Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

IBC - GPU Build REDCINE-X Beta 21

Grant Limberg said:
A 7990 is actually 2x7970's on a single card, and shows up to the OS as 2 cards. We don't yet support multiple GPUs. Additionally, we've run into some issues with AMD's HD 7900 series that causes them to run much slower than they would otherwise. We're working on resolving that issue with AMD now, but there's no ETA on when they'll be delivering a fix for end users.

That said, I have a 7970 GHz Edition at home and love it :)
Click to expand...

Hey Grant,

Has Red overcome some of the issues found with AMD 7900 series card in relation to RCx? What sort of capabilities does 7990 have now on RCx vs the likes of Titan/690/680?

thanks
 
Last edited:
Kalulu Ngilo said:
Hey Grant,

Has Red had overcome some of the issues found with AMD 7900 series card in relation to RCx? What sort of capabilities does 7990 have now on RCx vs the likes of Titan/690/680?

thanks
Click to expand...

Unfortunately, AMD still hasn't publicly released the fix for the issue we found.
 
Likely the cards are working but they are using an OpenCL implementation instead of CUDA

I'll leave it to the wizards at Red to let us know what will ultimately work better on nvidia cards when both options are available.

im pretty stoked with what I'm getting with opencl on my Titan right now.
 
Dan Kanes said:
Likely the cards are working but they are using an OpenCL implementation instead of CUDA

I'll leave it to the wizards at Red to let us know what will ultimately work better on nvidia cards when both options are available.

im pretty stoked with what I'm getting with opencl on my Titan right now.
Click to expand...

Ok that's cool. I know the folks at RED are mad geniuses! The Titan does speed up the workflow in all of my apps. It's amazing. Hopefully there will be a Maxwell Titan later this year. 6000+ cuda cores sounds sweet.
 
Ok, this is extremely frustrating. OpenCL works only occasionally. Tried it the first time, kept crashing. Two weeks ago, I was able to get "OpenCL enabled" working on my MBP w/ retina once, and it did perfectly fine exporting a bunch of footage (over 600GB of data).

Now, I opened RCX back up, and I tried exporting, and it crashed. Doesn't matter if it's set to exporting project, bin, etc. - everything just crashes.

Using OpenCL on a 2013 MBPr w/ GeForce 650 CL.
 
Brendan,

Which version of MacOS are you using? Also, if you could turn on verbose logging in the preferences and get it to crash, then send us the logs at redcinex at red dot com, that would be very helpful.

Thanks.
 
Hi Grant,

Thanks for the reply. I'm still using the 10.8.5 Mountain Lion OSX. I definitely know this ran on OpenCL before because my transcodes + exports were going insanely fast!
 
maybe a rude question, but why do you spend time making a CUDA implementation? CUDA was yesterday, OpenCL is today and tomorrow.
 
On a retina Macbook Pro I recommend a program called GFXcardStatus - it tells you which graphics card is enabled and lets you manually turn on or off the Nvidia card (secondary card) - in early resolve versions you had to use this to force external GPU on - no guarantees but this may be a similar situation - where if you force external GPU the OpenCL might "wake up" prior to running RCXpro.
 
Dan Kanes said:
On a retina Macbook Pro I recommend a program called GFXcardStatus - it tells you which graphics card is enabled and lets you manually turn on or off the Nvidia card (secondary card) - in early resolve versions you had to use this to force external GPU on - no guarantees but this may be a similar situation - where if you force external GPU the OpenCL might "wake up" prior to running RCXpro.
Click to expand...

Even if thats not the case GFXcardstatus is great to have....some programs force the dGPU when you want to save battery and you can switch it both ways. Although in power settings you can switch it the power save settings to "Performance" and it leaves the dGPU on no matter what. So far the latest build is playing nice with my retina MacBook Pro 15" 2013 with the Nvidia GT 650M. Just wish I could get FCP X 10.0.8 to play nice with the 6K .r3d's but have been transcoding to ProRes 4444 or TIFF stacks. Didn't realize how awesome it is to have 6K in even a 5K timeline let alone a 4K or 2K timeline! So much room to work in if you like to do allot of framing in post.

As for the guy who asked about CUDA, CUDA is still widely supported and to a certain extent more mature....I respect AMD GPU's but if I had to build a new rig I'd honestly go for Nvidia cards over any amount of AMD muscle. Dual GTX 680's, or an Nvidia Titan (or 3...lol) are amazing...guys that do rendering in Maya and C4D are talking about cutting 2 day render times to a couple of hours if not less!
 
because Nvidia has always been better than AMD/ATI.
Apple is trying to force OpenCL because of there deals with AMD and it has nothing to do with which is better OpenCL or CUDA.
 
roryhinds said:
because Nvidia has always been better than AMD/ATI.
Apple is trying to force OpenCL because of there deals with AMD and it has nothing to do with which is better OpenCL or CUDA.
Click to expand...

OpenCL has nothing to do with AMD/ATI. It's an open standard for calculations, CUDA is not. Like the Glide API, CUDA will be substituted by an open standard, and OpenCL has matured and will be that standard.
 
What exactly is being done in OpenCL ? The decompress , or the debayer ?
I believe it's the debayer. Can someone confirm ?
 
You must log in or register to reply here.
Back
Top