index
:
~gongzg/beignet
1.0rc
OpenCL20
Release_v0.9.x
Release_v1.0
blender
boolbug
double
draft_function_call
enqueue
fix
fixliveness
fixsched
forhsw
gl_sharing
hack_egl_display
hsw_darktable
image_refine
llvm-3.5
loop_opt
loop_unroll
loopspill
master
merge_gl_sharing
mixsimd16
opencl-1.1
optimize
optimize2
prepare_spill16
spillsimd16
stable
test_math
userptr
vectoropt
vectoropt_test
vload_opt
Unnamed repository; edit this file 'description' to name the repository.
UNKNOWN
summary
refs
log
tree
commit
diff
log msg
author
committer
range
Age
Commit message (
Expand
)
Author
Files
Lines
2014-11-03
GBE: remove useless debug info.
vectoropt
Zhigang Gong
1
-12
/
+2
2014-11-03
GBE: One simple redundant instruction elimination.
Zhigang Gong
1
-0
/
+46
2014-11-03
GBE: fix disassembly bug.
Zhigang Gong
1
-28
/
+0
2014-11-03
GBE: further optimize vector mov.
Zhigang Gong
1
-11
/
+25
2014-11-03
GBE: analyze vector intereference to save some extra MOV if possible.
Zhigang Gong
1
-10
/
+73
2014-11-03
GBE: workaround register allocation fail caused by custom loop unroll.
Ruiling Song
1
-7
/
+12
2014-11-03
add utest for llvm intrinsic call usub_with_overflow funtion.
Luo Xionghu
2
-35
/
+117
2014-11-03
enable llvm intrinsic call usub_with_overflow funtion.
Luo Xionghu
1
-1
/
+15
2014-11-03
use env to set environment variables for GBE_BIN_GENERATER
Andreas Beckmann
1
-1
/
+1
2014-11-03
fix some typos
Andreas Beckmann
2
-4
/
+4
2014-10-31
Remove intel_gpgpu_check_binded_buf_address()
Zhenyu Wang
1
-9
/
+10
2014-10-29
IVB/HSW/BYT: Revert the Dynamic state Base Addr and relative buffers address ...
Yang Rong
1
-7
/
+45
2014-10-29
BDW: Fix bwd 32*32 scalar multiplication bug.
Yang Rong
1
-6
/
+4
2014-10-28
Fix a size assert when setup bti.
Yang Rong
2
-2
/
+14
2014-10-28
Refine the intel gpgpu delete.
Yang Rong
5
-82
/
+169
2014-10-28
Fix the compare_image_2d_and_1d_array test case bug
Junyan He
1
-5
/
+16
2014-10-28
add utest compiler_overflow for llvm intrinsic function.
Luo Xionghu
3
-0
/
+93
2014-10-28
add llvm Intrinsic call support.
Luo Xionghu
1
-2
/
+64
2014-10-27
GBE: fix a wrong type of cl_device_info.
Zhigang Gong
1
-1
/
+1
2014-10-27
GBE: set default address space to -1 to avoid incorrect unroll hint.
Zhigang Gong
1
-1
/
+1
2014-10-24
Fix AUX buffer for page alignment
Zhenyu Wang
1
-3
/
+6
2014-10-24
Use pread/pwrite for buffer enqueue read/write
Zhenyu Wang
4
-26
/
+10
2014-10-24
Clear batch buffer pointer after unmap
Zhenyu Wang
1
-1
/
+2
2014-10-24
Make use of write enable flag for mem bo map
Zhenyu Wang
6
-22
/
+27
2014-10-24
GBE: handle dead loop BBs in liveness analysis.
Zhigang Gong
2
-1
/
+11
2014-10-24
create GIT_SHA1 without any dependency
Meng Mengmeng
2
-5
/
+4
2014-10-23
Add the disasm support for Gen8
Junyan He
2
-530
/
+582
2014-10-23
GBE: fix regression caused by simple block optimization.
Zhigang Gong
1
-0
/
+28
2014-10-23
add beignet GIT_HAL1 if there is .git directory
Meng Mengmeng
3
-2
/
+35
2014-10-23
GBE: add basic PHINode support in legalize pass.
Zhigang Gong
1
-0
/
+66
2014-10-23
GBE: increase maximum src/dst operands to 32.
Zhigang Gong
4
-9
/
+9
2014-10-23
GBE: add Selection instruction handler at legalize pass.
Zhigang Gong
1
-0
/
+54
2014-10-23
Add the test case for image 2d array fill
Junyan He
3
-0
/
+98
2014-10-23
Add the test case for image 1d array fill
Junyan He
3
-0
/
+85
2014-10-23
Fix the bug of 1D array slice pitch
Junyan He
5
-14
/
+50
2014-10-23
use global flag 0.0 to control unstructured simple block.
Luo Xionghu
3
-23
/
+88
2014-10-20
GBE: disable custom loop unroll for LLVM 3.3/3.4.
Zhigang Gong
2
-1
/
+6
2014-10-17
GBE: Add a customized loop unrolling handling mechanism.
Zhigang Gong
4
-2
/
+243
2014-10-17
GBE: fix a legalize pass bug when bitcast wide integer to incompaitble vector.
Zhigang Gong
1
-4
/
+12
2014-10-17
Refine the the error handling in function cl_command_queue_ND_range_gen7.
Yang Rong
1
-3
/
+5
2014-10-17
Fix a HSW regression.
Yang Rong
1
-1
/
+1
2014-10-17
GBE: do intrinsics lowering pass earlier.
Zhigang Gong
1
-0
/
+1
2014-10-17
Fix a upsample regression.
Yang Rong
5
-28
/
+18
2014-10-16
BDW: Also need set Shader Channel Select for constant buffer in BDW.
Yang Rong
1
-0
/
+6
2014-10-16
BDW: Fix load/store half error.
Yang Rong
5
-5
/
+22
2014-10-15
BDW: Change BDW's max work group size to 512.
Yang Rong
1
-3
/
+5
2014-10-15
Fix a HSW constant buffer regression.
Yang Rong
1
-1
/
+35
2014-10-15
GBE: fix a bug in legalize pass.
Zhigang Gong
1
-7
/
+5
2014-10-15
Fit the printf bug in loop
Junyan He
4
-141
/
+283
2014-10-14
Fix HSW thread_n <= 64 assert.
Yang Rong
1
-3
/
+7
[next]