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

New Nvidia GTX 970 and 980 - cMP?

Ok, With 10.10.1 and 343.01.02f01, it still doesn't work, but using screen sharing I notice that the nvidia web driver is not loading. I chose it after install and restarted, but it's switching back to the OS X driver during boot-up.

How can I make the nvidia web driver 'stick'?
 
Got one in my other 5.1, which has your card and a SC version under 10.10.3.

Drivers are nvidia's latest for both web driver and cuda. Also have a GT120 in that box as all it's used for is rendering. I have it coming into the RGB input on one of my monitors off the GT120 and control it from my main 5.1 via teleport. Works fine. That one is a stock dual 2.6 4 core. Had zero issues getting it to work. Formatted drive, installed 10.10.3 via FireWire transfer, installed the nvidia stuff while on the GT120, dropped the cards in and rebooted and started rendering. I do NOT think it's the card model. You might have a specific INSTANCE of the card that has a problem (quickly ruled out by simply putting it into another box), but I have 3 different variants of the EVGA GTX 970 ACX 2.0's including yours and have zero issues with any of them, under 10.10.2 and 10.10.3.

I think you're more than likely chasing the wrong issue as there are a lot of people using these card with the exact same drivers on the same versions of the OS without issue, so it must be something else. The simplest solution is to check the card in another box. If it's good there then it's either the box hardware, the power connectors, your installation hardware wise or OS wise. Can't be anything else.
 
mitchino said:
Ok, With 10.10.1 and 343.01.02f01, it still doesn't work, but using screen sharing I notice that the nvidia web driver is not loading. I chose it after install and restarted, but it's switching back to the OS X driver during boot-up.

How can I make the nvidia web driver 'stick'?
Click to expand...

In terminal type "nvram -p" and see what shows up under "boot args"
 
terminal he say:

bluetoothInternalControllerInfo %15%82%ac%05%00%00%11Z|%c3%a1T%a6_
prev-lang:kbd en:2
fmm-computer-name Mitch%e2%80%99s Mac Pro
bluetoothActiveControllerInfo %15%82%ac%05%00%00%00%00%11Z|%c3%a1T%a6_
SystemAudioVolume %80
LocationServicesEnabled %01
EFIBluetoothDelay %b8%0b
Mitchs-Mac-Pro:~ Mitch$
 
took my 10.10.1 startup disk and the GTX970 and installed them in my other Mac Pro, a 2008 3.1 model. On this model, the nvidia driver does load, but still get a black screen. I'm using an old Hanns G monitor on this mac, and interestingly, when I unplug it from the GTX970, it says 'no signal', when I plug it back in that message goes away but the screen is still black.

Did a nvram -p on this mac and got:

bluetoothInternalControllerInfo %06%82%ac%05%00%00 ]%00%1eR%dbN%0b
prev-lang:kbd en:2
fmm-computer-name Mitch%e2%80%99s Mac Pro
boot-gamma d"%00%00i$%00%00%00%00%00%002%00%00%00%00%00%00%00%02%00%80%80%80%80%81%81%80%80%02%00%80%80%80%80%81%81%80%80%02%00%80%80%80%80%81%81%80%80
SystemAudioVolumeDB %00
bluetoothActiveControllerInfo %06%82%ac%05%00%00%00%00 ]%00%1eR%dbN%0b
SystemAudioVolume @
efi-boot-device-data %02%01%0c%00%d0A%03%0a%00%00%00%00%01%01%06%00%02%1f%03%12%0a%00%04%00%00%00%00%00%04%01*%00%02%00%00%00(@%06%00%00%00%00%00%10%c4%1e%1d%00%00%00%00%c0@%f0s%05%e7%15B%86%e1^%9bK0%18%d0%02%02%7f%ff%04%00
efi-boot-device <array><dict><key>IOMatch</key><dict><key>IOProviderClass</key><string>IOMedia</string><key>IOPropertyMatch</key><dict><key>UUID</key><string>73F040C0-E705-4215-86E1-5E9B4B3018D0</string></dict></dict><key>BLLastBSDName</key><string>disk0s2</string></dict></array>%00
boot-args nvda_drv=1
EFIBluetoothDelay %b8%0b
500GB-SD:~ Mitch$
 
Does anyone know the correct procedure for installing the drivers and card (980Ti) on my Mac Pro? Do I install the GPU drivers first, then the card, or vice versa. Also, when do I install the CUDA driver System Pref, first or last?

Thanks.
 
Upgrader said:
Does anyone know the correct procedure for installing the drivers and card (980Ti) on my Mac Pro? Do I install the GPU drivers first, then the card, or vice versa. Also, when do I install the CUDA driver System Pref, first or last?

Thanks.
Click to expand...
Driver first. Then card. Cuba anytime.
 
AmanO said:
Also, your speculation on how long Mac OSes are going to officially support nVidia?
I would have to look, but no Macs on sale now use them anymore, right?
Click to expand...

I think all of the newer models sell without Nvidia graphics today. Nevertheless, OS X will have to continue to include Nvidia driver support for many years to come in order to cover all of the existing Macs out there with Nvidia GPUs.

I do wonder about the future of Nvidia on the Mac though. In the past, Apple has switched between Nvidia and AMD frequently. More recently however they seem to be expunging Nvidia from the entire Mac lineup, and sticking to it. They certainly are showing zero interest in including Nvidia's Maxwell drivers into OS X by default. :(
 
AmanO said:
Can anyone write that utilization into an app or do you have to have a license?

Also, your speculation on how long Mac OSes are going to officially support nVidia?
I would have to look, but no Macs on sale now use them anymore, right?
Click to expand...

ATI and Nvidia are the only two GPU manufacturers left in the game. Nobody else is close. So Apple and Microsoft need them for e foreseeable future. Apple can't make Metal look impressive without them otherwise Windows users will laugh.
 
I'm shooting to upgrade my 4,1 to a flashed GTX970 soon. Is there any benefit to keeping my current Radeon 4870 installed as well? I don't have any other PCI cards installed, nor do I have any (current) reason to install any. I typically only use one display.
 
PostProdMan said:
With the recent announcements of Nvidia's newest GPUs, is there any chance these could work in Mavericks or Yosemite?

Their low price points @ $299 (970 Ref) $329 AIB, and $549 for the 980, and their low TDP of 148W and 175W, they seem to be tailor made for the wattage requirements in the cMP.

Any thoughts or insights?
Click to expand...

No problems at all with the 970's.

I run two of them (EVGA GTX 970 SC ACX 2.0 and EVGA GTX 970 FTW ACX 2.0 in an unmodified 5.1 with nothing other than splitter power cables I got off eBay. I've got 2 boxes set up like this and they work extremely well for Blender / Rhino.

2 cards under full load renders can run 24/7 without getting close to the current limit on a 5.1 era power supply.

IMO for render operations there isn't enough performance delta with the 980's (like about 8 - 12% in gpu renders) to justify the additional cost / current draw of the 980's.
 
NorCalLights said:
I'm shooting to upgrade my 4,1 to a flashed GTX970 soon. Is there any benefit to keeping my current Radeon 4870 installed as well? I don't have any other PCI cards installed, nor do I have any (current) reason to install any. I typically only use one display.
Click to expand...

uh....The 4870 does not need supplemental power? What do you do on it? I think you last sentence pretty much answered your own question.
 
NorCalLights said:
I'm shooting to upgrade my 4,1 to a flashed GTX970 soon. Is there any benefit to keeping my current Radeon 4870 installed as well? I don't have any other PCI cards installed, nor do I have any (current) reason to install any. I typically only use one display.
Click to expand...

No, just keep it safe at somewhere else is good enough, especially it's a flashed card.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back