This revision was automatically updated to reflect the committed changes.
Closed by commit rC336643: [X86] Lowering integer truncation intrinsics to
native IR (authored by mike.dvoretsky, committed by ).
Changed prior to commit:
https://reviews.llvm.org/D48712?vs=153471&id=154765#toc
Repositor
craig.topper accepted this revision.
craig.topper added a comment.
This revision is now accepted and ready to land.
LGTM
https://reviews.llvm.org/D48712
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/li
RKSimon added a comment.
LGTM - @craig.topper any comments?
https://reviews.llvm.org/D48712
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
RKSimon added a comment.
Please can you create a llvm side parallel patch that updates the relevant
fast-isel tests
Comment at: clang/lib/Headers/avx512vlbwintrin.h:1501
+ (__v8qi){0, 0, 0, 0, 0, 0, 0, 0}, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11,
+ 12, 13, 14, 15);
}
-
mike.dvoretsky added inline comments.
Comment at: clang/lib/Headers/avx512vlbwintrin.h:1501
+ (__v8qi){0, 0, 0, 0, 0, 0, 0, 0}, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11,
+ 12, 13, 14, 15);
}
RKSimon wrote:
> Are we happy with using illegal types like this?
RKSimon added inline comments.
Comment at: clang/lib/Headers/avx512vlbwintrin.h:1501
+ (__v8qi){0, 0, 0, 0, 0, 0, 0, 0}, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11,
+ 12, 13, 14, 15);
}
Are we happy with using illegal types like this? What about flipping the
mike.dvoretsky updated this revision to Diff 153471.
mike.dvoretsky marked 2 inline comments as done.
mike.dvoretsky added a comment.
Updated per comments. Typedefs for intermediate short vectors moved into the
bodies of the functions using them.
https://reviews.llvm.org/D48712
Files:
clang/
craig.topper added inline comments.
Comment at: clang/lib/Headers/avx512vlintrin.h:33
+typedef short __v2hi __attribute__((__vector_size__(4)));
+typedef char __v4qi __attribute__((__vector_size__(4)));
Can you just do a local typedef in the functions that need
craig.topper added inline comments.
Comment at: clang/lib/Headers/avx512vlintrin.h:7421
+ __builtin_convertvector((__v4si)__A, __v4qi), (__v4qi){0, 0, 0, 0}, 0, 1,
+ 2, 3, 4, 5, 6, 7, 7, 7, 7, 7, 7, 7, 7, 7);
}
If you need to add more zeroes than the w
mike.dvoretsky updated this revision to Diff 153277.
mike.dvoretsky added a comment.
Uploaded the correct diff.
https://reviews.llvm.org/D48712
Files:
clang/lib/Headers/avx512vlbwintrin.h
clang/lib/Headers/avx512vlintrin.h
clang/test/CodeGen/avx512vl-builtins.c
clang/test/CodeGen/avx512
mike.dvoretsky created this revision.
mike.dvoretsky added a reviewer: craig.topper.
Herald added a subscriber: cfe-commits.
This patch lowers the _mm[256|512]_cvtepi{64|32|16}_epi{32|16|8} intrinsics to
native IR in cases where the result's length is less than 128 bits.
The resulting IR is fold
11 matches
Mail list logo
