summaryrefslogtreecommitdiff
AgeCommit message (Expand)AuthorFilesLines
2016-05-27i965: Expose GL 4.3 on Gen8+.i965-simd32-csFrancisco Jerez2-2/+2
2016-05-27i965/fs: Skip gen4 pre/post-send dependency workaronds for the first/last block.Francisco Jerez1-2/+2
2016-05-27i965/fs: Allow scalar source regions on SNB math instructions.Francisco Jerez3-17/+8
2016-05-27i965/fs: Fix constant combining for instructions that cannot accept source mods.Francisco Jerez1-4/+3
2016-05-27i965/fs: Extend remove_duplicate_mrf_writes() to handle non-VGRF to MRF copies.Francisco Jerez1-8/+8
2016-05-27i965/fs: Extend compute_to_mrf() to coalesce VGRFs initialized by multiple si...Francisco Jerez1-10/+36
2016-05-27i965/fs: Teach compute_to_mrf about the COMPR4 address transformation.Francisco Jerez1-1/+23
2016-05-27i965/fs: Refactor compute_to_mrf() to split search and rewrite into separate ...Francisco Jerez1-10/+25
2016-05-27i965/fs: Fix compute-to-mrf VGRF region coverage condition.Francisco Jerez1-3/+6
2016-05-27i965/fs: Simplify and improve accuracy of compute_to_mrf() by using regions_o...Francisco Jerez1-47/+13
2016-05-27i965/fs: Teach regions_overlap() about COMPR4 MRF regions.Francisco Jerez1-3/+17
2016-05-27i965/fs: Reindent emit_zip().Francisco Jerez1-14/+14
2016-05-27i965/fs: Skip SIMD lowering destination zipping if possible.Francisco Jerez1-0/+81
2016-05-27i965/fs: Skip SIMD lowering source unzipping for regular scalar regions.Francisco Jerez1-0/+7
2016-05-27i965/fs: Factor out region zipping and unzipping from the SIMD lowering pass.Francisco Jerez1-60/+84
2016-05-27i965/fs: Generalize regions_overlap() from copy propagation to handle non-VGR...Francisco Jerez2-12/+42
2016-05-27i965/fs: Refactor offset() into a separate function taking the width as argum...Francisco Jerez2-16/+21
2016-05-27SQUASH: i965/fs: Add basic dataflow check to opt_sampler_eot().Francisco Jerez2-0/+15
2016-05-27i965/fs: Implement opt_sampler_eot() in terms of logical sends.Francisco Jerez1-56/+23
2016-05-27i965/fs: Fix UB list sentinel dereference in opt_sampler_eot().Francisco Jerez1-6/+9
2016-05-27i965/fs: Take opt_redundant_discard_jumps out of the optimization loop.Francisco Jerez1-2/+1
2016-05-27i965/fs: Run SIMD and logical send lowering after the optimization loop.Francisco Jerez1-4/+25
2016-05-27i965/fs: Add FS_OPCODE_FB_WRITE_LOGICAL to has_side_effects().Francisco Jerez1-0/+1
2016-05-27i965/fs: Allow constant propagation into logical send sources.Francisco Jerez1-0/+34
2016-05-27i965/fs: Let CSE handle logical sampler sends as expressions.Francisco Jerez1-0/+13
2016-05-27i965/fs: Pass a BAD_FILE register to the logical FB write when oMask is unused.Francisco Jerez2-2/+3
2016-05-27i965: Update compute workgroup size limit calculation for SIMD32.Francisco Jerez1-11/+3
2016-05-27i965: Add do32 debug option.Francisco Jerez3-1/+3
2016-05-27i965/fs: Build 32-wide compute shader when needed.Francisco Jerez1-0/+26
2016-05-27i965/fs: Extend back-end interface for limiting the shader dispatch width.Francisco Jerez3-23/+22
2016-05-27i965/fs: Implement SIMD32 register allocation support.Francisco Jerez3-8/+9
2016-05-27i965/fs: Remove pre-Gen7 register allocation class micro-optimization.Francisco Jerez1-18/+3
2016-05-27i965/fs: Don't mutate multi-component arguments in sampler payload set-up.Francisco Jerez1-58/+32
2016-05-27i965/fs: Fix multiple ACP interference during copy propagation.Francisco Jerez1-6/+2
2016-05-27i965/fs: Fix cmod propagation not to propagate non-identity cmod into CMP(N).Francisco Jerez1-0/+12
2016-05-27i965/fs: Estimate number of registers written correctly in opt_register_renam...Francisco Jerez1-2/+2
2016-05-27i965/fs: Add (sub)reg_offset asserts to brw_reg_from_fs_reg.Francisco Jerez1-0/+2
2016-05-27i965/fs: Reset reg_offset of the original destination to zero in compute_to_m...Francisco Jerez1-0/+1
2016-05-27i965/fs: Skip remove_duplicate_mrf_writes() during SIMD32 runs.Francisco Jerez1-1/+1
2016-05-27i965/fs: Use SIMD8 SSBO GET_BUFFER_SIZE message regardless of the dispatch wi...Francisco Jerez1-22/+18
2016-05-27i965/fs: Don't emit duplicated SSBO GET_BUFFER_SIZE instruction unnecessarily.Francisco Jerez1-1/+0
2016-05-27i965/fs: Emit fixed width memory fence opcode regardless of the dispatch width.Francisco Jerez1-2/+3
2016-05-27i965/fs: Return 32 bit mask from fs_builder::sample_mask().Francisco Jerez1-1/+3
2016-05-27i965/fs: Emit fixed-width null register regardless of the dispatch width.Francisco Jerez1-8/+4
2016-05-27i965/fs: Fix half() to handle more exotic register files.Francisco Jerez1-21/+4
2016-05-27i965/fs: Fix horiz_offset() to handle ARF and HW GRF register files.Francisco Jerez1-4/+10
2016-05-27i965/fs: Clean up remaining uses of fs_inst::reads_flag and ::writes_flag.Francisco Jerez5-24/+12
2016-05-27i965/fs: Keep track of flag dependencies with byte granularity during schedul...Francisco Jerez1-10/+31
2016-05-27i965/fs: Track flag register liveness with byte granularity.Francisco Jerez2-25/+9
2016-05-27i965/fs: Define methods to calculate the flag subset read or written by an fs...Francisco Jerez2-11/+67
generated by cgit v1.2.3 (git 2.39.1) at 2024-06-08 12:20:45 +0000