New Nvidia GTX 970 and 980 - cMP?

[logicalways]

EVGA GTX 980 SC + Yosemite 10.10 GM Works

Just received my GTX 980 and installed it into my 2009 8core Mac Pro 4,1. everything seems to work great. Upgraded from an EVGA GTX 660. The card runs pretty warm compared to the GTX 660 which is to be expected. running 3x Asus VG248QE monitors at 144hz with surround gaming is pretty smooth.

 

[ToroidalZeus]

Just received my GTX 980 and installed it into my 2009 8core Mac Pro 4,1. everything seems to work great. Upgraded from an EVGA GTX 660. The card runs pretty warm compared to the GTX 660 which is to be expected. running 3x Asus VG248QE monitors at 144hz with surround gaming is pretty smooth.

nice, can't wait to try out my gtx 970

 

[Pending]

Just received my GTX 980 and installed it into my 2009 8core Mac Pro 4,1. everything seems to work great. Upgraded from an EVGA GTX 660. The card runs pretty warm compared to the GTX 660 which is to be expected. running 3x Asus VG248QE monitors at 144hz with surround gaming is pretty smooth.

Are you getting full speed pci 2.0 in OSX Yos?
Do you use bootcamp windows too?

 

[Ironjer]

are you getting full speed pci 2.0 in osx yos?
Do you use bootcamp windows too?

+1

 

[hemon]

Just received my GTX 980 and installed it into my 2009 8core Mac Pro 4,1. everything seems to work great. Upgraded from an EVGA GTX 660. The card runs pretty warm compared to the GTX 660 which is to be expected. running 3x Asus VG248QE monitors at 144hz with surround gaming is pretty smooth.

Hi,

how did you get the 980 to work? I've it too and Yosemite PB4. I installed the OSX with "nv_disable=1", then the nvidia web drivers and it doesn't work because of the no signal to the monitor. Which port are you using? I'm on HDMI. Can you explain please the steps you did for bringing the 980 to work?

Cheers.

 

[Pending]

+1

So a fully working GTX980....must be the first!
Which exact GTX980 card is it?

 

[Asgorath]

Hi,

how did you get the 980 to work? I've it too and Yosemite PB4. I installed the OSX with "nv_disable=1", then the nvidia web drivers and it doesn't work because of the no signal to the monitor. Which port are you using? I'm on HDMI. Can you explain please the steps you did for bringing the 980 to work?

Cheers.

Try a different port? Plenty of folks are reporting that the 980 cards basically work fine, aside from OpenCL etc.

 

[hemon]

Try a different port? Plenty of folks are reporting that the 980 cards basically work fine, aside from OpenCL etc.

I would but I just have a HDMI Cable.

 

[thornslack]

Originally Posted by pending View Post
are you getting full speed pci 2.0 in osx yos?
Do you use bootcamp windows too?

Also curious about this. I assume bootcamp would see the card as PCIe revision 1. I wonder how much the 980 performance gets crimped by this.

 

[MacVidCards]

Also curious about this. I assume bootcamp would see the card as PCIe revision 1. I wonder how much the 980 performance gets crimped by this.

Tonight I'll see if I can turn on PCIE 2.0

As an update, so far no luck with EFI. But it took several months to crack GK110 so everyone needs to be patient.

 

[logicalways]

Sorry guys didn't get notifications that people replied to my post. The card is in the bottom pcie slot plugged into my monitors using display port. I've just pulled it out because I wanted to put it into my PC for a few test. Ran heaven unigine 4.0 benchmark on ultra with AA 8x and tressellation at extreme. Everything set to max except v-sync was off. Final Score 62fps.

 

[hemon]

Sorry guys didn't get notifications that people replied to my post. The card is in the bottom pcie slot plugged into my monitors using display port. I've just pulled it out because I wanted to put it into my PC for a few test. Ran heaven unigine 4.0 benchmark on ultra with AA 8x and tressellation at extreme. Everything set to max except v-sync was off. Final Score 62fps.

That's interesting! Tomorrow I will buy a cable for the display port and I will try it, I will report the result here .

Edit: Video acceleration works on Yosemite too?

Cheers S.

 

[N19h7m4r3]

Sorry guys didn't get notifications that people replied to my post. The card is in the bottom pcie slot plugged into my monitors using display port. I've just pulled it out because I wanted to put it into my PC for a few test. Ran heaven unigine 4.0 benchmark on ultra with AA 8x and tressellation at extreme. Everything set to max except v-sync was off. Final Score 62fps.

Edit: Video acceleration works on Yosemite too?

Why did you run the test in Windows?

Is that a hackintosh, or just a custom windows machine, especially judging by the CPU as well.

How does this pertain to OS X performance and support?

 

[hemon]

Why did you run the test in Windows?

Is that a hackintosh, or just a custom windows machine, especially judging by the CPU as well.

How does this pertain to OS X performance and support?

I'm on a Hackintosh . I work with Yosemite and play games with Windows where the performance are gorgeous

 

[N19h7m4r3]

I'm on a Hackintosh . I work with Yosemite and play games with Windows where the performance are gorgeous

In windows yes, I more interested in OS X performance for OpenGL, and OpenCL though.

EDIT:
Here's two D700's in Unigine Heaven. Strange how much of an average FPS difference that 1 minimum FPS can make. Even with a higher Max FPS.

 

[logicalways]

Update: GTX 980 MacPro

So I tried the EVGA GTX 980 SC in both my PC (i5 3570K, 16gigs Ram) and in my 2009 8core Mac Pro 4,1 with 32gigs of ram using the Unigine Heaven Benchmarks set to Ultra & Extreme. Interesting to see the two results.

I also ran the same benchmark test in Bootcamp using Windows 8.1 as well on my Mac Pro 4,1 and got a result of 55 fps.


i5 PC: 62fps
MacPro: 49fps
Bootcamp: 55fps

 

[MacVidCards]

In OSX, using cmd/shift/4 gets you a little cursor to drag.

puts a png on your desktop

Easiest screenshots ever.

cmd/shift/3 does whole screen, iirc

some news coming on efi, bad news first, the "old" PCIE 2.0 trick doesn't work anymore

Nvidia has moved the goalpost, and without a Mac Edition maxwell, may be very difficult indeed to find

but there is some good news.....

 

[logicalways]

Can't wait to get some better results!

 

[Pressure]

In OSX, using cmd/shift/4 gets you a little cursor to drag.

puts a png on your desktop

Easiest screenshots ever.

cmd/shift/3 does whole screen, iirc

some news coming on efi, bad news first, the "old" PCIE 2.0 trick doesn't work anymore

Nvidia has moved the goalpost, and without a Mac Edition maxwell, may be very difficult indeed to find

but there is some good news.....

If you use cmd+shift+4 and then hit space, it will select the window of whatever you hover the cursor over.

 

[Tutor]

So I tried the EVGA GTX 980 SC in both my PC (i5 3570K, 16gigs Ram) and in my 2009 8core Mac Pro 4,1 with 32gigs of ram using the Unigine Heaven Benchmarks set to Ultra & Extreme. Interesting to see the two results.

I also ran the same benchmark test in Bootcamp using Windows 8.1 as well on my Mac Pro 4,1 and got a result of 55 fps.


i5 PC: 62fps
MacPro: 49fps
Bootcamp: 55fps

As is your experience, in Heaven 4 instance Direct3d11 renders currently prevail over OpenGL renders on the GTX 980s. So, scoring in Windows is higher than scoring in OSX. Also, in Heaven (regardless of OS), the GTX 980s aren't much faster than the GTX 780 TIs [ http://www.digitaltrends.com/computing/nvidia-geforce-gtx-980-review/ ].

 

[N19h7m4r3]

As is your experience, in Heaven 4 instance Direct3d11 renders currently prevails over OpenGL renders on the GTX 980s. So, scoring in Windows is higher. Also, in Heaven (regardless of OS), the GTX 980s aren't much faster than the GTX 780 TIs [ http://www.digitaltrends.com/computing/nvidia-geforce-gtx-980-review/ ].

It's very interesting that the 2x D700's score much better on the minimum FPS than the 780Ti, and essentially match Max FPS; but the scores and overall/average FPS is so different.

 

[Tutor]

It's very interesting that the 2x D700's score much better on the minimum FPS than the 780Ti, and essentially match Max FPS; but the scores and overall/average FPS is so different.

When I ran my Heaven test recently, I did notice a couple of burps which probably accompanied the lower lows in my 780Ti, but otherwise the benchmark ran like a beautiful movie. I currently have > 60 GPUs and my experience with these, as well as others that I've owned, has been that there are variances, vagaries and ghost in all of them even when they have the same model number and, for those that display it, the same lot number. Maybe, they're remnants of imperfect manufacturing and binning processes.

The takeaway for cMP owners is to deeply consider the particular aspect he/she finds most important in a GPU depending on his/her most critical application and pick the GPU that best suits that need. Moreover, Heaven didn't measure my most crucial need. So we must not only know are GPU needs, but we must also be acutely aware of how and what measures the likelihood that a particular GPU will satisfy them. We shouldn't let ourselves be fooled by the market makers into believing that they know what is best for each of our particular circumstance(s) - those market makers are clueless about our particular need(s). And since our needs could be plenty, one GPU could be a compromise or we might need different one for something else of extreme importance. Not all GPUs excel at the same task(s).

 

[N19h7m4r3]

SNIP

Oh, quite true. It's just interesting that even with better minimums the scores vary so much. Just shows you simply can't rely on benchmark scores for all intents and purposes.

It doesn't make sense for a higher minimum frame rate, and same maximum to net a lower score.

The same happens in Cinebench as well, where some lower GPUs can match top tier ones.

Like you a person needs to really check their needs before picking a GPU, and some real world comparisons and tests in the applications a person requires would be far more beneficial.

 

[Tutor]

Oh, quite true. It's just interesting that even with better minimums the scores vary so much. Just shows you simply can't rely on benchmark scores for all intents and purposes.

It doesn't make sense for a higher minimum frame rate, and same maximum to net a lower score.

The same happens in Cinebench as well, where some lower GPUs can match top tier ones.

Like you a person needs to really check their needs before picking a GPU, and some real world comparisons and tests in the applications a person requires would be far more beneficial.

You deserve and receive a thumbs up from me. Heaven's score is just a summary. It doesn't give us a complete description of how many extreme lows and highs there were or even where they exactly occurred - they/it could have occurred where one would normally have expected (e.g., scene density) or they/it could have occurred where a particular GPU has a weakness, etc. Moreover, I can imagine that for some tasks, a GPU with high minimums may be preferable to one with a higher overall score, but lower minimums (and the same might apply to the highs), i.e., low variance may be essential. Thus, on a comparative basis, that imperfect glimpse that Heaven gives to us may still be used profitably by someone who has critically examined his/her needs.

 

[hemon]

It works with the display port!!!!

just photos doesn´t...