KonstantinSeurer
/
mesa
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

gallivm: Hide AVX support when requested by LP_NATIVE_VECTOR_WIDTH or unsupported by LLVM. 

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
This commit is contained in:
José Fonseca 2012-10-17 15:38:50 +01:00
parent 54821c0e99
commit aa2067c757
1 changed files with 10 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
src/gallium/auxiliary/gallivm/lp_bld_init.c
View File

@ -458,6 +458,15 @@ lp_build_init(void)
lp_native_vector_width = debug_get_num_option("LP_NATIVE_VECTOR_WIDTH",
lp_native_vector_width);
if (lp_native_vector_width <= 128) {
/* Hide AVX support, as often LLVM AVX instrinsics are only guarded by
* "util_cpu_caps.has_avx" predicate, and lack the
* "lp_native_vector_width > 128" predicate. And also to ensure a more
* consistent behavior, allowing one to test SSE2 on AVX machines.
*/
util_cpu_caps.has_avx = 0;
}
gallivm_initialized = TRUE;
#if 0
@ -465,6 +474,7 @@ lp_build_init(void)
util_cpu_caps.has_sse3 = 0;
util_cpu_caps.has_ssse3 = 0;
util_cpu_caps.has_sse4_1 = 0;
util_cpu_caps.has_avx = 0;
#endif
}