LGTM, will push latter, thanks.
On Wed, Apr 23, 2014 at 10:56:50AM +0800, Ruiling Song wrote:
Untyped read seems better than byte gather read.
Some performance test in opencv got doubled after the patch.
Signed-off-by: Ruiling Song ruiling.s...@intel.com
---
LGTM, will push latter, thanks.
On Wed, Apr 23, 2014 at 02:31:29PM +0800, Ruiling Song wrote:
1. This would free unused register asap, so it becomes easy to allocate
contiguous registers.
2. We previously met many hidden register liveness issue. Let's try
to reuse the expired register
From: Junyan He junyan...@linux.intel.com
We use the floatn's assigment to do the copy.
128 pattern size is according to double16, and because
the double problem on our platform, we use to float16
to handle this.
unaligned cases is not optimized now, just use the char
assigment.
Signed-off-by:
From: Junyan He junyan...@linux.intel.com
Signed-off-by: Junyan He junyan...@linux.intel.com
---
utests/CMakeLists.txt | 1 +
utests/enqueue_fill_buf.cpp | 90 +
2 files changed, 91 insertions(+)
create mode 100644 utests/enqueue_fill_buf.cpp
Signed-off-by: Guo Yejun yejun@intel.com
---
CMake/FindLLVM.cmake | 2 +-
backend/src/CMakeLists.txt | 8
2 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/CMake/FindLLVM.cmake b/CMake/FindLLVM.cmake
index 97ee7db..556b3a9 100644
--- a/CMake/FindLLVM.cmake
+++