[Bug target/82460] AVX512: choose between vpermi2d and vpermt2d to save mov instructions. Also, fails to optimize away shifts before shuffle

["jakub at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82460

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Author: jakub
Date: Tue Oct 24 19:35:37 2017
New Revision: 254059

URL: https://gcc.gnu.org/viewcvs?rev=254059&root=gcc&view=rev
Log:
        PR target/82460
        * config/i386/sse.md (UNSPEC_VPERMI2, UNSPEC_VPERMI2_MASK): Remove.
        (VPERMI2, VPERMI2I): New mode iterators.
        (<avx512>_vpermi2var<mode>3_maskz): Remove 3 define_expand patterns.
        (<avx512>_vpermi2var<mode>3<sd_maskz_name>): Remove 3 define_insn
        patterns.
        (<avx512>_vpermi2var<mode>3_mask): New define_expand using VPERMI2
        mode iterator.  Remove 3 old define_insn patterns.
        (*<avx512>_vpermi2var<mode>3_mask): 2 new define_insn patterns.
        (<avx512>_vpermt2var<mode>3_maskz): Adjust 1 define_expand to use
        VPERMI2 mode iterator, remove the other two expanders.
        (<avx512>_vpermt2var<mode>3<sd_maskz_name>): Adjust 1 define_insn
        to use VPERMI2 mode iterator, add another alternative for vpermi2*
        instructions, remove the other two patterns.
        (<avx512>_vpermt2var<mode>3_mask): Adjust 1 define_insn to use VPERMI2
        mode iterator, remove the other two patterns.
        * config/i386/i386.c (ix86_expand_vec_perm_vpermi2): Renamed to ...
        (ix86_expand_vec_perm_vpermt2): ... this.  Swap mask and op0
        arguments, use gen_*vpermt2* expanders instead of gen_*vpermi2*
        and adjust argument order accordingly.
        (ix86_expand_vec_perm): Adjust caller.
        (expand_vec_perm_1): Likewise.
        (expand_vec_perm_vpermi2_vpshub2): Rename to ...
        (expand_vec_perm_vpermt2_vpshub2): ... this.
        (ix86_expand_vec_perm_const_1): Adjust caller.
        (ix86_vectorize_vec_perm_const_ok): Adjust comments.

        * gcc.target/i386/pr82460-1.c: New test.
        * gcc.target/i386/pr82460-2.c: New test.
        * gcc.target/i386/avx512f-vpermt2pd-1.c: Adjust scan-assembler*
        regexps to allow vpermt2* to vpermi2* replacement or vice versa
        where possible.
        * gcc.target/i386/avx512vl-vpermt2pd-1.c: Likewise.
        * gcc.target/i386/avx512f-vpermt2d-1.c: Likewise.
        * gcc.target/i386/vect-pack-trunc-2.c: Likewise.
        * gcc.target/i386/avx512vl-vpermt2ps-1.c: Likewise.
        * gcc.target/i386/avx512vl-vpermt2q-1.c: Likewise.
        * gcc.target/i386/avx512f-vpermt2ps-1.c: Likewise.
        * gcc.target/i386/avx512vl-vpermt2d-1.c: Likewise.
        * gcc.target/i386/avx512bw-vpermt2w-1.c: Likewise.
        * gcc.target/i386/avx512vbmi-vpermt2b-1.c: Likewise.
        * gcc.target/i386/avx512f-vpermt2q-1.c: Likewise.

Added:
    trunk/gcc/testsuite/gcc.target/i386/pr82460-1.c
    trunk/gcc/testsuite/gcc.target/i386/pr82460-2.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/i386/i386.c
    trunk/gcc/config/i386/sse.md
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/gcc.target/i386/avx512bw-vpermt2w-1.c
    trunk/gcc/testsuite/gcc.target/i386/avx512f-vpermt2d-1.c
    trunk/gcc/testsuite/gcc.target/i386/avx512f-vpermt2pd-1.c
    trunk/gcc/testsuite/gcc.target/i386/avx512f-vpermt2ps-1.c
    trunk/gcc/testsuite/gcc.target/i386/avx512f-vpermt2q-1.c
    trunk/gcc/testsuite/gcc.target/i386/avx512vbmi-vpermt2b-1.c
    trunk/gcc/testsuite/gcc.target/i386/avx512vl-vpermt2d-1.c
    trunk/gcc/testsuite/gcc.target/i386/avx512vl-vpermt2pd-1.c
    trunk/gcc/testsuite/gcc.target/i386/avx512vl-vpermt2ps-1.c
    trunk/gcc/testsuite/gcc.target/i386/avx512vl-vpermt2q-1.c
    trunk/gcc/testsuite/gcc.target/i386/vect-pack-trunc-2.c