Advanced search

Message boards : Wish list : AMD port using HIP

Author Message
ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2674
Credit: 894,555,724
RAC: 565,665
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 45307 - Posted: 17 Nov 2016 | 21:42:00 UTC
Last modified: 17 Nov 2016 | 21:42:57 UTC

I know we've talked about AMD GPU support a lot and so far it has not worked out well. Maybe their new "HIP" in the Open Compute / Boltzmann / ROCm 1.3 initiative could be worth a new try? Usig it to port CAFFE from CUDA to AMD-something was presumably surprisingly easy and faster than OpenCL out of the box.

MrS
____________
Scanning for our furry friends since Jan 2002

[VENETO] boboviz
Send message
Joined: 10 Sep 10
Posts: 105
Credit: 253,107
RAC: 44
Level

Scientific publications
wat
Message 46412 - Posted: 31 Jan 2017 | 10:21:18 UTC - in response to Message 45307.

Multiscale Lab, the lab of gpugrid, some times ago published this tool: SWAN, porting cuda to opencl...

[VENETO] boboviz
Send message
Joined: 10 Sep 10
Posts: 105
Credit: 253,107
RAC: 44
Level

Scientific publications
wat
Message 47195 - Posted: 11 May 2017 | 16:05:35 UTC - in response to Message 45307.

Maybe their new "HIP" in the Open Compute / Boltzmann / ROCm 1.3 initiative could be worth a new try?


Now ROCm is at 1.5 version.....(with full support to opencl)

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2674
Credit: 894,555,724
RAC: 565,665
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 47386 - Posted: 7 Jun 2017 | 21:41:05 UTC

Apparently there is another way to compile CUDA codes to OpenCL. I understand the project team is pretty busy keeping up with the current crunching power, though.

MrS
____________
Scanning for our furry friends since Jan 2002

[VENETO] boboviz
Send message
Joined: 10 Sep 10
Posts: 105
Credit: 253,107
RAC: 44
Level

Scientific publications
wat
Message 48073 - Posted: 30 Oct 2017 | 14:49:11 UTC - in response to Message 45307.

Maybe their new "HIP" in the Open Compute / Boltzmann / ROCm 1.3 initiative could be worth a new try?


With the latest version (1.4) there are a lot of new features, bugfix, documentation, etc
HIP

[VENETO] boboviz
Send message
Joined: 10 Sep 10
Posts: 105
Credit: 253,107
RAC: 44
Level

Scientific publications
wat
Message 48784 - Posted: 2 Feb 2018 | 17:36:49 UTC

Very interesting
rocRAND

[VENETO] boboviz
Send message
Joined: 10 Sep 10
Posts: 105
Credit: 253,107
RAC: 44
Level

Scientific publications
wat
Message 49240 - Posted: 6 Apr 2018 | 8:54:34 UTC

Another interesting news about converting Cuda into OpenCl
LLVM Cuda to HIP

[VENETO] boboviz
Send message
Joined: 10 Sep 10
Posts: 105
Credit: 253,107
RAC: 44
Level

Scientific publications
wat
Message 49241 - Posted: 6 Apr 2018 | 9:09:20 UTC - in response to Message 48073.

With the latest version (1.4) there are a lot of new features, bugfix, documentation, etc
HIP


Uh, i forget the new version of HIP (1.5) with new features


- Support threadIdx, blockIdx, blockDim directly (no need for hipify conversions in kernels.) HIP Kernel syntax is now identical to CUDA kernel syntax - no need for extra parms or conversions.
- Remove need for manual destructors
- Provide printf in device code
- Support for globals when using module API
- hipify-clang now supports using newer versions of clang
- HIP texture support equivalent to CUDA texture driver APIs

etc

Post to thread

Message boards : Wish list : AMD port using HIP