| Commit message (Expand) | Author | Age | Files | Lines |
|---|
| * | util: Move util_is_power_of_two to bitscan.h and rename to util_is_power_of_t... | Ian Romanick | 2018-03-29 | 1 | -2/+2 |
| * | nvir: dont optimize mad with subops to shladd | Karol Herbst | 2018-02-24 | 1 | -1/+2 |
| * | nv50/ir: Fix unused var warnings in release build | Rhys Kidd | 2017-12-29 | 1 | -1/+3 |
| * | nv50/ir: move LateAlgebraicOpt to the very end | Ilia Mirkin | 2017-11-26 | 1 | -1/+1 |
| * | nv50/ir: add optimization for modulo by a non-power-of-2 value | Ilia Mirkin | 2017-11-26 | 1 | -0/+15 |
| * | nv50/ir: optimize signed integer modulo by pow-of-2 | Ilia Mirkin | 2017-11-25 | 1 | -3/+28 |
| * | nv50/ir: clean up saturated values immediately | Ilia Mirkin | 2017-08-12 | 1 | -1/+6 |
| * | nv50/ir: fix ConstantFolding with saturation | Karol Herbst | 2017-08-09 | 1 | -0/+8 |
| * | nv50/ir: disable mul+add to mad for precise instructions | Karol Herbst | 2017-07-21 | 1 | -2/+3 |
| * | nv50/ir: fix combineLd/St to update existing records as necessary | Ilia Mirkin | 2017-06-26 | 1 | -0/+8 |
| * | nv50/ir: adjust overlapping logic to take fileIndex-relative offsets | Ilia Mirkin | 2017-06-26 | 1 | -1/+5 |
| * | nv50/ir: VFETCH is also considered a load for MemoryOpt | Ilia Mirkin | 2017-06-26 | 1 | -1/+1 |
| * | nv50/ir: Properly fold constants in SPLIT operation | Pierre Moreau | 2017-06-25 | 1 | -3/+4 |
| * | nv50/ir: also do PostRaLoadPropagation for FMA | Karol Herbst | 2017-03-31 | 1 | -0/+1 |
| * | gm107/ir: add LIMM form of mad | Karol Herbst | 2017-03-31 | 1 | -2/+1 |
| * | gk110/ir: add LIMM form of mad | Karol Herbst | 2017-03-31 | 1 | -1/+1 |
| * | nv50/ir: implement mad post ra folding for nvc0+ | Karol Herbst | 2017-03-31 | 1 | -4/+47 |
| * | nv50/ir: restructure and rename postraconstantfolding pass | Karol Herbst | 2017-03-31 | 1 | -58/+63 |
| * | nvc0/ir: also do ConstantFolding for FMA | Karol Herbst | 2017-03-31 | 1 | -0/+1 |
| * | nv50/ir: check for origin insn in findOriginForTestWithZero | Pierre Moreau | 2017-03-09 | 1 | -0/+2 |
| * | nv50/ir: convert an ATOM.EXCH without a destination into a store | Ilia Mirkin | 2017-02-11 | 1 | -0/+5 |
| * | nvc0/ir: add support for all the new int64 tgsi opcodes | Ilia Mirkin | 2017-02-09 | 1 | -1/+5 |
| * | nv50/ir: Split 64-bit integer MAD/MUL operations | Pierre Moreau | 2017-02-09 | 1 | -0/+116 |
| * | nvc0: add support for MUL_ZERO_WINS property | Ilia Mirkin | 2017-01-23 | 1 | -0/+1 |
| * | nv50/ir: optimize shl + and | Ilia Mirkin | 2017-01-16 | 1 | -0/+11 |
| * | nv50/ir: start LocalCSE with getFirst to merge PHI instructions | Karol Herbst | 2016-10-25 | 1 | -1/+1 |
| * | nv50/ir: it appears that OP_DISCARD can't take a join modifier | Ilia Mirkin | 2016-10-22 | 1 | -0/+1 |
| * | nv50/ir: constant fold OP_SPLIT | Tobias Klausmann | 2016-10-14 | 1 | -0/+18 |
| * | nvc0/ir: be more careful about preserving modifiers in SHLADD creation | Ilia Mirkin | 2016-10-13 | 1 | -7/+5 |
| * | nv50/ir: optimize ADD(SHL(a, b), c) to SHLADD(a, b, c) | Samuel Pitoiset | 2016-10-12 | 1 | -0/+87 |
| * | nv50/ir: fix wrong check when optimizing MAD to SHLADD | Samuel Pitoiset | 2016-10-07 | 1 | -1/+1 |
| * | nv50/ir: optimize sub(a, 0) to a | Karol Herbst | 2016-10-06 | 1 | -0/+3 |
| * | nv50/ir: optimize SHLADD(a, b, c) to MOV((a << b) + c) | Samuel Pitoiset | 2016-09-29 | 1 | -0/+3 |
| * | nv50/ir: optimize SHLADD(a, b, 0x0) to SHL(a, b) | Samuel Pitoiset | 2016-09-29 | 1 | -0/+8 |
| * | nv50/ir: optimize IMAD to SHLADD in presence of power of 2 | Samuel Pitoiset | 2016-09-29 | 1 | -0/+7 |
| * | nv50/ir: optimize SUB(a, b) to MOV(a - b) | Samuel Pitoiset | 2016-09-18 | 1 | -0/+10 |
| * | nv50/ir: allow to swap sources for OP_SUB | Samuel Pitoiset | 2016-07-22 | 1 | -1/+6 |
| * | nv50: fix alphatest for non-blendable formats | Ilia Mirkin | 2016-07-16 | 1 | -0/+4 |
| * | nvc0/ir: handle a load's reg result not being used for locked variants | Ilia Mirkin | 2016-05-26 | 1 | -2/+8 |
| * | Treewide: Remove Elements() macro | Jan Vesely | 2016-05-17 | 1 | -1/+1 |
| * | nouveau: codegen: combineLd/St do not combine indirect loads | Hans de Goede | 2016-04-25 | 1 | -0/+7 |
| * | nv50/ir: do not try to attach JOIN ops to ATOM | Samuel Pitoiset | 2016-04-07 | 1 | -1/+1 |
| * | nv50/ir: avoid folding mul + add if the mul has a dnz | Ilia Mirkin | 2016-03-13 | 1 | -3/+2 |
| * | nv50/ir: we can't do the add to mad conversion when the mul saturates | Karol Herbst | 2016-02-16 | 1 | -0/+3 |
| * | nv50/ir: optimize neg(and(set, 1)) to set | Karol Herbst | 2016-02-16 | 1 | -0/+32 |
| * | nv50/ir: get rid of memory stores with nop values | Ilia Mirkin | 2016-01-30 | 1 | -0/+6 |
| * | nv50/ir: fix false global CSE on instructions with multiple defs | Ilia Mirkin | 2016-01-30 | 1 | -0/+2 |
| * | nv50/ir: add support for BUFFER accesses | Ilia Mirkin | 2016-01-29 | 1 | -1/+1 |
| * | nv50/ir: optimize mad/fma with third argument 0 to mul | Karol Herbst | 2016-01-28 | 1 | -0/+21 |
| * | nv50/ir: run DCE backwards | Karol Herbst | 2016-01-28 | 1 | -3/+3 |
