video card - C1 speed vs 30 bit color
I'm building a new PC to run OS X and have a GPU question.
I know the R9 is on the C1 supported list and that's what I plan to use. I'd also like to work toward getting 30 bit color while working in photoshop and maybe in the future, C1, if supported. I plan to use a low end 2GB Quadro video card to get the 30 bit. The monitors would be plugged into the Quadro.
Assuming I can even get it to work, can C1 in OS X somehow select a preferred GPU, in this case the R9?
Any other info on how C1 manages GPU's like how it treats the onboard CPU GPU or how it treats dual GPU would be appreciated.
I know the R9 is on the C1 supported list and that's what I plan to use. I'd also like to work toward getting 30 bit color while working in photoshop and maybe in the future, C1, if supported. I plan to use a low end 2GB Quadro video card to get the 30 bit. The monitors would be plugged into the Quadro.
Assuming I can even get it to work, can C1 in OS X somehow select a preferred GPU, in this case the R9?
Any other info on how C1 manages GPU's like how it treats the onboard CPU GPU or how it treats dual GPU would be appreciated.
0
-
CO uses (as a rule of thumb, exceptions occur) all available GPU's.
However, combining a fast and a slow GPU is bad for performance, when compared to just using the fast one.
Say GPU A performs at index 100, and the slow one, GPU B, at index 50. Final "performance" will land somewhere between "faster than GPU B" and "slightly slower than GPU A".
In short, we only recommend using identical cards, when using multiple GPU's.0 -
[quote="Christian Gruner" wrote:
CO uses (as a rule of thumb, exceptions occur) all available GPU's.
and
In short, we only recommend using identical cards, when using multiple GPU's.
How does that relate to laptops with switchable graphics?
The CO OpenCl benchmark on my own laptop report:
2016-01-09 04:55:04.050> OpenCL initialization...
2016-01-09 04:55:04.238> OpenCL : found platform Intel(R) OpenCL, OpenCL Version : OpenCL 1.2
2016-01-09 04:55:04.238> OpenCL : found platform AMD Accelerated Parallel Processing, OpenCL Version : OpenCL 2.0 AMD-APP (1912.5)
2016-01-09 04:55:04.285> OpenCL Device : Intel(R) HD Graphics 4600
2016-01-09 04:55:04.285> OpenCL Driver Version : 10.18.15.4281
2016-01-09 04:55:04.285> OpenCL Compute Units : 20
2016-01-09 04:55:04.363> OpenCL : Loading kernels
2016-01-09 04:55:04.535> OpenCL : Loading kernels finished
2016-01-09 04:55:04.535> OpenCL : Benchmarking
2016-01-09 04:55:05.238> OpenCL : Initialization completed
2016-01-09 04:55:05.238> OpenCL benchMark : 0.790160
2016-01-09 04:55:05.253> OpenCL Device : Oland
2016-01-09 04:55:05.253> OpenCL Driver Version : 1912.5 (VM)
2016-01-09 04:55:05.253> OpenCL Compute Units : 6
2016-01-09 04:55:05.378> OpenCL : Loading kernels
2016-01-09 04:55:05.925> OpenCL : Loading kernels finished
2016-01-09 04:55:05.925> OpenCL : Benchmarking
2016-01-09 04:55:06.257> OpenCL : Initialization completed
2016-01-09 04:55:06.257> OpenCL benchMark : 0.378964
As can be seen, the AMD graphics card is approximately twice as fast as the i7-4800MQ integrated graphics.
Does the mean that the integrated graphics is dragging down the OpenCl performance of this laptop?0 -
That is correct, or at least not making it faster than the faster card. 0
Post is closed for comments.
Comments
3 comments