Branch | Commit message | Author | Age |
add-f-d-d | i965/fs: add :f :d :d | Matt Turner | 10 years |
arb_gpu_shader5 | i965/fs: Add a peephole pass to combine mul/imul_high. | Matt Turner | 12 years |
ballot | nir: Narrow unnecessary 64-bit operations to 32-bits | Matt Turner | 8 years |
bug-109404 | intel/compiler: Avoid propagating inequality cmods if types are different | Matt Turner | 6 years |
compact | i965: Replace illegal compacted NOP with valid compact instruction | Zhenyu Wang | 10 years |
dead/csel | CSEL | Matt Turner | 10 years |
dead/double-scheduler | Works? | Matt Turner | 9 years |
dead/g45-transposed-read | i965: Add support for URB transposed reads present on g45 through ILK. | Eric Anholt | 9 years |
dead/i965-sched-v3 | i965/sched: use Sethi-Ullman numbering | Connor Abbott | 9 years |
dead/lround | mesa: Optimize (UN)CLAMPED_FLOAT_TO_USHORT macros. | Matt Turner | 10 years |
dead/lround2 | mesa: Close fp on error path. | Matt Turner | 9 years |
dead/lvn | WIP | Matt Turner | 9 years |
dead/nir-vec4-rc | fuck! | Matt Turner | 10 years |
dead/pspom | i965/fs: Recognize constants can be loaded by subtracting from 1.0. | Matt Turner | 8 years |
dead/spf | i965/blorp: Enable SPF when the shader contains no control flow. | Matt Turner | 11 years |
dead/validate-old | ! | Matt Turner | 10 years |
dead/vec4-neg-mul-cse | i965/vec4: Allow CSE to handle MULs with negated arguments. | Matt Turner | 8 years |
dead/vec4-rc | i965/vec4: Rewrite register coalescing pass. | Matt Turner | 10 years |
dead/vector-logic | glsl: Add vector logic extension | Matt Turner | 9 years |
deferred/locking | mesa: Replace uses of Shared->Mutex with hash-table mutexes | Matt Turner | 10 years |
deferred/no-peephole-sel | i965/fs: Remove fs_visitor::opt_peephole_sel(). | Matt Turner | 10 years |
double-scheduler | Works? | Matt Turner | 9 years |
experiment/brwinst-struct | preprocess and store brw_inst.i | Matt Turner | 11 years |
experiment/cfg | i965: Mark is_tex() and friends as const. | Matt Turner | 11 years |
experiment/demorgan-reverse | nir: Do opposite formulation of DeMorgan's Laws. | Matt Turner | 10 years |
experiment/fsgenerator | WIP | Matt Turner | 11 years |
experiment/glsl_ir_ssa | glsl: convert to and from SSA form in the compiler | Connor Abbott | 11 years |
experiment/opt-1-minus-x-into-swiz | Optimize 1 - vec4 into a swizzle of the vec4. | Matt Turner | 11 years |
experiment/sample_lz | i965/fs/skl+: Recognize sample_c_lz. | Matt Turner | 9 years |
experiment/single-i965-gen | i965/build: Add configure option to build i965_dri.so for a single Gen. | Matt Turner | 11 years |
experiment/ssa | fixup! i965/fs: add ssa_drop_subscripts(). | Matt Turner | 11 years |
experiment/ssa-before-fixups | fixup! i965/fs: Add ssa_rename_variables(). | Matt Turner | 11 years |
for-mark | i965: Always compile fp64 funcs when needed | Matt Turner | 6 years |
for-rafael | intel/compiler: Improve fix_3src_operand() | Matt Turner | 6 years |
fp64 | intel/compiler: Lower more 64-bit MOV/SEL operations | Matt Turner | 6 years |
jenkins | fixup! conversion instructions | Ian Romanick | 6 years |
lvn | fixup! i965/fs: Add local value numbering optimization pass. | Matt Turner | 10 years |
master | i965: Don't allocate uniform space for samplers | Matt Turner | 8 years |
musl | meson: Search for execinfo.h | Jory Pratt | 6 years |
sched | i965/sched: switch to bottom-up scheduling | Connor Abbott | 9 years |
sched-eot | eot | Matt Turner | 9 years |
single-i965-gen | i965/build: Add configure option to build i965_dri.so for a single Gen. | Matt Turner | 11 years |
tag/oct-18-2013 | i965: track number of basic blocks and cfg calculations | Matt Turner | 11 years |
tag/oct-2-2014 | util: s/INLINE/inline/ in u_atomic.h. | Matt Turner | 10 years |
tag/oct-2-2014-2 | i965: Give compile stats through KHR_debug. | Matt Turner | 10 years |
vec4-lower-int-mul | i965/vec4: Unrestrict constant propagation into integer multiply. | Matt Turner | 10 years |
vf-component-packing | WIP for location_frac | Matt Turner | 8 years |
wip/arb_shader_group_vote | 1 | Matt Turner | 8 years |
wip/bool | i965: Use the nir_move_comparisons pass. | Kenneth Graunke | 9 years |
wip/const-vf | match imm | Matt Turner | 10 years |
wip/const-vf2 | match imm | Matt Turner | 9 years |
wip/dot | dp2 -> add | Matt Turner | 9 years |
wip/gen4-optimize-int-conv | i965/fs: Optimize conversions to int. | Matt Turner | 10 years |
wip/lround | xlib: Use float type in calculation. | Matt Turner | 9 years |
wip/nir-gvn | i965/cfg: Ignore non-CF instructions in unreachable blocks. | Matt Turner | 9 years |
wip/nir-optimization-progress | i965: Replace OPT_V() with OPT(). | Matt Turner | 8 years |
wip/packing | glsl: Optimize packUnorm4x8. | Matt Turner | 10 years |
wip/pending-scheduler | i965: Don't check whether brw_saturate_imm() changed the value. | Matt Turner | 8 years |
wip/predicate-block | don't predicate min/max | Matt Turner | 9 years |
wip/sample_lz | i965/fs/skl+: Recognize sample_c_lz. | Matt Turner | 9 years |
wip/sched-eot | eot | Matt Turner | 9 years |
wip/unorm | mesa: Use same trick for *CLAMPED_FLOAT_TO_USHORT. | Matt Turner | 10 years |
wip/validate | i965: Split gen_device_info_i965 between brw and cl. | Matt Turner | 8 years |
wip/vec4-builder | i965/vec4: Make is_high_sampler a static function. | Matt Turner | 9 years |
wip/vec4-lower-int-mul | i965/vec4: Unrestrict constant propagation into integer multiply. | Matt Turner | 8 years |
|