PATCHES: R600/SI: CodeGen patches for HSA Runtime

Matt Arsenault Matthew.Arsenault at amd.com
Wed Nov 12 11:52:53 PST 2014


On 11/12/2014 09:24 AM, Tom Stellard wrote:
> On Fri, Nov 07, 2014 at 01:45:05PM -0800, Matt Arsenault wrote:
>> >On 11/07/2014 01:31 PM, Tom Stellard wrote:
>>> > >On Thu, Nov 06, 2014 at 03:01:53PM -0800, Matt Arsenault wrote:
>>>>> > >>>On 11/06/2014 02:23 PM, Tom Stellard wrote:
>>>>>>> > >>>> >Hi,
>>>>>>> > >>>> >
>>>>>>> > >>>> >Attached are patches to enable the R600 backend to emit code for the HSA runtime.
>>>>>>> > >>>> >Please review.
>>>>>>> > >>>> >
>>>>>>> > >>>> >-Tom
>>>>>>> > >>>> >
>>> > >Hi,
>>> > >
>>> > >Here are some updated patches.  I'm going to leave all the alignments at 0
>>> > >until I  can figure out exactly what they are for.
>>> > >
>>> > >-Tom
> Hi Matt,
>
> Here are patches with the refactoring changes and the improved AmdKernelCodeT
> debugging output.
>
> -Tom
>

LGTM

> diff --git a/lib/Target/R600/AMDKernelCodeT.h b/lib/Target/R600/AMDKernelCodeT.h
> new file mode 100644
> index 0000000..5b25582
> --- /dev/null
> +++ b/lib/Target/R600/AMDKernelCodeT.h
> @@ -0,0 +1,704 @@
> +//===-- AMDGPUKernelCodeT.h - Print AMDGPU assembly code ---------*- C++ -*-===//
> +//
> +//                     The LLVM Compiler Infrastructure
> +//
> +// This file is distributed under the University of Illinois Open Source
> +// License. See LICENSE.TXT for details.
> +//
> +//===----------------------------------------------------------------------===//
> +/// \file AMDKernelCodeT.h
> +//===----------------------------------------------------------------------===//
> +
> +#ifndef AMDKERNELCODET_H
> +#define AMDKERNELCODET_H
> +
> +#include <stddef.h>
> +#include <stdint.h>
<cstddef>
<cstdint>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20141112/63823232/attachment.html>


More information about the llvm-commits mailing list