Compostella, Jeremy d25fd8710d BaseMemoryLibSse2: Take advantage of write combining buffers ... The current SSE2 implementation of the ZeroMem(), SetMem(), SetMem16(), SetMem32 and SetMem64 functions is writing 16 bytes per 16 bytes. It hurts the performances so bad that this is even slower than a simple 'rep stos' (4% slower) in regular DRAM. To take full advantages of the 'movntdq' instruction it is better to "queue" a total of 64 bytes in the write combining buffers. This patch implement such a change. Below is a table where I measured (with 'rdtsc') the time to write an entire 100MB RAM buffer. These functions operate almost two times faster. | Function | Arch | Untouched | 64 bytes | Result | |----------+------+-----------+----------+--------| | ZeroMem | Ia32 | 17765947 | 9136062 | 1.945x | | ZeroMem | X64 | 17525170 | 9233391 | 1.898x | | SetMem | Ia32 | 17522291 | 9137272 | 1.918x | | SetMem | X64 | 17949261 | 9176978 | 1.956x | | SetMem16 | Ia32 | 18219673 | 9372062 | 1.944x | | SetMem16 | X64 | 17523331 | 9275184 | 1.889x | | SetMem32 | Ia32 | 18495036 | 9273053 | 1.994x | | SetMem32 | X64 | 17368864 | 9285885 | 1.870x | | SetMem64 | Ia32 | 18564473 | 9241362 | 2.009x | | SetMem64 | X64 | 17506951 | 9280148 | 1.886x | Signed-off-by: Jeremy Compostella <jeremy.compostella@intel.com> Reviewed-by: Liming Gao <gaoliming@byosoft.com.cn>		2020-10-16 01:12:05 +00:00
..
Ia32	BaseMemoryLibSse2: Take advantage of write combining buffers	2020-10-16 01:12:05 +00:00
X64	BaseMemoryLibSse2: Take advantage of write combining buffers	2020-10-16 01:12:05 +00:00
BaseMemoryLibSse2.inf	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
BaseMemoryLibSse2.uni	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
CompareMemWrapper.c	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
CopyMemWrapper.c	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
IsZeroBufferWrapper.c	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
MemLibGuid.c	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
MemLibInternals.h	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
ScanMem8Wrapper.c	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
ScanMem16Wrapper.c	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
ScanMem32Wrapper.c	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
ScanMem64Wrapper.c	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
SetMem16Wrapper.c	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
SetMem32Wrapper.c	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
SetMem64Wrapper.c	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
SetMemWrapper.c	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00
ZeroMemWrapper.c	MdePkg: Replace BSD License with BSD+Patent License	2019-04-09 10:58:13 -07:00