[PATCH] D60610: [X86][SSE] Recognise vXi1 boolean anyof/allof reduction patterns

[Sanjay Patel via Phabricator via llvm-commits]

spatel accepted this revision.
spatel added a comment.
This revision is now accepted and ready to land.

LGTM - see inline for a minor potential follow-up.

If we can assume vXi1 IR, then I can probably abandon my recent movmsk efforts:
D59669 <https://reviews.llvm.org/D59669>
D59912 <https://reviews.llvm.org/D59912>



================
Comment at: test/CodeGen/X86/vector-compare-all_of.ll:1295-1296
 ; AVX1-NEXT:    vpcmpgtq %xmm0, %xmm1, %xmm0
-; AVX1-NEXT:    vpackssdw %xmm2, %xmm0, %xmm1
-; AVX1-NEXT:    vpackssdw %xmm0, %xmm2, %xmm0
-; AVX1-NEXT:    vpand %xmm1, %xmm0, %xmm0
-; AVX1-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[1,1,2,3]
-; AVX1-NEXT:    vpand %xmm0, %xmm1, %xmm0
-; AVX1-NEXT:    vpextrb $0, %xmm0, %eax
-; AVX1-NEXT:    # kill: def $al killed $al killed $eax
+; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
+; AVX1-NEXT:    vmovmskpd %ymm0, %eax
+; AVX1-NEXT:    cmpb $15, %al
----------------
I think the other path prefers to use packss and a 128-bit movmsk for AVX1 here which could be a slight win since it avoids ymm?


Repository:
  rL LLVM

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D60610/new/

https://reviews.llvm.org/D60610