 675M = 3337 score (not the 675MX in the 2012 27" iMac 3.1GHz Core i5 but . . .)

GeForce GTX 680M = 6437 score (not the CTO 680MX available in the top 27" iMac but definitely faster than all above)

GeForce GTX 680 (desktop model) = 10590 (and that's not even overclocked like the one we have one in our Mac Pro)

Click to expand...

the 680m that you got is also without oc.

you have to realize that the 680m is a 670 and the 680mx is almost a 680. The main difference is that the clocks are extremely slow, like 720 for both on the core.

kaellar said:

680MX is already the most powerful GPU with TDP of 122W. Do you really think, it needs any overclocking higher than standard +125(GPU)/+1000(vRAM) MSI Afterburner could offer?

Click to expand...

the question is why not? there is a very large gap on what the 680mx can provide and a 680 can.

Mr MM said:

the question is why not? there is a very large gap on what the 680mx can provide and a 680 can.

Click to expand...

I don't think the issue is what the card is capable of, but what is thermally advisable in the iMac's form factor.

MatthewAMEL said:

I don't think the issue is what the card is capable of, but what is thermally advisable in the iMac's form factor.

Click to expand...

if this can happen on a m15x, I really think the thermal budget of the imac is going to be better

http://www.3dmark.com/3dm11/5170805

also, your computer wont come to harm, only if you raise the voltage, what will happen if you hit the thermal limit is the throttling of the 680mx

Mr MM said:

if this can happen on a m15x, I really think the thermal budget of the imac is going to be better

http://www.3dmark.com/3dm11/5170805

also, your computer wont come to harm, only if you raise the voltage, what will happen if you hit the thermal limit is the throttling of the 680mx

Click to expand...

oh really want an imac now!

Ill have to look at gpu-z but from what was reported by the nvidia control panel it appeared that the core and memory aren't down clocked and are running at stock nv frequencies.

The fan did kick in after a while of running arkham city but 75% of the time the machine was completely silent.

The back was cool to the touch, unlike other iMacs in the past. My initial thoughts are the cooling design in these is tremendous, and a stable healthy oc might be very feasible.

Hi barefeets /bow Here are the results you requested run in Win7. All of these have been run after a clean restart and with energy profile on max performance. They are using the current beta drivers from the Nvidia website as the stock apple drivers were around 10% slower in all cases. I'll have to reboot in a bit and try the OS X versions that are available for the same benchmarks. I have yet to OC. Maybe later.


Heaven (Native resolution, default settings.)

FPS: 51.9
Scores: 1307
Min FPS: 28.6
Max FPS: 115.3


LuxMark (Sala default, GPU only)

Samples/sec 430K
Rays/sec 4155K on 488.5k tris


Geeks 3D GPUTest (native resolution, default options)
FurMark
Score: 1821 points (FPS: 30)

GiMark
Score: 3484 points (FPS: 58)

barefeats said:

Run the Heaven OpenGL Benchmark at 2560x1440 --defaults to that when you launch it on the 27" iMac. Use other default settings as well. Click "Run" and then "Benchmark" - the average FPS will be displayed in a pop-up window after it "flys" through 26 scenes:
http://unigine.com/products/heaven/download/

If you are running Windows, it tests Direct-X speed.

An OpenCL GPU test is LuxMark. (Runs on OS X and Windows.) You can render three different scenes with GPU only, CPU only, or combined CPU/GPU. Please post your results in Ksamples/sec for the Default Sala scene using GPU only:
http://www.luxrender.net/wiki/LuxMark#Download

Bonus: Geeks3D GPUtest - Runs under OS X as well as Windows. Has 3 OpenGL tests. Run FurMark and GiMark. Set at 2560x1440 and default settings. Click "Run benchmark" button and wait for the average to be posted in a pop-up window:
http://www.geeks3d.com/20121113/gpu...gl-benchmark-furmark-lands-on-linux-and-os-x/

Click to expand...

Hey Fenn (and Barefeats).

It seems like these are the first compareable results right now. If I interpreted it the right way, the new iMac with the 680MX is almost as fast as a mid 2010' Mac Pro 3.33GHz Hex-Core with a GTX 680 (as tested by barefeats) at the Heaven Benchmark (~3 FPS/Second difference).

That would be some stunning results, right?

Edit: I just realized that the Mac Pro results are produced on MacOs vs. iMac results on Windows 7. Not that compareable as desired.

Fenn said:

Hi barefeets /bow Here are the results you requested run in Win7. All of these have been run after a clean restart and with energy profile on max performance. They are using the current beta drivers from the Nvidia website as the stock apple drivers were around 10% slower in all cases. I'll have to reboot in a bit and try the OS X versions that are available for the same benchmarks. I have yet to OC. Maybe later.


Heaven (Native resolution, default settings.)

FPS: 51.9
Scores: 1307
Min FPS: 28.6
Max FPS: 115.3


LuxMark (Sala default, GPU only)

Samples/sec 430K
Rays/sec 4155K on 488.5k tris


Geeks 3D GPUTest (native resolution, default options)
FurMark
Score: 1821 points (FPS: 30)

GiMark
Score: 3484 points (FPS: 58)

Click to expand...

Good work! Can anybody decipher these results for us? I tried finding some comparisons, but with all of the graphics card variations out there it is hard to make sense of it all.

Fenn said:

Hi barefeets /bow Here are the results you requested run in Win7.

Click to expand...

Thanks, FENN, for the results for Heaven, LuxMark, FurMark, and GiMark.

I'll boot up Windows on the old Mac Pro with the GTX 680 Classified to see how the numbers compare.

I can give you one comparison. On the Mac Pro with GTX 580 running Windows, the Heaven Score at 25x14 rez was 49 FPS vs your 52 FPS. I get 57 FPS on the 680 Classified but that's under OS X.

Let's just say at least in that case, the 680MX is truly running with the "big dogs."

I did some Googling, and came across this image which is from the Heaven Benchmark:

http://imgur.com/GST1K

From this link:

http://www.tomshardware.com/forum/372225-33-3dmark-heaven-benchmark-good-results


The spec mentioned is: i5 2500K overclocked to 4.2Ghz, GeForce 680 GTX, at 1920x1080. That looks VERY comparable to the iMac scores posted above, and don't forget the iMac is running at 1440p, not 1080p as that image is.

So... that means the 680MX is a beast from what I can tell!

tacokisses said:

Do you still have 3DMark installed? If so, maybe run once more with your energy profile set to High Performance?

The next thing is to upload gameplay videos to youtube

Click to expand...

Turning energy profile to High Performance didn't change anything performance wise with 3DMark.
Not videos yet, but here are a few screenshots from the Batman: Arkham City built-in benchmark.

2011 iMac Radeon HD 6970 - 2560x1440, maximum settings, FXAA (High), PhysX (Off)
Minimum 8, Maximum 34, Average 26
http://img208.imageshack.us/img208/2277/2012121300001r.jpg

2012 iMac Geforce 680MX - 2560x1440, maximum settings, FXAA (High), PhysX (Off)
Minimum 14, Maximum 85, Average 68
http://img138.imageshack.us/img138/9673/2012121300001a.jpg

This means that the 680MX has a 2.43X higher Maximum framerate, and a 2.62X higher Average framerate. This also means that in Bootcamp at least, when the 680MX drivers mature, we might be able to see close to 3 times the performance in comparison to the high-end 2011 iMac. Astounding hardware!

Also for comparison, with the 680MX, Batman: Arkham City now brings hardware accelerated PhysX to the iMac at completely playable framerates. Arkham City with PhysX on normal DRASTICALLY improves the realism of the game. Smoke and mist displace when characters pass through, fabric blows with the wind, sparks bounce off of all surfaces, newspaper pages swirl around corners, the immersion is just unbelievable.

2012 iMac Geforce 680MX - 2560x1440, maximum settings, FXAA (High), PhysX (Normal)
Minimum 23, Maximum 72, Average 55
http://img405.imageshack.us/img405/6696/2012121300002u.jpg

Even with PhysX set to High, which recommends the use of a dedicated GPU (!), you can still get framerates twice that of the old 2011 iMac. This thing literally will not choke.

2012 iMac Geforce 680MX - 2560x1440, maximum settings, FXAA (High), PhysX (High)
Minimum 24, Maximum 72, Average 51
http://img849.imageshack.us/img849/977/2012121300004.jpg


Also, here is my GPU-Z

ahh this is exciting! can't wait to see some more direct comparisons

even more excited about receiving mine now

More Heaven comparisons

Heaven under OS X at 2560x1440,

Retina 15" MacBook Pro 2.7 w/ GT 650M = 17 FPS

2011 iMac 27" Core i7 3.4GHz w/ HD 6970 = 32 FPS

2010 Mac Pro 6-core w/ HD 5870 = 37 FPS

2010 Mac Pro 6-core w/ GTX 680 Classified = 57 FPS

Under Windows,

FENN's 2012 iMac 27" Core i7 3.4GHz w/ GTX 680MX = 52 FPS

barefeats said:

Heaven under OS X at 2560x1440,

Retina 15" MacBook Pro 2.7 w/ GT 650M = 17 FPS

2011 iMac 27" Core i7 3.4GHz w/ HD 6970 = 32 FPS

2010 Mac Pro 6-core w/ HD 5870 = 37 FPS

2010 Mac Pro 6-core w/ GTX 680 Classified = 57 FPS

Under Windows,

FENN's 2012 iMac 27" Core i7 3.4GHz w/ GTX 680MX = 52 FPS

Click to expand...

wow, so thats like 60% better than the 2011

This is pretty amazing