BranchCommit messageAuthorAge
Release_v0.9.xBump to 0.9.3.Zhigang Gong5 weeks
masterGBE: disable custom loop unroll for LLVM 3.3/3.4.Zhigang Gong2 days
opencl-1.1relax the build dependency on Gen GPUGuo Yejun4 months
stableBump to 0.9.3.Zhigang Gong5 weeks
Release_v0.9.3commit 80e721cfa6...Zhigang Gong5 weeks
Release_v0.9.2commit 39c8a4331d...Zhigang Gong3 months
Release_v0.9.1commit b140444bd5...Yi Sun4 months
Release_v0.9commit 5ab90babd2...Zhigang Gong4 months
Release_v0.8commit f206e98227...Zhigang Gong8 months
Release_v0.3commit acfac8ab53...Homer Hsing12 months
Release_v0.2commit b512f1e560...Zhigang Gong16 months
Release_v0.1commit e6b503e2b9...Zhigang Gong18 months
AgeCommit messageAuthorFilesLines
2 daysGBE: disable custom loop unroll for LLVM 3.3/3.4.HEADmasterZhigang Gong2-1/+6
5 daysGBE: Add a customized loop unrolling handling mechanism.Zhigang Gong4-2/+243
5 daysGBE: fix a legalize pass bug when bitcast wide integer to incompaitble vector.Zhigang Gong1-4/+12
5 daysRefine the the error handling in function cl_command_queue_ND_range_gen7.Yang Rong1-3/+5
5 daysFix a HSW regression.Yang Rong1-1/+1
5 daysGBE: do intrinsics lowering pass earlier.Zhigang Gong1-0/+1
5 daysFix a upsample regression.Yang Rong5-28/+18
6 daysBDW: Also need set Shader Channel Select for constant buffer in BDW.Yang Rong1-0/+6
6 daysBDW: Fix load/store half error.Yang Rong5-5/+22
7 daysBDW: Change BDW's max work group size to 512.Yang Rong1-3/+5