Re: [PATCH] Handle in-order reductions when SLP vectorizing non-loops

2023-08-09 Thread Jeff Law via Gcc-patches
On 8/9/23 07:51, Alexander Monakov wrote: On Wed, 9 Aug 2023, Richard Biener via Gcc-patches wrote: The following teaches the non-loop reduction vectorization code to handle non-associatable reductions. Using the existing FOLD_LEFT_PLUS internal functions might be possible but I'd have to

Re: [PATCH] Handle in-order reductions when SLP vectorizing non-loops

2023-08-09 Thread Alexander Monakov
On Wed, 9 Aug 2023, Richard Biener via Gcc-patches wrote: > The following teaches the non-loop reduction vectorization code to > handle non-associatable reductions. Using the existing FOLD_LEFT_PLUS > internal functions might be possible but I'd have to convince myself > that +0.0 + x[0] is a s

[PATCH] Handle in-order reductions when SLP vectorizing non-loops

2023-08-09 Thread Richard Biener via Gcc-patches
The following teaches the non-loop reduction vectorization code to handle non-associatable reductions. Using the existing FOLD_LEFT_PLUS internal functions might be possible but I'd have to convince myself that +0.0 + x[0] is a safe extra operation in ever rounding mode (I also have no way to test