| 2020-08-31 |
Dave Airlie | gallivm/nir: allow 64-bit arit ops |
blob | commitdiff | raw |
| 2019-10-17 |
Roland Scheidegger | gallivm: Fix saturated signed psub/padd intrinsics... |
blob | commitdiff | raw | diff to current |
| 2019-09-11 |
Adam Jackson | gallium: Require LLVM >= 3.9 |
blob | commitdiff | raw | diff to current |
| 2019-09-11 |
Adam Jackson | gallium: Require LLVM >= 3.6 |
blob | commitdiff | raw | diff to current |
| 2019-09-11 |
Adam Jackson | gallium: Require LLVM >= 3.4 |
blob | commitdiff | raw | diff to current |
| 2019-09-06 |
Eric Engestrom | gallivm: drop LLVM<3.3 code paths as no build system... |
blob | commitdiff | raw | diff to current |
| 2019-09-06 |
Eric Engestrom | gallivm: replace more complex 3.x version check with... |
blob | commitdiff | raw | diff to current |
| 2019-09-06 |
Eric Engestrom | gallivm: replace major llvm version checks with LLVM_VE... |
blob | commitdiff | raw | diff to current |
| 2019-08-29 |
Roland Scheidegger | gallivm: use fallback code for mul_hi with llvm >= 7.0 |
blob | commitdiff | raw | diff to current |
| 2019-06-28 |
Jose Fonseca | gallivm: Improve lp_build_rcp_refine. |
blob | commitdiff | raw | diff to current |
| 2019-04-17 |
Roland Scheidegger | gallivm: fix saturated signed add / sub with llvm 9 |
blob | commitdiff | raw | diff to current |
| 2019-01-24 |
Matt Turner | gallivm: Return true from arch_rounding_available(... |
blob | commitdiff | raw | diff to current |
| 2018-11-28 |
Matt Turner | gallivm: Use nextafterf(0.5, 0.0) as rounding constant |
blob | commitdiff | raw | diff to current |
| 2018-08-24 |
Roland Scheidegger | gallivm: don't use saturated unsigned add/sub intrinsic... |
blob | commitdiff | raw | diff to current |
| 2018-03-29 |
Ian Romanick | util: Move util_is_power_of_two to bitscan.h and rename... |
blob | commitdiff | raw | diff to current |
| 2017-11-21 |
Roland Scheidegger | llvmpipe: fix snorm blending |
blob | commitdiff | raw | diff to current |
| 2017-11-02 |
Tim Rowley | gallivm: allow arch rounding with avx512 |
blob | commitdiff | raw | diff to current |
| 2017-10-06 |
Roland Scheidegger | gallivm: don't use pabs intrinsic with llvm version... |
blob | commitdiff | raw | diff to current |
| 2017-01-24 |
Roland Scheidegger | gallivm: don't try to use fast rcp for fdiv |
blob | commitdiff | raw | diff to current |
| 2016-11-10 |
Nicolai Hähnle | gallivm: fix [IU]MUL_HI regression harder |
blob | commitdiff | raw | diff to current |
| 2016-11-08 |
Nicolai Hähnle | gallivm: fix [IU]MUL_HI regression |
blob | commitdiff | raw | diff to current |
| 2016-11-08 |
Roland Scheidegger | gallivm: introduce 32x32->64bit lp_build_mul_32_lohi... |
blob | commitdiff | raw | diff to current |
| 2016-10-18 |
Roland Scheidegger | gallivm: Use native packs and unpacks for the lerps |
blob | commitdiff | raw | diff to current |
| 2016-10-04 |
José Fonseca | gallivm: Basic AVX2 support. |
blob | commitdiff | raw | diff to current |
| 2016-06-20 |
Roland Scheidegger | gallivm: don't use integer min/max sse intrinsics with... |
blob | commitdiff | raw | diff to current |
| 2016-06-10 |
Jose Fonseca | gallivm: Never emit llvm.fmuladd on LLVM 3.3. |
blob | commitdiff | raw | diff to current |
| 2016-06-10 |
Jose Fonseca | gallivm: Use llvm.fmuladd.*. |
blob | commitdiff | raw | diff to current |
| 2016-05-27 |
Roland Scheidegger | gallivm: eliminate a unnecessary AND with unorm lerps |
blob | commitdiff | raw | diff to current |
| 2016-04-27 |
Brian Paul | gallivm: s/Elements/ARRAY_SIZE/ |
blob | commitdiff | raw | diff to current |
| 2016-04-26 |
Roland Scheidegger | gallivm: make sampling more robust against bogus coordi... |
blob | commitdiff | raw | diff to current |
| 2016-04-14 |
Jason Ekstrand | Merge remote-tracking branch 'public/master' into vulkan |
blob | commitdiff | raw | diff to current |
| 2016-04-13 |
Jose Fonseca | gallivm: Workaround LLVM PR 27332. |
blob | commitdiff | raw | diff to current |
| 2016-04-13 |
Roland Scheidegger | gallivm: use llvm.nearbyint instead of llvm.round. |
blob | commitdiff | raw | diff to current |
| 2016-04-04 |
Jason Ekstrand | Merge remote-tracking branch 'public/master' into vulkan |
blob | commitdiff | raw | diff to current |
| 2016-04-03 |
Jose Fonseca | gallivm: Introduce lp_format_intrinsic. |
blob | commitdiff | raw | diff to current |
| 2016-04-03 |
Jose Fonseca | gallivm: Use llvm.fabs. |
blob | commitdiff | raw | diff to current |
| 2016-04-03 |
Jose Fonseca | gallivm: Prefer backend agnostic intrinsic for rounding. |
blob | commitdiff | raw | diff to current |
| 2015-12-11 |
Jason Ekstrand | Merge remote-tracking branch 'mesa-public/master' into... |
blob | commitdiff | raw | diff to current |
| 2015-12-06 |
Edward O'Callaghan | gallium/auxiliary: Trivial code style cleanup |
blob | commitdiff | raw | diff to current |
| 2015-08-17 |
Jason Ekstrand | Merge remote-tracking branch 'mesa-public/master' into... |
blob | commitdiff | raw | diff to current |
| 2015-07-31 |
Marek Olšák | gallivm: add LLVMAttribute parameter to lp_build_intrinsic |
blob | commitdiff | raw | diff to current |
| 2015-07-21 |
Ilia Mirkin | gallium: replace INLINE with inline |
blob | commitdiff | raw | diff to current |
| 2015-07-01 |
Dave Airlie | gallivm: add fp64 support. (v2.1) |
blob | commitdiff | raw | diff to current |
| 2015-03-22 |
Jose Fonseca | gallivm: Silence unused variable warnings on release... |
blob | commitdiff | raw | diff to current |
| 2015-03-19 |
Brian Paul | gallivm: remove unused 'builder' variable |
blob | commitdiff | raw | diff to current |
| 2015-03-18 |
Jose Fonseca | gallivm: Use INFINITY directly. |
blob | commitdiff | raw | diff to current |
| 2014-11-26 |
José Fonseca | draw,gallivm,llvmpipe: Avoid implicit casts of 32-bit... |
blob | commitdiff | raw | diff to current |
| 2014-09-16 |
Richard Sandiford | gallivm: Fix uses of 2^24 |
blob | commitdiff | raw | diff to current |
| 2014-09-16 |
Richard Sandiford | gallivm: Add SNORM clamping to lp_build_{add, sub} |
blob | commitdiff | raw | diff to current |
| 2014-09-08 |
Roland Scheidegger | gallivm: (trivial) don't try to use rcp when the divisi... |
blob | commitdiff | raw | diff to current |
| 2014-09-08 |
Roland Scheidegger | gallivm: (trivial) fix min / max variable names |
blob | commitdiff | raw | diff to current |
| 2014-08-29 |
Roland Scheidegger | gallivm: fix somewhat broken NaN behavior for exp2 |
blob | commitdiff | raw | diff to current |
| 2014-05-14 |
José Fonseca | gallivm,draw,llvmpipe: Remove support for versions... |
blob | commitdiff | raw | diff to current |
| 2013-12-14 |
Roland Scheidegger | gallivm: fix pointer type for stmxcsr/ldmxcsr |
blob | commitdiff | raw | diff to current |
| 2013-12-10 |
Zack Rusin | llvmpipe: fix blending with half-float formats |
blob | commitdiff | raw | diff to current |
| 2013-11-14 |
Roland Scheidegger | gallivm,llvmpipe: fix float->srgb conversion to handle... |
blob | commitdiff | raw | diff to current |
| 2013-09-12 |
José Fonseca | llvmpipe: Remove the special path for TGSI_OPCODE_EXP. |
blob | commitdiff | raw | diff to current |
| 2013-08-21 |
Roland Scheidegger | gallivm: unify sin and cos implementation |
blob | commitdiff | raw | diff to current |
| 2013-08-15 |
Zack Rusin | draw: handle nan clipdistance |
blob | commitdiff | raw | diff to current |
| 2013-08-15 |
Roland Scheidegger | gallivm: revert accidentally commited hunk |
blob | commitdiff | raw | diff to current |
| 2013-08-15 |
Roland Scheidegger | gallivm: do per-sample depth comparison instead of... |
blob | commitdiff | raw | diff to current |
| 2013-07-19 |
Zack Rusin | gallivm: add a log function that handles edge cases |
blob | commitdiff | raw | diff to current |
| 2013-07-19 |
Zack Rusin | gallivm: handle -inf, inf and nan's in sin/cos instructions |
blob | commitdiff | raw | diff to current |
| 2013-07-19 |
Zack Rusin | gallivm: add a version of log2 which handles edge cases |
blob | commitdiff | raw | diff to current |
| 2013-07-19 |
Zack Rusin | gallivm: fix edge cases in exp2 |
blob | commitdiff | raw | diff to current |
| 2013-07-19 |
Zack Rusin | gallivm: handle nan's in min/max |
blob | commitdiff | raw | diff to current |
| 2013-07-17 |
Roland Scheidegger | gallivm: (trivial) simplify lp_build_cos/lp_build_sin... |
blob | commitdiff | raw | diff to current |
| 2013-07-13 |
Roland Scheidegger | gallivm: handle srgb-to-linear and linear-to-srgb conve... |
blob | commitdiff | raw | diff to current |
| 2013-07-13 |
Roland Scheidegger | gallivm: better support for fast rsqrt |
blob | commitdiff | raw | diff to current |
| 2013-07-10 |
Zack Rusin | util: treat denorm'ed floats like zero |
blob | commitdiff | raw | diff to current |
| 2013-06-10 |
Richard Sandiford | llvmpipe: Use saturating add/sub for UNORM formats |
blob | commitdiff | raw | diff to current |
| 2013-05-17 |
José Fonseca | gallivm: Eliminate 8.8 fixed point intermediates from... |
blob | commitdiff | raw | diff to current |
| 2013-05-17 |
José Fonseca | gallivm: Add and use lp_build_lerp_3d. |
blob | commitdiff | raw | diff to current |
| 2013-05-02 |
Adam Jackson | gallivm: Fix altivec intrinsics for 8xi16 add/sub |
blob | commitdiff | raw | diff to current |
| 2013-02-16 |
Roland Scheidegger | gallivm: fix issues with trunc/round/floor/ceil with... |
blob | commitdiff | raw | diff to current |
| 2013-02-09 |
Roland Scheidegger | gallivm: fix typo in lp_build_mul_norm |
blob | commitdiff | raw | diff to current |
| 2012-12-06 |
José Fonseca | gallivm: Fix lerping of (un)signed normalized numbers. |
blob | commitdiff | raw | diff to current |
| 2012-12-04 |
José Fonseca | gallivm: Re-add the kludge for lp_build_lerp of fixed... |
blob | commitdiff | raw | diff to current |
| 2012-12-04 |
José Fonseca | gallivm: Generalize lp_build_mul and lp_build_lerp... |
blob | commitdiff | raw | diff to current |
| 2012-11-29 |
Adhemerval Zanella | gallivm: Altivec floating-point rounding |
blob | commitdiff | raw | diff to current |
| 2012-11-29 |
Adhemerval Zanella | gallivm: Altivec vector add/sub intrisics |
blob | commitdiff | raw | diff to current |
| 2012-11-29 |
Adhemerval Zanella | gallivm: Altivec vector max/min intrisics |
blob | commitdiff | raw | diff to current |
| 2012-11-28 |
James Benton | gallivm: Add support for unorm16 in lp_build_mul. |
blob | commitdiff | raw | diff to current |
| 2012-10-17 |
José Fonseca | Revert "gallivm: Don't use llvm.x86.avx.max/min.ps... |
blob | commitdiff | raw | diff to current |
| 2012-10-12 |
José Fonseca | gallivm: Don't use llvm.x86.avx.max/min.ps.256 inadvert... |
blob | commitdiff | raw | diff to current |
| 2012-10-12 |
Roland Scheidegger | gallivm: fix rsqrt failures |
blob | commitdiff | raw | diff to current |
| 2012-07-17 |
Brian Paul | gallivm: silence uninitialized variable warnings |
blob | commitdiff | raw | diff to current |
| 2012-07-17 |
José Fonseca | gallivm,draw,llvmpipe: Support wider native registers. |
blob | commitdiff | raw | diff to current |
| 2012-05-21 |
James Benton | gallivm: Fixed erroneous optimisation in lp_build_min... |
blob | commitdiff | raw | diff to current |
| 2012-05-15 |
José Fonseca | gallivm: Fix lp_build_sgn for normalized/fixed-point... |
blob | commitdiff | raw | diff to current |
| 2012-05-11 |
Marek Olšák | Merge branch 'gallium-userbuf' |
blob | commitdiff | raw | diff to current |
| 2012-05-09 |
Dave Airlie | gallivm: implement iabs/issg opcode. |
blob | commitdiff | raw | diff to current |
| 2012-04-05 |
James Benton | gallivm: Updated lp_build_log2_approx to use a more... |
blob | commitdiff | raw | diff to current |
| 2012-04-05 |
James Benton | gallivm: Updated lp_build_polynomial to compute odd... |
blob | commitdiff | raw | diff to current |
| 2012-03-05 |
Roland Scheidegger | gallivm: fix floating type in lp_build_mod helper |
blob | commitdiff | raw | diff to current |
| 2012-02-28 |
Dave Airlie | gallivm: only do rcp/mul for floating |
blob | commitdiff | raw | diff to current |
| 2012-02-28 |
Dave Airlie | gallivm: add frem support to the lp_build_mod helper. |
blob | commitdiff | raw | diff to current |
| 2012-02-28 |
Dave Airlie | gallivm: add integer and unsigned mod arit functions... |
blob | commitdiff | raw | diff to current |
| 2012-02-20 |
José Fonseca | llvmpipe: Don't assume vector is 4 wide in lp_build_sin... |
blob | commitdiff | raw | diff to current |
| 2011-10-16 |
José Fonseca | llvmpipe: Use lp_build_ifloor_fract for exp2 calculation. |
blob | commitdiff | raw | diff to current |
| next |
projects / mesa.git / history