Displaying 3 results from an estimated 3 matches for "0f706507".
Did you mean:
070507
2013 Feb 04
0
[LLVMdev] [Polly] Parallelizing outer loop containing inner reduction loop
Hi Dmitry,
[FORTRAN code]
> The difference between two versions is only in the way how inner
> reduction loop is implemented. In first case it is explicit and in
> second case - uses sum intrinsic, which is lowered to plain code by
> DragonEgg. In first case Polly successfully detects parallel outer
> loop, in second case - it can't.
Thanks for the test
2013 Feb 04
1
[LLVMdev] [Polly] Parallelizing outer loop containing inner reduction loop
.../ISL can't be aware of such things. I
thought more about pushing AllocaInst-s down closer to the nested array
header - would that work?
Thanks,
- D.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20130204/0f706507/attachment.html>
2013 Feb 03
2
[LLVMdev] [Polly] Parallelizing outer loop containing inner reduction loop
Oops, sorry for the message title, making it more descriptive now...
2013/2/3 Dmitry Mikushin <dmitry at kernelgen.org>
> Dear all,
>
> Yesterday, from the customer's code I observed relatively simple case,
> where Polly is unable to detect parallelizable outer loop. Consider two
> versions of Fortran routine:
>
> 1)
>
> subroutine filter(H, szh, X, szx, Y,