> On Jul 13, 2016, at 12:36, Marek Olšák <mar...@gmail.com> wrote: > > On Wed, Jul 13, 2016 at 9:25 PM, Tom Stellard <t...@stellard.net > <mailto:t...@stellard.net>> wrote: >> On Wed, Jul 13, 2016 at 03:20:55PM -0400, Tom Stellard wrote: >>> On Tue, Jul 12, 2016 at 10:52:35PM +0200, Marek Olšák wrote: >>>> From: Marek Olšák <marek.ol...@amd.com> >>>> >>>> This allows moving the loads arbitrarily in the Sinking pass. >>>> >>>> 26002 shaders in 14643 tests >>>> Totals: >>>> SGPRS: 2080160 -> 2080160 (0.00 %) >>>> VGPRS: 798875 -> 797826 (-0.13 %) >>>> Spilled SGPRs: 108485 -> 79165 (-27.03 %) >>>> Spilled VGPRs: 327 -> 327 (0.00 %) >>>> Scratch VGPRs: 1656 -> 1652 (-0.24 %) dwords per thread >>>> Code Size: 36127192 -> 35559780 (-1.57 %) bytes >>>> LDS: 767 -> 767 (0.00 %) blocks >>>> Max Waves: 212464 -> 212672 (0.10 %) >>>> Wait states: 0 -> 0 (0.00 %) >>>> >>>> PERCENTAGES / App Shaders SGPRs VGPRs SpillSGPR SpillVGPR >>>> Scratch CodeSize MaxWaves Waits >>>> (unknown) 4 . . . . . >>>> . . . >>>> 0ad 6 . . . . . >>>> . . . >>>> alien_isolation 2938 . 0.04 % -8.53 % . . >>>> -0.71 % -0.06 % . >>>> anholt 10 . . . . . >>>> . . . >>>> batman_arkham_origins 589 . -0.58 % -79.54 % . . >>>> -6.72 % 0.57 % . >>>> bioshock-infinite 1769 . -0.65 % -89.32 % . . >>>> -4.73 % 0.48 % . >>>> borderlands2 3968 . -0.31 % -51.21 % . . >>>> -4.09 % 0.22 % . >>>> brutal-legend 338 . -0.03 % -2.95 % . . >>>> -0.06 % . . >>>> civilization_beyond.. 116 . . -14.17 % . . >>>> -0.88 % . . >>>> counter_strike_glob.. 1142 . . . . . >>>> . . . >>>> dirt-showdown 541 . -0.56 % -40.14 % . >>>> -3.45 % -1.82 % 0.35 % . >>>> dolphin 22 . . . . . >>>> 0.16 % . . >>>> dota2 1747 . . . . . >>>> 0.01 % . . >>>> europa_universalis_4 76 . -0.23 % -42.11 % . . >>>> -0.96 % . . >>>> f1-2015 774 . -0.09 % -28.89 % . . >>>> -2.60 % 0.09 % . >>>> furmark-0.7.0 4 . . . . . >>>> . . . >>>> gimark-0.7.0 10 . . . . . >>>> . . . >>>> glamor 16 . . . . . >>>> . . . >>>> humus-celshading 4 . . . . . >>>> . . . >>>> humus-domino 6 . . . . . >>>> . . . >>>> humus-dynamicbranching 24 . 0.71 % . . . >>>> 0.29 % -0.45 % . >>>> humus-hdr 10 . . . . . >>>> . . . >>>> humus-portals 2 . . . . . >>>> . . . >>>> humus-volumetricfog.. 6 . . . . . >>>> . . . >>>> left_4_dead_2 1762 . . . . . >>>> . . . >>>> metro_2033_redux 2670 . -0.10 % -7.15 % . . >>>> -0.03 % . . >>>> nexuiz 80 . . . . . >>>> . . . >>>> pixmark-julia-fp32 2 . . . . . >>>> . . . >>>> pixmark-julia-fp64 2 . . . . . >>>> . . . >>>> pixmark-piano-0.7.0 2 . . . . . >>>> . . . >>>> pixmark-volplosion-.. 2 . . . . . >>>> . . . >>>> plot3d-0.7.0 8 . . . . . >>>> . . . >>>> portal 474 . . . . . >>>> . . . >>>> sauerbraten 7 . . . . . >>>> . . . >>>> serious_sam_3_bfe 392 . . -13.20 % . . >>>> -1.81 % . . >>>> supertuxkart 4 . . . . . >>>> . . . >>>> talos_principle 324 . -0.21 % -18.39 % . . >>>> -2.73 % 0.14 % . >>>> team_fortress_2 808 . . . . . >>>> . . . >>>> tesseract 430 . 0.08 % -68.57 % . . >>>> -0.45 % . . >>>> tessmark-0.7.0 6 . . . . . >>>> . . . >>>> thea 172 . . . . . >>>> 0.03 % . . >>>> ue4_effects_cave 299 . -0.04 % -10.15 % . . >>>> -0.25 % 0.04 % . >>>> ue4_elemental 586 . -0.02 % -13.93 % . . >>>> -0.13 % 0.02 % . >>>> ue4_lightroom_inter.. 74 . -0.17 % -70.00 % . . >>>> -1.27 % . . >>>> ue4_realistic_rende.. 92 . . -32.58 % . . >>>> -0.35 % . . >>>> unigine_heaven 322 . 0.12 % -54.17 % . . >>>> -1.42 % -0.12 % . >>>> unigine_sanctuary 264 . . . . . >>>> . . . >>>> unigine_tropics 210 . . . . . >>>> . . . >>>> unigine_valley 278 . -0.15 % -40.74 % . . >>>> -2.00 % 0.09 % . >>>> unity 72 . . . . . >>>> 0.03 % . . >>>> warsow 176 . . . . . >>>> . . . >>>> warzone2100 4 . . . . . >>>> 0.13 % . . >>>> witcher2 1040 . -0.03 % -86.28 % . . >>>> -0.28 % 0.01 % . >>>> xcom_enemy_within 1236 . -0.24 % -63.54 % . . >>>> -0.93 % 0.18 % . >>>> yofrankie 82 . -0.61 % -100.00 % . . >>>> -0.83 % 0.41 % . >>>> ----------------------------------------------------------------------------------------------------------- >>>> Total 26002 . -0.13 % -27.03 % . >>>> -0.24 % -1.57 % 0.10 % . >>>> --- >>>> src/gallium/drivers/radeonsi/si_shader.c | 15 +++++++++++---- >>>> 1 file changed, 11 insertions(+), 4 deletions(-) >>>> >>>> diff --git a/src/gallium/drivers/radeonsi/si_shader.c >>>> b/src/gallium/drivers/radeonsi/si_shader.c >>>> index 31140b1..b23c7c6 100644 >>>> --- a/src/gallium/drivers/radeonsi/si_shader.c >>>> +++ b/src/gallium/drivers/radeonsi/si_shader.c >>>> @@ -33,6 +33,7 @@ >>>> #include "gallivm/lp_bld_arit.h" >>>> #include "gallivm/lp_bld_bitarit.h" >>>> #include "gallivm/lp_bld_flow.h" >>>> +#include "gallivm/lp_bld_misc.h" >>>> #include "radeon/r600_cs.h" >>>> #include "radeon/radeon_llvm.h" >>>> #include "radeon/radeon_elf_util.h" >>>> @@ -5311,11 +5312,17 @@ static void si_create_function(struct >>>> si_shader_context *ctx, >>>> for (i = 0; i <= last_sgpr; ++i) { >>>> LLVMValueRef P = LLVMGetParam(ctx->radeon_bld.main_fn, i); >>>> >>>> - /* We tell llvm that array inputs are passed by value to allow >>>> Sinking pass >>>> - * to move load. Inputs are constant so this is fine. */ >>>> - if (i <= last_array_pointer) >>>> + /* The combination of: >>>> + * - ByVal >>>> + * - dereferenceable >>>> + * - tbaa >>>> + * allows the optimization passes to move loads and reduces >>>> + * SGPR spilling significantly. >>>> + */ >>>> + if (i <= last_array_pointer) { >>>> LLVMAddAttribute(P, LLVMByValAttribute); >>>> - else >>>> + lp_add_attr_dereferenceable(P, UINT64_MAX); >>> >>> I'm not sure if anything bad will happen passing UINT64_MAX here, but >>> is there any way we can get the buffer size here, so we can >>> pass a more accurate value? >>> >> >> Actually, these are all array pointers, so we should be able to get the >> size pretty easily. > > Matt on IRC suggested -1 (the unsigned maximum). I don't know if > indirect indexing will fail the isDereferenceable test with the real > size. Is there a reason to use the real size other than "so it doesn't > look silly in the printed IR"? In my opinion, we should set > dereferenceable(UINT64_MAX) on all pointers, because the GPU doesn't > terminate the program on bad pointer dereference like CPUs do. > > Marek
-1 is usually used in places that need an unknown size, e.g. lifetime intrinsics
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
