Re: [PATCH] move x86 to use gather/scatter internal functions

2021-08-17 Thread Hongtao Liu via Gcc-patches
On Wed, Aug 18, 2021 at 11:24 AM Hongtao Liu wrote: > > On Tue, Aug 17, 2021 at 10:43 PM Richard Biener via Gcc-patches > wrote: > > > > On Tue, Aug 17, 2021 at 3:29 PM Richard Biener via Gcc-patches > > wrote: > > > > > > This is an attempt to start moving the x86 backend to use > > > standard

Re: [PATCH] [i386] Add x86 tune to enable v2df vector reduction by paddpd.

2021-08-17 Thread Hongtao Liu via Gcc-patches
On Tue, Aug 17, 2021 at 5:06 PM liuhongt wrote: > > Hi: > This patch add a new x86 tune named X86_TUNE_V2DF_REDUCTION_PREFER_HADDPD > to enable haddpd for v2df vector reduction, the tune is disabled by default. > > Bootstrapped and regtested on x86_64-linux-gnu{-m32,} > Ok for trunk? > Pushe

Re: Optional machine prefix for programs in for -B dirs

2021-08-17 Thread John Ericson
OK I have polished off my code in light of previous discussion and will submit it in follow-up emails. As mentioned before, this patch series is on top of the non-behavior-changing cleanup I previously submitted in https://gcc.gnu.org/pipermail/gcc-patches/2021-August/576725.html The first patch

[PATCH 3/3] find_a_program: Only search for prefixed paths in undisambiguated dirs

2021-08-17 Thread John Ericson
This means, we might search for: - path/$machine/$version/prog - path/$machine/prog - path/$machine-prog But not - path/$machine/$version/$machine-prog because disambiguating $machine twice is unnecessary. This does mean we less liberal in what we accept than LLVM, but that's OK. The down side

[PATCH 1/3] find_a_program: First search with machine prefix

2021-08-17 Thread John Ericson
This matches the behavior of Clang, and makes it easier to work with cross compilers without heeding to hard-code paths at build time. --- gcc/gcc.c | 78 --- 1 file changed, 68 insertions(+), 10 deletions(-) diff --git a/gcc/gcc.c b/gcc/gcc.c i

[PATCH 2/3] driver: for_each_pass: Pass to callback whether dir is machine-disambiguated

2021-08-17 Thread John Ericson
We will use this in the subsequent diff to control what basenames we search for. In machine-specific subdirectories, we should just look for the original basename, but in machine-agnostic subdirectories, we might additionally look for prefixed disambiguated names, as an alternate method of keeping

<    1   2