|
|
#1 | |
|
Registered User
Join Date: Nov 2009
Posts: 2
|
I have two systems, one with OSX 10.6.1 and one with a form of debian Linux.
On OSX, compiling my OpenCL program takes around two seconds, but on my Linux box, it takes nearly 18! My OpenCL program has many kernels inside of it, I was wonder why there is such a huge difference in compile times and if there is a way to fix it. |
|
|
|
|
|
|
#2 | |
|
Registered User
Join Date: Oct 2009
Location: Stockholm
Posts: 33
|
The compiler runs on the Intel processor, not on the video card ...
|
|
|
|
|
|
|
#3 | |
|
Registered User
Join Date: Aug 2004
Posts: 301
|
Quote:
|
|
|
|
|
|
|
#4 | |
|
Registered User
Join Date: Nov 2009
Posts: 2
|
jma, I know that it runs on the CPU when compiling....
Deanjo, Should be. I guess there is a massive speed difference between OSX's LLVM/Clang compiling and whatever is being used in Linux? |
|
|
|
|
|
|
#5 |
|
Registered User
Join Date: Dec 2007
Posts: 64
|
Sonicadvance1 everything depends on your settings (optimization and compilation flags) and core you're compiling.
If you're using LLVM/Clang for C code - it should be faster than gcc. Also as far as I'm aware neither LLVM/Clang not gcc can't take advantage of OpenCL yet (correct me if I'm wrong here). |
|
|
|
![]() |
| Thread Tools | |
|
|