This patch updates code of avx512 support for xbzrle_encode_buffer
function. We modified runtime check of avx512 and simplified algorithm.
Besides, we provide benchmark in xbzrle-bench.c for performance
comparison.
Signed-off-by: ling xu <ling1.xu@intel.com>
Co-authored-by: Zhou Zhao <zhou.zhao@intel.com>
Co-authored-by: Jun Jin <jun.i.jin@intel.com>
ling xu (2):
Update AVX512 support for xbzrle_encode_buffer
Test code and benchmark code
meson.build | 16 ++
meson_options.txt | 2 +
migration/ram.c | 35 ++-
migration/xbzrle.c | 130 +++++++++++
migration/xbzrle.h | 4 +
tests/bench/meson.build | 4 +
tests/bench/xbzrle-bench.c | 468 +++++++++++++++++++++++++++++++++++++
tests/unit/test-xbzrle.c | 39 +++-
8 files changed, 690 insertions(+), 8 deletions(-)
create mode 100644 tests/bench/xbzrle-bench.c
--
2.25.1