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

lexR

macrumors regular
Original poster
Dec 12, 2013
210
23
UK
Hi guys,

I have recently heavily upgraded my 4.1 2009 MP from an 8 core 2.66 to a 5.1 12 core 3.46 with 48gb of ram and PCIe SSD. I also upgraded my ATI 5870 to a nvidia GTX 770.

My question is I have found myself using FCPX and motion a lot recently plus I have been creating models in Autodesk fusion, render times and responsiveness does 'seem' that blistering fast but unfortunately I have only really started to get into 3D modelling and HD video editing so I have no benchmark, (I think Fusion is quite a poor piece of software when it come to optimising the hardware as its a bit clunky and only uses 100% of one core and barely any ram in editing and building, but yet as soon as you start to render all cores kick in). So does anybody have advice on a possible bottleneck on my specs above? I am guessing the card and was thinking of possibly 2 GTX 970's but unsure if they have ironed out all the kinks yet?

Basically I would like my machine to scream through FCPX and 3D modelling, which i feel the core components can do but its like i missing something...:( happy to run some bench marks if anyone with a similar set up but, say, a gtx 980 running in FCPX to see if i can claw back some time.

thanks in advance
 
Two 970s use up to 450w peak consumption when under heavy load. They would have to be powered externally otherwise they will fry your mobo.

Everyone here knows I'm an Nvidia man (since way back to Riva 128 days) but if you want a GPU for FCPX then it would have to be a Radeon. They are more optimised for Open CL even though the GTX 980 did improve performance a lot. Part of the problem is that Nvidia wants developers to focus on CUDA.

If you can wait until February-March then you can make a comparison between the new Radeon 380/390 series and GeForce 980 Ti. Benchmarks for them have leaked and put the Radeons slightly ahead (but they consume more power).

All the next generation GPUs come in at 205w and under so they should work in a 5,1.
 
Everyone here knows I'm an Nvidia man (since way back to Riva 128 days) but if you want a GPU for FCPX then it would have to be a Radeon. They are more optimised for Open CL even though the GTX 980 did improve performance a lot. Part of the problem is that Nvidia wants developers to focus on CUDA.

Ahh, thanks for the info, i have seen post from people running dual 5770's with ok results for FCPX and trying to reason to myself a possible swap from my new 'ish' GTX 770 back to the older 5770's :( Open CL is the engine in Autodesk Fusion too which may help if i swap to radeon in the future...

thanks a lot for the advice, i am currently also swapping the single SSD on a PCIe card for a sonnet tempo pro and doubling up in RAID 0 for quicker read write time which may help...
 
Ahh, thanks for the info, i have seen post from people running dual 5770's with ok results for FCPX and trying to reason to myself a possible swap from my new 'ish' GTX 770 back to the older 5770's :( Open CL is the engine in Autodesk Fusion too which may help if i swap to radeon in the future...

thanks a lot for the advice, i am currently also swapping the single SSD on a PCIe card for a sonnet tempo pro and doubling up in RAID 0 for quicker read write time which may help...

Another advantage with your classic Mac Pro is you can swap between a Nvidia or Radeon at any time whenever you use a software or perform a task that is optimised for either GPUs. I always keep both a Nvidia or Radeon for specified tasks that may arise.
 
Another advantage with your classic Mac Pro is you can swap between a Nvidia or Radeon at any time whenever you use a software or perform a task that is optimised for either GPUs. I always keep both a Nvidia or Radeon for specified tasks that may arise.

yes that is true, from researching you get the feeling Apple have lent towards AMD more so than nvidia with regard to FCPX and progression through to the nMP so logic says that i would have a better experience with a radeon setup but it just such a mine field with regard to what 'works' fully with the cMP with regard to PCIe 2.0 and dual card, internal power usage etc etc
 
yes that is true, from researching you get the feeling Apple have lent towards AMD more so than nvidia with regard to FCPX and progression through to the nMP so logic says that i would have a better experience with a radeon setup but it just such a mine field with regard to what 'works' fully with the cMP with regard to PCIe 2.0 and dual card, internal power usage etc etc

Yep you're right. Probably for business or marketing reasons Apple chose to work closely with AMD Radeon gpus than Nvidia. I try do some researches too when using non-Mac cards and I would prefer having it flashed professionally. I use Adobe apps with Nvidia cards but I also keep a Radeon card when the need arises.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.