aboutsummaryrefslogtreecommitdiffstats
path: root/src/compiler/nir/nir_search_helpers.h
Commit message (Collapse)AuthorAgeFilesLines
* nir/algebraic: add patterns for a >> #b << #bRhys Perry2020-01-291-0/+26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fixes compilation of a Battlefront 2 shader with ACO by removing VGPR spilling. The reassociation makes it worse on LLVM though. pipeline-db (ACO): Totals from affected shaders: SGPRS: 10704 -> 10688 (-0.15 %) VGPRS: 18736 -> 18528 (-1.11 %) Spilled SGPRs: 70 -> 70 (0.00 %) Spilled VGPRs: 0 -> 0 (0.00 %) Private memory VGPRs: 0 -> 0 (0.00 %) Scratch size: 0 -> 0 (0.00 %) dwords per thread Code Size: 909696 -> 885796 (-2.63 %) bytes LDS: 225 -> 225 (0.00 %) blocks Max Waves: 1115 -> 1129 (1.26 %) pipeline-db (LLVM): Totals from affected shaders: SGPRS: 8472 -> 8424 (-0.57 %) VGPRS: 14284 -> 14368 (0.59 %) Spilled SGPRs: 0 -> 0 (0.00 %) Spilled VGPRs: 442 -> 503 (13.80 %) Private memory VGPRs: 0 -> 0 (0.00 %) Scratch size: 268 -> 396 (47.76 %) dwords per thread Code Size: 862568 -> 853028 (-1.11 %) bytes LDS: 0 -> 0 (0.00 %) blocks Max Waves: 971 -> 964 (-0.72 %) Signed-off-by: Rhys Perry <[email protected]> Reviewed-by: Connor Abbott <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2271>
* util: rename list_empty() to list_is_empty()Timothy Arceri2019-10-281-4/+4
| | | | | | | This makes it clear that it's a boolean test and not an action (eg. "empty the list"). Reviewed-by: Eric Engestrom <[email protected]>
* nir/search: fix the PoT helpersRob Clark2019-10-181-3/+6
| | | | | | | | | | Otherwise, if the base type is (for example) uint32, we would incorrectly think that PoT optimizations could not apply. Signed-off-by: Rob Clark <[email protected]> Reviewed-by: Jason Ekstsrand <[email protected]> Reviewed-by: Kristian H. Kristensen <[email protected]> Reviewed-by: Eduardo Lima Mitev <[email protected]>
* nir/search: Fix possible NULL dereference in is_fsignIan Romanick2019-10-171-1/+1
| | | | | Reviewed-by: Caio Marcelo de Oliveira Filho <[email protected]> Fixes: 09705747d72 ("nir/algebraic: Reassociate fadd into fmul in DPH-like pattern")
* nir: Keep the range analysis HT around intra-pass until we make a change.Eric Anholt2019-10-041-22/+30
| | | | | | | | | This lets us memoize range analysis work across instructions. Reduces runtime of shader-db on Intel by -30.0288% +/- 2.1693% (n=3). Fixes: 405de7ccb6cb ("nir/range-analysis: Rudimentary value range analysis pass") Reviewed-by: Ian Romanick <[email protected]> Reviewed-by: Connor Abbott <[email protected]>
* nir/range-analysis: Rudimentary value range analysis passIan Romanick2019-08-051-0/+49
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Most integer operations are omitted because dealing with integer overflow is hard. There are a few things that could be smarter if there was a small amount more tracking of ranges of integer types (i.e., operands are Boolean, operand values fit in 16 bits, etc.). The changes to nir_search_helpers.h are included in this patch to simplify reordering the changes to nir_opt_algebraic.py. v2: Memoize range analysis results. Without this, some shaders appear to get stuck in infinite loops. v3: Rebase on many months of Mesa changes, including 1-bit Boolean changes. v4: Rebase on "nir: Drop imov/fmov in favor of one mov instruction". v5: Use nir_alu_srcs_equal for detecting (a*a). Previously just the SSA value was compared, and this incorrectly matched (a.x*a.y). v6: Many code improvements including (but not limited to) better names, more comments, and better use of helper functions. All suggested by Caio. Rework the handling of several opcodes to use a table for mapping source ranges to a result range. This change fixed a bug that caused fmax(gt_zero, ge_zero) to be incorrectly recognized as ge_zero. Slightly tighten the range of fmul by recognizing that x*x is gt_zero if x is gt_zero. Add similar handling for -x*x. v7: Use _______ in the tables as an alias for unknown. Suggested by Caio. Reviewed-by: Caio Marcelo de Oliveira Filho <[email protected]>
* nir/algebraic: Reassociate fadd into fmul in DPH-like patternIan Romanick2019-07-111-0/+24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Moving the add to the other end of the sequence allows it to be fused into an FMA. Ice Lake total instructions in shared programs: 17173074 -> 16933147 (-1.40%) instructions in affected programs: 7938745 -> 7698818 (-3.02%) helped: 35583 HURT: 90 helped stats (abs) min: 1 max: 716 x̄: 6.75 x̃: 6 helped stats (rel) min: 0.10% max: 53.04% x̄: 5.29% x̃: 3.45% HURT stats (abs) min: 1 max: 41 x̄: 2.46 x̃: 1 HURT stats (rel) min: 0.32% max: 8.33% x̄: 1.41% x̃: 0.77% 95% mean confidence interval for instructions value: -6.80 -6.65 95% mean confidence interval for instructions %-change: -5.32% -5.22% Instructions are helped. total cycles in shared programs: 360881386 -> 359533568 (-0.37%) cycles in affected programs: 189489144 -> 188141326 (-0.71%) helped: 27250 HURT: 6707 helped stats (abs) min: 1 max: 21997 x̄: 62.15 x̃: 16 helped stats (rel) min: <.01% max: 70.69% x̄: 4.04% x̃: 2.35% HURT stats (abs) min: 1 max: 3507 x̄: 51.56 x̃: 14 HURT stats (rel) min: <.01% max: 77.26% x̄: 2.72% x̃: 1.27% 95% mean confidence interval for cycles value: -44.70 -34.68 95% mean confidence interval for cycles %-change: -2.75% -2.65% Cycles are helped. total spills in shared programs: 8943 -> 8829 (-1.27%) spills in affected programs: 625 -> 511 (-18.24%) helped: 6 HURT: 3 total fills in shared programs: 21815 -> 21719 (-0.44%) fills in affected programs: 1653 -> 1557 (-5.81%) helped: 7 HURT: 10 LOST: 11 GAINED: 3 Skylake and Broadwell had similar results. (Skylake shown) total instructions in shared programs: 15271996 -> 15040882 (-1.51%) instructions in affected programs: 7193699 -> 6962585 (-3.21%) helped: 33985 HURT: 30 helped stats (abs) min: 1 max: 260 x̄: 6.80 x̃: 6 helped stats (rel) min: 0.10% max: 30.00% x̄: 5.54% x̃: 3.85% HURT stats (abs) min: 1 max: 41 x̄: 4.00 x̃: 3 HURT stats (rel) min: 0.20% max: 2.16% x̄: 1.46% x̃: 1.72% 95% mean confidence interval for instructions value: -6.87 -6.72 95% mean confidence interval for instructions %-change: -5.59% -5.48% Instructions are helped. total cycles in shared programs: 355520785 -> 354253799 (-0.36%) cycles in affected programs: 185869148 -> 184602162 (-0.68%) helped: 25824 HURT: 6287 helped stats (abs) min: 1 max: 21997 x̄: 61.66 x̃: 16 helped stats (rel) min: <.01% max: 42.05% x̄: 4.18% x̃: 2.41% HURT stats (abs) min: 1 max: 3327 x̄: 51.76 x̃: 14 HURT stats (rel) min: <.01% max: 101.62% x̄: 2.80% x̃: 1.28% 95% mean confidence interval for cycles value: -44.70 -34.21 95% mean confidence interval for cycles %-change: -2.87% -2.76% Cycles are helped. total spills in shared programs: 8835 -> 8818 (-0.19%) spills in affected programs: 613 -> 596 (-2.77%) helped: 5 HURT: 2 total fills in shared programs: 21738 -> 21744 (0.03%) fills in affected programs: 1348 -> 1354 (0.45%) helped: 5 HURT: 11 LOST: 0 GAINED: 12 Haswell total instructions in shared programs: 13447102 -> 13381508 (-0.49%) instructions in affected programs: 3770735 -> 3705141 (-1.74%) helped: 11999 HURT: 29 helped stats (abs) min: 1 max: 409 x̄: 5.60 x̃: 3 helped stats (rel) min: 0.10% max: 20.00% x̄: 2.38% x̃: 1.87% HURT stats (abs) min: 3 max: 750 x̄: 54.90 x̃: 3 HURT stats (rel) min: 0.12% max: 125.30% x̄: 9.96% x̃: 1.82% 95% mean confidence interval for instructions value: -5.71 -5.19 95% mean confidence interval for instructions %-change: -2.39% -2.30% Instructions are helped. total cycles in shared programs: 376342236 -> 375690458 (-0.17%) cycles in affected programs: 155699021 -> 155047243 (-0.42%) helped: 8397 HURT: 2876 helped stats (abs) min: 1 max: 20248 x̄: 109.87 x̃: 18 helped stats (rel) min: <.01% max: 40.71% x̄: 2.23% x̃: 1.49% HURT stats (abs) min: 1 max: 15414 x̄: 94.15 x̃: 22 HURT stats (rel) min: <.01% max: 432.49% x̄: 3.15% x̃: 1.41% 95% mean confidence interval for cycles value: -67.64 -48.00 95% mean confidence interval for cycles %-change: -0.99% -0.74% Cycles are helped. total spills in shared programs: 23134 -> 23184 (0.22%) spills in affected programs: 1675 -> 1725 (2.99%) helped: 13 HURT: 11 total fills in shared programs: 34550 -> 34686 (0.39%) fills in affected programs: 1421 -> 1557 (9.57%) helped: 13 HURT: 11 LOST: 0 GAINED: 11 Ivy Bridge total instructions in shared programs: 12019642 -> 11987285 (-0.27%) instructions in affected programs: 1532236 -> 1499879 (-2.11%) helped: 5522 HURT: 110 helped stats (abs) min: 1 max: 312 x̄: 6.22 x̃: 3 helped stats (rel) min: 0.16% max: 20.00% x̄: 2.46% x̃: 1.88% HURT stats (abs) min: 1 max: 750 x̄: 18.07 x̃: 3 HURT stats (rel) min: 0.09% max: 125.30% x̄: 3.42% x̃: 1.15% 95% mean confidence interval for instructions value: -6.25 -5.24 95% mean confidence interval for instructions %-change: -2.43% -2.26% Instructions are helped. total cycles in shared programs: 180214667 -> 179761900 (-0.25%) cycles in affected programs: 31448723 -> 30995956 (-1.44%) helped: 7191 HURT: 2838 helped stats (abs) min: 1 max: 17680 x̄: 88.47 x̃: 17 helped stats (rel) min: <.01% max: 50.45% x̄: 2.16% x̃: 1.40% HURT stats (abs) min: 1 max: 15540 x̄: 64.63 x̃: 24 HURT stats (rel) min: 0.02% max: 435.17% x̄: 3.10% x̃: 1.51% 95% mean confidence interval for cycles value: -53.34 -36.95 95% mean confidence interval for cycles %-change: -0.81% -0.53% Cycles are helped. total spills in shared programs: 3599 -> 3642 (1.19%) spills in affected programs: 1180 -> 1223 (3.64%) helped: 12 HURT: 2 total fills in shared programs: 4031 -> 4162 (3.25%) fills in affected programs: 876 -> 1007 (14.95%) helped: 12 HURT: 2 LOST: 6 GAINED: 5 Sandy Bridge total instructions in shared programs: 10850686 -> 10822890 (-0.26%) instructions in affected programs: 1247986 -> 1220190 (-2.23%) helped: 4699 HURT: 102 helped stats (abs) min: 1 max: 104 x̄: 6.02 x̃: 3 helped stats (rel) min: 0.15% max: 17.65% x̄: 2.44% x̃: 1.88% HURT stats (abs) min: 1 max: 16 x̄: 4.70 x̃: 3 HURT stats (rel) min: 0.09% max: 3.85% x̄: 1.11% x̃: 1.10% 95% mean confidence interval for instructions value: -6.10 -5.47 95% mean confidence interval for instructions %-change: -2.42% -2.30% Instructions are helped. total cycles in shared programs: 154044149 -> 153920095 (-0.08%) cycles in affected programs: 26037392 -> 25913338 (-0.48%) helped: 5974 HURT: 2521 helped stats (abs) min: 1 max: 1802 x̄: 35.42 x̃: 16 helped stats (rel) min: <.01% max: 35.80% x̄: 1.43% x̃: 0.84% HURT stats (abs) min: 1 max: 862 x̄: 34.73 x̃: 20 HURT stats (rel) min: 0.01% max: 36.33% x̄: 1.67% x̃: 0.85% 95% mean confidence interval for cycles value: -16.31 -12.90 95% mean confidence interval for cycles %-change: -0.56% -0.45% Cycles are helped. total spills in shared programs: 2876 -> 2957 (2.82%) spills in affected programs: 592 -> 673 (13.68%) helped: 6 HURT: 35 total fills in shared programs: 3157 -> 3134 (-0.73%) fills in affected programs: 402 -> 379 (-5.72%) helped: 6 HURT: 0 LOST: 5 GAINED: 11 Reviewed-by: Matt Turner <[email protected]>
* nir/algebraic: Add helpers and a rule involving wrappingCaio Marcelo de Oliveira Filho2019-06-261-0/+12
| | | | | | | | The helpers are needed so we can use the syntax `instr(cond)` in the algebraic rules. Add simple rule for dropping a pair of mul-div of the same value when wrapping is guaranteed to not happen. Reviewed-by: Jason Ekstrand <[email protected]>
* nir_algebraic: Add basic optimizations for umul_low and imadsh_mix16Eduardo Lima Mitev2019-06-071-0/+46
| | | | | | | | | | | | | | | | | | For umul_low (al * bl), zero is returned if the low 16-bits word of either source is zero. for imadsh_mix16 (ah * bl << 16 + c), c is returned if either 'ah' or 'bl' is zero. A couple of nir_search_helpers are added: is_upper_half_zero() returns true if the highest word of all components of an integer NIR alu src are zero. is_lower_half_zero() returns true if the lowest word of all components of an integer nir alu src are zero. Reviewed-by: Eric Anholt <[email protected]>
* nir/algebraic: Commute 1-fsat(a) to fsat(1-a) for all non-fmul instructionsIan Romanick2019-05-141-0/+16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The goal is to avoid having an extra MOV instruction to perform the saturate. Doing the subtraction first allows the saturate to be applied to the ADD instruction making the MOV unnecessary. Values generated in different block and values from non-ALU instructions (e.g., texture instructions) almost always need the extra MOV. Multiply instructions are restricted because doing this rearrangement can interfere with the generation of flrp and ffma instructions. v2: Now that the final method has been selected, squash three commits into one. All Intel platforms has similar results. (Ice Lake shown) total instructions in shared programs: 17223214 -> 17219386 (-0.02%) instructions in affected programs: 1524376 -> 1520548 (-0.25%) helped: 2686 HURT: 26 helped stats (abs) min: 1 max: 32 x̄: 1.44 x̃: 1 helped stats (rel) min: 0.03% max: 16.67% x̄: 0.54% x̃: 0.37% HURT stats (abs) min: 1 max: 2 x̄: 1.69 x̃: 2 HURT stats (rel) min: 0.33% max: 1.67% x̄: 0.54% x̃: 0.35% 95% mean confidence interval for instructions value: -1.46 -1.36 95% mean confidence interval for instructions %-change: -0.56% -0.50% Instructions are helped. total cycles in shared programs: 360811571 -> 360791896 (<.01%) cycles in affected programs: 103650214 -> 103630539 (-0.02%) helped: 1557 HURT: 675 helped stats (abs) min: 1 max: 1773 x̄: 41.44 x̃: 16 helped stats (rel) min: <.01% max: 26.77% x̄: 1.37% x̃: 0.64% HURT stats (abs) min: 1 max: 1513 x̄: 66.44 x̃: 14 HURT stats (rel) min: <.01% max: 46.16% x̄: 2.00% x̃: 0.49% 95% mean confidence interval for cycles value: -14.82 -2.81 95% mean confidence interval for cycles %-change: -0.50% -0.20% Cycles are helped. LOST: 2 GAINED: 0 Reviewed-by: Matt Turner <[email protected]> [v1] Reviewed-by: Thomas Helland <[email protected]>
* nir/algebraic: Eliminate useless fsat() on operand of comparison w/value in ↵Ian Romanick2019-05-141-0/+30
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | (0, 1) v2: Fix copy-and-paste bug in a cmp b vs b cmp a cases. All Gen7+ platforms had similar results. (Ice Lake shown) total instructions in shared programs: 17224337 -> 17224269 (<.01%) instructions in affected programs: 13578 -> 13510 (-0.50%) helped: 68 HURT: 0 helped stats (abs) min: 1 max: 1 x̄: 1.00 x̃: 1 helped stats (rel) min: 0.31% max: 3.12% x̄: 0.84% x̃: 0.42% 95% mean confidence interval for instructions value: -1.00 -1.00 95% mean confidence interval for instructions %-change: -1.05% -0.63% Instructions are helped. total cycles in shared programs: 360826090 -> 360825137 (<.01%) cycles in affected programs: 94867 -> 93914 (-1.00%) helped: 58 HURT: 1 helped stats (abs) min: 2 max: 28 x̄: 17.74 x̃: 18 helped stats (rel) min: 0.08% max: 3.17% x̄: 1.39% x̃: 1.22% HURT stats (abs) min: 76 max: 76 x̄: 76.00 x̃: 76 HURT stats (rel) min: 2.86% max: 2.86% x̄: 2.86% x̃: 2.86% 95% mean confidence interval for cycles value: -19.53 -12.78 95% mean confidence interval for cycles %-change: -1.56% -1.08% Cycles are helped. No changes on any other Intel platform. Reviewed-by: Matt Turner <[email protected]> Reviewed-by: Thomas Helland <[email protected]>
* Revert "nir: add late opt to turn inot/b2f combos back to bcsel"Ian Romanick2019-05-141-16/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This reverts commit 7acc8652268205a266068ea4d059eccce43e1f78. With these optimizations in place, the extra constant folding added in the next commit extends some live ranges of 0.0 and ±1.0 constants, and that causes several hundred shaders to have more spills and fills. I believe this optimization we made basically irrelevant by 7725d609387 "intel/fs: Emit better code for b2f(inot(a)) and b2i(inot(a))". All Gen7.5+ platforms had similar results. (Ice Lake shown) total instructions in shared programs: 17225303 -> 17224634 (<.01%) instructions in affected programs: 879402 -> 878733 (-0.08%) helped: 679 HURT: 1 helped stats (abs) min: 1 max: 1 x̄: 1.00 x̃: 1 helped stats (rel) min: 0.03% max: 0.93% x̄: 0.24% x̃: 0.05% HURT stats (abs) min: 10 max: 10 x̄: 10.00 x̃: 10 HURT stats (rel) min: 0.45% max: 0.45% x̄: 0.45% x̃: 0.45% 95% mean confidence interval for instructions value: -1.02 -0.95 95% mean confidence interval for instructions %-change: -0.26% -0.22% Instructions are helped. total cycles in shared programs: 360842595 -> 360828542 (<.01%) cycles in affected programs: 110443594 -> 110429541 (-0.01%) helped: 389 HURT: 265 helped stats (abs) min: 1 max: 7525 x̄: 162.81 x̃: 28 helped stats (rel) min: <.01% max: 18.66% x̄: 1.11% x̃: 0.11% HURT stats (abs) min: 1 max: 7614 x̄: 185.96 x̃: 48 HURT stats (rel) min: <.01% max: 25.08% x̄: 0.95% x̃: 0.10% 95% mean confidence interval for cycles value: -75.65 32.67 95% mean confidence interval for cycles %-change: -0.49% -0.06% Inconclusive result (value mean confidence interval includes 0). total spills in shared programs: 12159 -> 12161 (0.02%) spills in affected programs: 13 -> 15 (15.38%) helped: 0 HURT: 1 total fills in shared programs: 25207 -> 25208 (<.01%) fills in affected programs: 25 -> 26 (4.00%) helped: 0 HURT: 1 Ivy Bridge total instructions in shared programs: 12082019 -> 12082013 (<.01%) instructions in affected programs: 1033 -> 1027 (-0.58%) helped: 6 HURT: 0 helped stats (abs) min: 1 max: 1 x̄: 1.00 x̃: 1 helped stats (rel) min: 0.41% max: 0.83% x̄: 0.61% x̃: 0.59% 95% mean confidence interval for instructions value: -1.00 -1.00 95% mean confidence interval for instructions %-change: -0.78% -0.45% Instructions are helped. total cycles in shared programs: 179849270 -> 179849157 (<.01%) cycles in affected programs: 4735 -> 4622 (-2.39%) helped: 4 HURT: 0 helped stats (abs) min: 2 max: 74 x̄: 28.25 x̃: 18 helped stats (rel) min: 0.13% max: 6.53% x̄: 2.85% x̃: 2.36% 95% mean confidence interval for cycles value: -82.73 26.23 95% mean confidence interval for cycles %-change: -7.98% 2.28% Inconclusive result (value mean confidence interval includes 0). Sandy Bridge total instructions in shared programs: 10882750 -> 10882748 (<.01%) instructions in affected programs: 266 -> 264 (-0.75%) helped: 2 HURT: 0 Iron Lake total cycles in shared programs: 188609440 -> 188609448 (<.01%) cycles in affected programs: 4320 -> 4328 (0.19%) helped: 0 HURT: 2 GM45 total cycles in shared programs: 129016868 -> 129016872 (<.01%) cycles in affected programs: 2302 -> 2306 (0.17%) helped: 0 HURT: 1 Reviewed-by: Matt Turner <[email protected]>
* nir/algebraic: Push unary operations into source operands of fsat sourceIan Romanick2019-05-141-0/+19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Pushing a unary operation, like fneg, into the operation that generates its operand allows the fsat to be applied to the inner instruction instead of on a separate instruction that performs the unary operation. This changes fmul ssa_100, ssa_99, ssa_98 fmov.sat ssa_101, -ssa_100 into fmul.sat ssa_100, -ssa_99, ssa_98 Ice Lake, Skylake, and Broadwell had similar results. (Ice Lake shown) total instructions in shared programs: 17228658 -> 17228584 (<.01%) instructions in affected programs: 3163 -> 3089 (-2.34%) helped: 49 HURT: 0 helped stats (abs) min: 1 max: 2 x̄: 1.51 x̃: 2 helped stats (rel) min: 0.58% max: 9.09% x̄: 3.69% x̃: 3.51% 95% mean confidence interval for instructions value: -1.66 -1.37 95% mean confidence interval for instructions %-change: -4.37% -3.00% Instructions are helped. total cycles in shared programs: 360937144 -> 360936431 (<.01%) cycles in affected programs: 24029 -> 23316 (-2.97%) helped: 47 HURT: 2 helped stats (abs) min: 4 max: 18 x̄: 15.34 x̃: 16 helped stats (rel) min: 0.69% max: 6.18% x̄: 3.78% x̃: 4.27% HURT stats (abs) min: 4 max: 4 x̄: 4.00 x̃: 4 HURT stats (rel) min: 0.34% max: 0.67% x̄: 0.50% x̃: 0.50% 95% mean confidence interval for cycles value: -16.05 -13.05 95% mean confidence interval for cycles %-change: -4.07% -3.15% Cycles are helped. All Gen7 and earlier platforms had similar results. (Haswell shown) total instructions in shared programs: 13536059 -> 13535884 (<.01%) instructions in affected programs: 8797 -> 8622 (-1.99%) helped: 150 HURT: 0 helped stats (abs) min: 1 max: 2 x̄: 1.17 x̃: 1 helped stats (rel) min: 0.40% max: 11.11% x̄: 3.51% x̃: 1.96% 95% mean confidence interval for instructions value: -1.23 -1.11 95% mean confidence interval for instructions %-change: -3.97% -3.05% Instructions are helped. total cycles in shared programs: 357696119 -> 357694193 (<.01%) cycles in affected programs: 50216 -> 48290 (-3.84%) helped: 109 HURT: 14 helped stats (abs) min: 2 max: 92 x̄: 18.97 x̃: 16 helped stats (rel) min: 0.26% max: 19.09% x̄: 7.37% x̃: 5.37% HURT stats (abs) min: 2 max: 26 x̄: 10.14 x̃: 5 HURT stats (rel) min: 0.18% max: 4.73% x̄: 1.84% x̃: 0.92% 95% mean confidence interval for cycles value: -19.27 -12.05 95% mean confidence interval for cycles %-change: -7.34% -5.31% Cycles are helped. Reviewed-by: Matt Turner <[email protected]>
* nir: Add partial redundancy elimination for comparesIan Romanick2019-03-281-0/+27
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This pass attempts to dectect code sequences like if (x < y) { z = y - x; ... } and replace them with sequences like t = x - y; if (t < 0) { z = -t; ... } On architectures where the subtract can generate the flags used by the if-statement, this saves an instruction. It's also possible that moving an instruction out of the if-statement will allow nir_opt_peephole_select to convert the whole thing to a bcsel. Currently only floating point compares and adds are supported. Adding support for integer will be a challenge due to integer overflow. There are a couple possible solutions, but they may not apply to all architectures. v2: Fix a typo in the commit message and a couple typos in comments. Fix possible NULL pointer deref from result of push_block(). Add missing (-A + B) case. Suggested by Caio. v3: Fix is_not_const_zero to work correctly with types other than nir_type_float32. Suggested by Ken. v4: Add some comments explaining how this works. Suggested by Ken. Reviewed-by: Kenneth Graunke <[email protected]>
* nir/algebraic: Replace i2b used by bcsel or if-statement with comparisonIan Romanick2019-03-011-0/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | All of the helped shaders are in Deus Ex. I looked at a couple shaders, and they have a pattern like: vec1 32 ssa_373 = i2b32 ssa_345.w vec1 32 ssa_374 = bcsel ssa_373, ssa_20, ssa_0 ... vec1 32 ssa_377 = ine ssa_345.w, ssa_0 if ssa_377 { ... vec1 32 ssa_416 = i2b32 ssa_385.w vec1 32 ssa_417 = bcsel ssa_416, ssa_386, ssa_374 ... } The massive help occurs because the i2b32 is removed, then other passes determine that ssa_374 must be ssa_20 inside the if-statement allowing the first bcsel to also be deleted. v2: Rebase on 1-bit Boolean changes. v3: Fix i2b32 vs ine problem in if-statement replacement. Noticed by Bas. Skylake total instructions in shared programs: 15241394 -> 15186287 (-0.36%) instructions in affected programs: 890583 -> 835476 (-6.19%) helped: 355 HURT: 0 helped stats (abs) min: 1 max: 497 x̄: 155.23 x̃: 149 helped stats (rel) min: 0.09% max: 16.49% x̄: 6.10% x̃: 6.59% 95% mean confidence interval for instructions value: -165.07 -145.39 95% mean confidence interval for instructions %-change: -6.42% -5.77% Instructions are helped. total cycles in shared programs: 373846583 -> 371023357 (-0.76%) cycles in affected programs: 118972102 -> 116148876 (-2.37%) helped: 343 HURT: 14 helped stats (abs) min: 45 max: 118284 x̄: 8332.32 x̃: 6089 helped stats (rel) min: 0.03% max: 38.19% x̄: 2.48% x̃: 1.77% HURT stats (abs) min: 120 max: 4126 x̄: 2482.79 x̃: 3019 HURT stats (rel) min: 0.16% max: 17.37% x̄: 2.13% x̃: 1.11% 95% mean confidence interval for cycles value: -8723.28 -7093.12 95% mean confidence interval for cycles %-change: -2.57% -2.02% Cycles are helped. total spills in shared programs: 32401 -> 23465 (-27.58%) spills in affected programs: 24457 -> 15521 (-36.54%) helped: 343 HURT: 0 total fills in shared programs: 37866 -> 31765 (-16.11%) fills in affected programs: 18889 -> 12788 (-32.30%) helped: 343 HURT: 0 Broadwell and Haswell had similar results. (Haswell shown) Haswell total instructions in shared programs: 13764783 -> 13750679 (-0.10%) instructions in affected programs: 1176256 -> 1162152 (-1.20%) helped: 334 HURT: 21 helped stats (abs) min: 1 max: 358 x̄: 42.59 x̃: 47 helped stats (rel) min: 0.09% max: 11.81% x̄: 1.30% x̃: 1.37% HURT stats (abs) min: 1 max: 61 x̄: 5.76 x̃: 1 HURT stats (rel) min: 0.03% max: 1.84% x̄: 0.17% x̃: 0.03% 95% mean confidence interval for instructions value: -43.99 -35.47 95% mean confidence interval for instructions %-change: -1.35% -1.08% Instructions are helped. total cycles in shared programs: 386511910 -> 385402528 (-0.29%) cycles in affected programs: 143831110 -> 142721728 (-0.77%) helped: 327 HURT: 39 helped stats (abs) min: 16 max: 25219 x̄: 3519.74 x̃: 3570 helped stats (rel) min: <.01% max: 10.26% x̄: 0.95% x̃: 0.96% HURT stats (abs) min: 16 max: 4881 x̄: 1065.95 x̃: 997 HURT stats (rel) min: <.01% max: 16.67% x̄: 0.70% x̃: 0.24% 95% mean confidence interval for cycles value: -3375.59 -2686.60 95% mean confidence interval for cycles %-change: -0.92% -0.64% Cycles are helped. total spills in shared programs: 100480 -> 97846 (-2.62%) spills in affected programs: 84702 -> 82068 (-3.11%) helped: 316 HURT: 21 total fills in shared programs: 96877 -> 94369 (-2.59%) fills in affected programs: 69167 -> 66659 (-3.63%) helped: 316 HURT: 9 No changes on Ivy Bridge or earlier platforms. Reviewed-by: Kenneth Graunke <[email protected]>
* nir/search_helpers: Use nir_src_is_const and friendsJason Ekstrand2018-10-221-27/+22
| | | | | | | This not only makes them safe for more bit sizes but it also fixes a bug in is_zero_to_one where it would return true for constant NaN. Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* util: Add and use util_is_power_of_two_nonzeroIan Romanick2018-03-291-9/+4
| | | | | Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Eduardo Lima Mitev <[email protected]>
* nir: Don't condition 'a-b < 0' -> 'a < b' on is_not_used_by_conditionalIan Romanick2018-03-261-15/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Now that i965 recognizes that a-b generates the same conditions as 'a < b', there is no reason to condition this transformation on 'is not used by conditional.' Since this was the only user of the is_not_used_by_conditional function, delete it. All Gen6+ platforms had similar results. (Skylake shown) total instructions in shared programs: 14400775 -> 14400595 (<.01%) instructions in affected programs: 36712 -> 36532 (-0.49%) helped: 182 HURT: 26 helped stats (abs) min: 1 max: 2 x̄: 1.13 x̃: 1 helped stats (rel) min: 0.15% max: 1.82% x̄: 0.70% x̃: 0.62% HURT stats (abs) min: 1 max: 1 x̄: 1.00 x̃: 1 HURT stats (rel) min: 0.24% max: 1.02% x̄: 0.82% x̃: 0.90% 95% mean confidence interval for instructions value: -0.97 -0.76 95% mean confidence interval for instructions %-change: -0.59% -0.43% Instructions are helped. total cycles in shared programs: 532929592 -> 532926345 (<.01%) cycles in affected programs: 478660 -> 475413 (-0.68%) helped: 187 HURT: 22 helped stats (abs) min: 2 max: 200 x̄: 20.99 x̃: 18 helped stats (rel) min: 0.23% max: 24.10% x̄: 1.48% x̃: 1.03% HURT stats (abs) min: 1 max: 214 x̄: 30.86 x̃: 11 HURT stats (rel) min: 0.01% max: 23.06% x̄: 3.12% x̃: 0.86% 95% mean confidence interval for cycles value: -19.50 -11.57 95% mean confidence interval for cycles %-change: -1.42% -0.58% Cycles are helped. GM45 and Iron Lake had similar results. (Iron Lake shown) total cycles in shared programs: 177851578 -> 177851810 (<.01%) cycles in affected programs: 24408 -> 24640 (0.95%) helped: 2 HURT: 4 helped stats (abs) min: 4 max: 4 x̄: 4.00 x̃: 4 helped stats (rel) min: 0.42% max: 0.47% x̄: 0.44% x̃: 0.44% HURT stats (abs) min: 24 max: 108 x̄: 60.00 x̃: 54 HURT stats (rel) min: 0.52% max: 1.62% x̄: 1.04% x̃: 1.02% 95% mean confidence interval for cycles value: -7.75 85.08 95% mean confidence interval for cycles %-change: -0.39% 1.49% Inconclusive result (value mean confidence interval includes 0). Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* nir: Silence unused parameter warningsIan Romanick2018-01-101-2/+2
| | | | | | | | | | | | | | | | In file included from src/compiler/nir/nir_opt_algebraic.c:4:0: src/compiler/nir/nir_search_helpers.h: In function ‘is_not_const’: src/compiler/nir/nir_search_helpers.h:118:59: warning: unused parameter ‘num_components’ [-Wunused-parameter] is_not_const(nir_alu_instr *instr, unsigned src, unsigned num_components, ^~~~~~~~~~~~~~ src/compiler/nir/nir_search_helpers.h:119:29: warning: unused parameter ‘swizzle ’ [-Wunused-parameter] const uint8_t *swizzle) ^~~~~~~ Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Alejandro Piñeiro <[email protected]>
* nir: shuffle constants to the topTimothy Arceri2017-04-241-0/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | V2: mark float opts as inexact If one of the inputs to an mul/add is the result of another mul/add there is a chance that we can reuse the result of that mul/add in other calls if we do the multiplication in the right order. Also by attempting to move all constants to the top we increase the chance of constant folding. For example it is a fairly common pattern for shaders to do something similar to this: const float a = 0.5; in vec4 b; in float c; ... b.x = b.x * c; b.y = b.y * c; ... b.x = b.x * a + a; b.y = b.y * a + a; So by simply detecting that constant a is part of the multiplication in ffma and switching it with previous fmul that updates b we end up with: ... c = a * c; ... b.x = b.x * c + a; b.y = b.y * c + a; Shader-db results BDW: total instructions in shared programs: 13011050 -> 12967888 (-0.33%) instructions in affected programs: 4118366 -> 4075204 (-1.05%) helped: 17739 HURT: 1343 total cycles in shared programs: 246717952 -> 246410716 (-0.12%) cycles in affected programs: 166870802 -> 166563566 (-0.18%) helped: 18493 HURT: 7965 total spills in shared programs: 14937 -> 14560 (-2.52%) spills in affected programs: 9331 -> 8954 (-4.04%) helped: 284 HURT: 33 total fills in shared programs: 20211 -> 19671 (-2.67%) fills in affected programs: 12586 -> 12046 (-4.29%) helped: 286 HURT: 33 LOST: 39 GAINED: 33 Some of the hurt will go away when we shuffle things back down to the bottom in the following patch. It's also noteworthy that almost all of the spill changes are in Deus Ex both hurt and helped. Reviewed-by: Elie Tournier <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]>
* nir: add flt comparision simplificationTimothy Arceri2017-04-241-0/+15
| | | | | | | | | | | | | | | | | | | | | | | | | Didn't turn out as useful as I'd hoped, but it will help alot more on i965 by reducing regressions when we drop brw_do_channel_expressions() and brw_do_vector_splitting(). I'm not sure how much sense 'is_not_used_by_conditional' makes on platforms other than i965 but since this is a new opt it at least won't do any harm. shader-db BDW: total instructions in shared programs: 13029581 -> 13029415 (-0.00%) instructions in affected programs: 15268 -> 15102 (-1.09%) helped: 86 HURT: 0 total cycles in shared programs: 247038346 -> 247036198 (-0.00%) cycles in affected programs: 692634 -> 690486 (-0.31%) helped: 183 HURT: 27 Reviewed-by: Elie Tournier <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]>
* nir: Stop using apostrophes to pluralize.Matt Turner2017-03-231-2/+2
| | | | | Reviewed-by: Dylan Baker <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* nir: optimise min/max fadd combosTimothy Arceri2017-01-141-0/+22
| | | | | | | | | | | | | | | | shader-db results BDW: total instructions in shared programs: 13060410 -> 13060313 (-0.00%) instructions in affected programs: 24533 -> 24436 (-0.40%) helped: 88 HURT: 0 total cycles in shared programs: 256585692 -> 256586698 (0.00%) cycles in affected programs: 647290 -> 648296 (0.16%) helped: 35 HURT: 30 Reviewed-by: Matt Turner <[email protected]>
* nir: don't turn ieq/ine into inot if used by an ifTimothy Arceri2017-01-121-0/+6
| | | | | | | | | | | | | | | | | | | Otherwise we will end up with an extra instruction to compare the result of the inot. On BDW: total instructions in shared programs: 13060620 -> 13060481 (-0.00%) instructions in affected programs: 103379 -> 103240 (-0.13%) helped: 127 HURT: 0 total cycles in shared programs: 256590950 -> 256587408 (-0.00%) cycles in affected programs: 11324730 -> 11321188 (-0.03%) helped: 114 HURT: 21 Reviewed-by: Jason Ekstrand <[email protected]>
* nir: add late opt to turn inot/b2f combos back to bcselTimothy Arceri2017-01-121-0/+16
| | | | | | | | | | | | | | | | | | | | | | | | | | | We turn these from bcsel into inot/b2f combos in order for other optimisation passes to get further. Once we have finished turn the ones that remain and are used in more than a single expression back into a bcsel. On BDW: total instructions in shared programs: 13060965 -> 13060297 (-0.01%) instructions in affected programs: 835701 -> 835033 (-0.08%) helped: 670 HURT: 2 total cycles in shared programs: 256599536 -> 256598006 (-0.00%) cycles in affected programs: 114655488 -> 114653958 (-0.00%) helped: 419 HURT: 240 LOST: 0 GAINED: 1 The 2 HURT is because inserting bcsel creates the only use of const 1.0 in two shaders from tri-of-friendship-and-madness. Reviewed-by: Jason Ekstrand <[email protected]>
* nir: Move fsat outside of fmin/fmax if second arg is 0 to 1.Matt Turner2016-12-121-0/+23
| | | | | | | | | | instructions in affected programs: 550 -> 544 (-1.09%) helped: 6 cycles in affected programs: 6952 -> 6850 (-1.47%) helped: 6 Reviewed-by: Jason Ekstrand <[email protected]>
* nir/algebraic: support for power-of-two optimizationsRob Clark2016-06-031-0/+94
Some optimizations, like converting integer multiply/divide into left/ right shifts, have additional constraints on the search expression. Like requiring that a variable is a constant power of two. Support these cases by allowing a fxn name to be appended to the search var expression (ie. "a#32(is_power_of_two)"). Signed-off-by: Rob Clark <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-02 07:25:50 +0000