llvmpipe: Fallback to element-wise comparisons when no comparison intrinsic is available.
authorJosé Fonseca <jfonseca@vmware.com>
Mon, 31 Aug 2009 09:22:36 +0000 (10:22 +0100)
committerJosé Fonseca <jfonseca@vmware.com>
Mon, 31 Aug 2009 09:22:36 +0000 (10:22 +0100)
commit241c3a1d8001fc5a30e2af4b4636b48e6f99690a
tree2a14a75e5e10c450cfb441bc5e55aba4df378653
parent45fb66ab7bc1cbb150e055826dc61b542739cc35
llvmpipe: Fallback to element-wise comparisons when no comparison intrinsic is available.

Although selection of vector elements is valid LLVM IR, no machine target
supports it yet.

This is a last-resort option, but it allows llvmpipe to be used on any
target supported by LLVM without modifications. Obviously better
performance is attainable by emitting SIMD intrinsics where otherwise
LLVM doesn't.
src/gallium/drivers/llvmpipe/lp_bld_logic.c