3808 commits

Author	SHA1	Message	Date
Georg Lehmann	46c1bd1147	aco: add a dedicated pass for better float MODE insertion ... Foz-DB Navi48: Totals from 14 (0.02% of 80251) affected shaders: Instrs: 13998 -> 11684 (-16.53%) CodeSize: 104464 -> 86260 (-17.43%) Latency: 108722 -> 106667 (-1.89%) InvThroughput: 100332 -> 100324 (-0.01%) VClause: 621 -> 595 (-4.19%); split: -4.99%, +0.81% VALU: 6875 -> 6871 (-0.06%) SALU: 3256 -> 1015 (-68.83%) VOPD: 1328 -> 1332 (+0.30%) Removes the s_setreg spam in FSR4. Reviewed-by: Rhys Perry <pendingchaos02@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35746>	2025-07-10 13:48:50 +00:00
Daniel Schürmann	610a19cf31	aco/isel: allow to select SGPR defs for vectorized bcsel and logical operations ... No fossil changes. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35784>	2025-07-09 14:10:37 +00:00
Daniel Schürmann	d7477111d2	aco: split vectorized bcsel and bitwise logic VGPR definitions ... This has a slightly negative effect on parallel-rdp, but positively affects FSR4. Totals from 14 (0.02% of 79839) affected shaders: (Navi48) Instrs: 63543 -> 63646 (+0.16%); split: -0.01%, +0.17% CodeSize: 352888 -> 353608 (+0.20%); split: -0.02%, +0.23% Latency: 1822354 -> 1825036 (+0.15%) InvThroughput: 364683 -> 365738 (+0.29%); split: -0.04%, +0.32% Copies: 9299 -> 9363 (+0.69%); split: -0.11%, +0.80% PreVGPRs: 1381 -> 1394 (+0.94%) VALU: 34511 -> 34575 (+0.19%); split: -0.03%, +0.21% Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35784>	2025-07-09 14:10:36 +00:00
Daniel Schürmann	764ee3a834	radv: don't lower subdword phis to scalar ... Totals from 193 (0.24% of 79839) affected shaders: (Navi48) MaxWaves: 6004 -> 6024 (+0.33%) Instrs: 169276 -> 166784 (-1.47%); split: -3.01%, +1.53% CodeSize: 940608 -> 915768 (-2.64%); split: -4.29%, +1.64% VGPRs: 8012 -> 7716 (-3.69%); split: -3.99%, +0.30% SpillVGPRs: 185 -> 0 (-inf%) Scratch: 13568 -> 0 (-inf%) Latency: 2159787 -> 2147084 (-0.59%); split: -2.86%, +2.28% InvThroughput: 664022 -> 395859 (-40.38%); split: -42.59%, +2.21% VClause: 2998 -> 2880 (-3.94%); split: -4.27%, +0.33% SClause: 3117 -> 3120 (+0.10%) Copies: 21290 -> 16278 (-23.54%); split: -24.74%, +1.20% Branches: 4757 -> 4760 (+0.06%); split: -0.34%, +0.40% PreSGPRs: 7369 -> 7378 (+0.12%); split: -0.11%, +0.23% PreVGPRs: 4257 -> 3859 (-9.35%); split: -9.94%, +0.59% VALU: 83173 -> 79804 (-4.05%); split: -5.68%, +1.63% SALU: 36672 -> 37318 (+1.76%); split: -0.02%, +1.78% VMEM: 4012 -> 3762 (-6.23%); split: -6.83%, +0.60% SMEM: 4300 -> 4303 (+0.07%) Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35784>	2025-07-09 14:10:36 +00:00
Daniel Schürmann	fc2fcac04e	aco: allow vectorized nir_op_mov ... nir_lower_phis_to_scalar() can create these with the next commit. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35784>	2025-07-09 14:10:36 +00:00
Daniel Schürmann	3f35b1329e	aco: allow subdword vector-definitions on some VALU instructions ... Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35784>	2025-07-09 14:10:36 +00:00
Daniel Schürmann	025306a95d	aco/isel: refactor emission of bitwise logical operations ... Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35784>	2025-07-09 14:10:36 +00:00
Georg Lehmann	9e8ba10447	aco/vn: remove dead instructions early ... Dead p_create_vector/p_split_vector left behind by instruction selection slow down the other passes and negatively affect extract labels in aco_optimizer. Foz-DB GFX1201: Totals from 964 (1.20% of 80251) affected shaders: MaxWaves: 29206 -> 29030 (-0.60%); split: +0.08%, -0.68% Instrs: 669369 -> 668842 (-0.08%); split: -0.16%, +0.09% CodeSize: 3385192 -> 3383216 (-0.06%); split: -0.13%, +0.07% VGPRs: 46788 -> 46848 (+0.13%); split: -0.85%, +0.97% Latency: 3985660 -> 3892742 (-2.33%); split: -2.54%, +0.21% InvThroughput: 538296 -> 536761 (-0.29%); split: -0.38%, +0.10% VClause: 8336 -> 8418 (+0.98%); split: -0.17%, +1.15% SClause: 17111 -> 17120 (+0.05%); split: -0.20%, +0.25% Copies: 44393 -> 44239 (-0.35%); split: -1.25%, +0.91% PreSGPRs: 45417 -> 45419 (+0.00%) PreVGPRs: 30401 -> 31644 (+4.09%); split: -0.00%, +4.09% VALU: 348282 -> 348167 (-0.03%); split: -0.15%, +0.12% SALU: 121454 -> 121410 (-0.04%); split: -0.04%, +0.01% Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35825>	2025-07-09 07:23:09 +00:00
Georg Lehmann	82af226690	aco: remove unused swap_srcs from emit_vop3p_instruction ... Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35825>	2025-07-09 07:23:09 +00:00
Georg Lehmann	96793fb0c1	aco/isel: implement 16bit vec2 shifts ... The source bit size mismatch is a bit annoying, but it's still worth it to vectorize these. Foz-DB Navi48: Totals from 85 (0.11% of 80251) affected shaders: Instrs: 119073 -> 118827 (-0.21%); split: -0.21%, +0.00% CodeSize: 669604 -> 667552 (-0.31%); split: -0.31%, +0.00% VGPRs: 4796 -> 4736 (-1.25%) Latency: 1907685 -> 1901983 (-0.30%); split: -0.32%, +0.02% InvThroughput: 642603 -> 640680 (-0.30%); split: -0.33%, +0.03% VClause: 2088 -> 2091 (+0.14%) Copies: 18300 -> 18394 (+0.51%); split: -0.01%, +0.52% Branches: 3452 -> 3440 (-0.35%) VALU: 63378 -> 63144 (-0.37%); split: -0.37%, +0.00% SALU: 23065 -> 23076 (+0.05%); split: -0.00%, +0.05% Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35825>	2025-07-09 07:23:08 +00:00
Daniel Schürmann	2c51a8870d	nir: add nir_vectorize_cb callback parameter to nir_lower_phis_to_scalar() ... Similar to nir_lower_alu_width(), the callback can return the desired number of components for a phi, or 0 for no lowering. The previous behavior of nir_lower_phis_to_scalar() with lower_all=true can be elicited via nir_lower_all_phis_to_scalar() while the previous behavior with lower_all=false now corresponds to nir_lower_phis_to_scalar() with NULL callback. Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io> Reviewed-by: Mel Henning <mhenning@darkrefraction.com> Reviewed-by: Georg Lehmann <dadschoorse@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35783>	2025-07-08 15:33:59 +00:00
Rhys Perry	34f1a8f707	aco: handle FPAtomicToDenormModeHazard ... This is quite unlikely to happen, but I guess it might be possible and it's relatively simple to work around. Signed-off-by: Rhys Perry <pendingchaos02@gmail.com> Reviewed-by: Georg Lehmann <dadschoorse@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35884>	2025-07-07 13:02:43 +00:00
Marek Olšák	4263b49778	ac/nir: remove ngg_scratch LDS ABI, allocate it in the lowering pass ... This is a cleanup. Old gs LDS layout: [es outputs][gs outputs][scratch] Old nogs LDS layout: [xfb/cull][scratch] New gs LDS layout: [es outputs][scratch|gs outputs] New nogs LDS layout: [scratch|xfb/cull] The LDS scratch is moved to the beginning of the preceding buffer in LDS, while the addresses in that LDS buffer are offset by the scratch size. It effectively merges the LDS scratch with the preceding buffer in LDS. Thanks to that, we no longer need the ngg_scratch ABI and the offset in a user SGPR. The lowering passes now return the LDS scratch size, which is used by the drivers to determine the final LDS size. The ngg_lds_layout SGPR is now unused without GS in RADV. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35352>	2025-07-02 20:27:41 +00:00
Rhys Perry	dce1d4ad4c	aco/ra: fix repeated compact_linear_vgprs() in get_reg() ... Signed-off-by: Rhys Perry <pendingchaos02@gmail.com> Fixes: b7738de4f9 ("aco/ra: rework linear VGPR allocation") Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/13431 Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35838>	2025-07-02 09:26:04 +00:00
Rhys Perry	21c4400278	aco: update ctx.block when inserting discard block ... Signed-off-by: Rhys Perry <pendingchaos02@gmail.com> Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/13432 Backport-to: 25.1 Reviewed-by: Georg Lehmann <dadschoorse@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35833>	2025-07-01 14:31:11 +00:00
Alyssa Rosenzweig	67237b6f1b	treewide: use nir_break_if ... Via Coccinelle patch: @@ expression builder, condition; @@ -nir_push_if(builder, condition); -{ -nir_jump(builder, nir_jump_break); -} -nir_pop_if(builder, NULL); +nir_break_if(builder, condition); Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io> Reviewed-by: Jesse Natalie <jenatali@microsoft.com> Reviewed-by: Mel Henning <mhenning@darkrefraction.com> Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35794>	2025-06-30 14:51:24 -04:00
Natalie Vock	af86cc37d5	aco/spill: Don't spill scratch_rsrc-related temps ... These temps are used to create the scratch_rsrc. Spilling them will never benefit anything, because assign_spill_slots will insert code that keeps them live. Since the spiller assumes all spilled variables to be dead, this can cause more variables being live than intended and spilling to fail. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35031>	2025-06-26 11:02:53 +00:00
Natalie Vock	acf29e403a	aco/spill: Add a null scratch offset if no scratch_offset arg exists ... Function callees' scratch_rsrc comes with the scratch offset pre-applied. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35031>	2025-06-26 11:02:53 +00:00
Natalie Vock	630913e1b4	aco: Introduce static_scratch_rsrc program member ... Function callees get their scratch resource as a parameter instead of generating it on-the-fly. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35031>	2025-06-26 11:02:53 +00:00
Natalie Vock	e006f68b11	aco/isel: Don't add scratch offset as gfx8- soffset if no offsets exist ... Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35031>	2025-06-26 11:02:53 +00:00
Natalie Vock	a5eba11657	aco/isel: Use stack pointer parameter in load/store_scratch ... Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35031>	2025-06-26 11:02:53 +00:00
Natalie Vock	4a62b342f3	aco: Add common utility to load scratch descriptor ... Also modifies the scratch descriptor to take the stack pointer into account. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35031>	2025-06-26 11:02:52 +00:00
Natalie Vock	cd2caa5e2b	aco/spill: Use scratch stack pointer ... Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35031>	2025-06-26 11:02:52 +00:00
Natalie Vock	22624d6f12	aco: Add scratch stack pointer ... Function callees shouldn't overwrite caller's stacks. Track where to write scratch data with a stack pointer. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35031>	2025-06-26 11:02:52 +00:00
Natalie Vock	be89c02be5	aco: Add pseudo instr to calculate a function callee's stack pointer ... Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35031>	2025-06-26 11:02:52 +00:00
Daniel Schürmann	7620957193	aco/ra: always set fill_operands=true when handling operands ... This makes the behavior consistent and less prone to error. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35735>	2025-06-26 10:05:07 +00:00
Daniel Schürmann	ee8424d839	aco/ra: always fill moved operands when handling vector-operands ... update_renames() assumes that killed operands are already removed from the register file, except for precolored and copy-kill operands. When dealing with vector-operands, however, unrelated operands might also be moved, in order to make space. Fixes: fb689f133e ('aco/ra: handle register assignment of vector-aligned operands') Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35735>	2025-06-26 10:05:07 +00:00
Samuel Pitoiset	e91029c82d	aco: consider that nir_tex_src_{coord,ddx} can be the first source ... Only -1 means it's not found, but 0 is still valid. Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35736>	2025-06-25 17:20:02 +00:00
Georg Lehmann	01d20680e2	aco/optimizer: generalize p_create_vector of split vector opt ... Foz-DB Navi48: Totals from 116 (0.14% of 80251) affected shaders: MaxWaves: 2965 -> 2972 (+0.24%) Instrs: 145933 -> 144632 (-0.89%); split: -0.91%, +0.02% CodeSize: 815968 -> 806512 (-1.16%); split: -1.20%, +0.04% VGPRs: 7240 -> 7144 (-1.33%); split: -1.66%, +0.33% Latency: 3065858 -> 3063802 (-0.07%); split: -0.11%, +0.05% InvThroughput: 745395 -> 743506 (-0.25%); split: -0.26%, +0.01% VClause: 3702 -> 3694 (-0.22%); split: -0.65%, +0.43% SClause: 3187 -> 3191 (+0.13%) Copies: 12716 -> 11804 (-7.17%); split: -7.42%, +0.25% Branches: 3501 -> 3503 (+0.06%) PreVGPRs: 5400 -> 5327 (-1.35%); split: -1.41%, +0.06% VALU: 76455 -> 75492 (-1.26%); split: -1.30%, +0.04% SALU: 23594 -> 23595 (+0.00%); split: -0.00%, +0.01% VOPD: 1478 -> 1527 (+3.32%); split: +4.67%, -1.35% Mostly helps FSR4. Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35674>	2025-06-25 11:03:30 +00:00
Georg Lehmann	001cd632ee	aco: select float8 to fp32 conversions ... Reviewed-by: Rhys Perry <pendingchaos02@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35434>	2025-06-23 07:59:27 +00:00
Georg Lehmann	19ca4be6b0	aco/isel: fix get_alu_src with 8bit vec2 source ... Reviewed-by: Rhys Perry <pendingchaos02@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35434>	2025-06-23 07:59:27 +00:00
Georg Lehmann	f047a67fba	nir,aco: optimize FP16_OFVL pattern created by vkd3d-proton ... Reviewed-by: Rhys Perry <pendingchaos02@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35434>	2025-06-23 07:59:27 +00:00
Georg Lehmann	9e6adcbca0	aco: select fp32 to float8 conversions ... Reviewed-by: Rhys Perry <pendingchaos02@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35434>	2025-06-23 07:59:26 +00:00
Georg Lehmann	3a45802514	aco/lower_to_hw: support saturating fp8 conversions ... Sadly amd only made this behavior controlable with global state. We add a new pseudo opcode for this purpose and change FP16_OVFL for each instruction. Ideally we would only do it once for clauses and after ilp scheduling, but this can be improved in the future. Reviewed-by: Rhys Perry <pendingchaos02@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35434>	2025-06-23 07:59:25 +00:00
Georg Lehmann	65650cfef8	aco: emit float8 wmma ... Reviewed-by: Rhys Perry <pendingchaos02@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35434>	2025-06-23 07:59:25 +00:00
Rhys Perry	325dfd809a	radv,aco: switch to shader statistics framework ... Signed-off-by: Rhys Perry <pendingchaos02@gmail.com> Gitlab: https://gitlab.freedesktop.org/mesa/mesa/-/issues/12756 Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35583>	2025-06-20 09:26:58 +00:00
Rhys Perry	2cfd2d3b1d	aco/tests: add lower_branches tests ... Signed-off-by: Rhys Perry <pendingchaos02@gmail.com> Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35202>	2025-06-19 10:58:39 +00:00
Rhys Perry	c45482e652	aco: validate that preds/succs match ... This isn't done in validate_cfg() because that's called less frequently. Signed-off-by: Rhys Perry <pendingchaos02@gmail.com> Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35202>	2025-06-19 10:58:39 +00:00
Rhys Perry	85db025cd7	aco: continue when try_remove_simple_block can't remove a predecessor ... We should update linear_preds so that the predecessors we can remove are actually removed. Signed-off-by: Rhys Perry <pendingchaos02@gmail.com> Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35202>	2025-06-19 10:58:38 +00:00
Rhys Perry	5344abbc56	aco/lower_branches: keep blocks with multiple logical successors ... It might be the case that both the branch and exec mask write in a divergent branch block are removed. try_remove_simple_block() might then try to remove it, but fail because it has multiple logical successors. Instead, just skip these blocks. Signed-off-by: Rhys Perry <pendingchaos02@gmail.com> Backport-to: 25.1 Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35202>	2025-06-19 10:58:38 +00:00
Georg Lehmann	001fe8c236	aco: optimize boolean phi with empty else block ... We can keep the else empty by handling the phi in the "then" block. Foz-DB Navi21: Totals from 921 (1.15% of 80065) affected shaders: Instrs: 4532598 -> 4527309 (-0.12%); split: -0.12%, +0.00% CodeSize: 24498484 -> 24481780 (-0.07%); split: -0.08%, +0.01% Latency: 41016915 -> 41020477 (+0.01%); split: -0.10%, +0.11% InvThroughput: 9998405 -> 9991873 (-0.07%); split: -0.08%, +0.02% SClause: 128261 -> 128267 (+0.00%) Copies: 409949 -> 408585 (-0.33%); split: -0.36%, +0.02% Branches: 169740 -> 169222 (-0.31%); split: -0.58%, +0.27% PreSGPRs: 64408 -> 64398 (-0.02%) VALU: 2972521 -> 2972518 (-0.00%) SALU: 673844 -> 668973 (-0.72%); split: -0.72%, +0.00% Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35165>	2025-06-19 07:32:43 +00:00
Georg Lehmann	88753ddd1d	aco: allow nir divergence to be printed again ... Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32990>	2025-06-19 07:02:20 +00:00
Samuel Pitoiset	d23de4918e	aco: add support for image f32 atomic add ... It's supported on GFX12. Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35493>	2025-06-13 08:47:59 +00:00
Pierre-Eric Pelloux-Prayer	3bcbd11a33	aco/isel: fix visit_tex handling of is_sparse ... For cases when less than 4 components are read, the original code would compute an incorrect dmask. eg: with a single component + is_sparse, the dmask was 0x13: - 0x 3 = coming from nir_def_components_read - 0x10 = the sparse bit While it should have at 2 bits set (1 for the color/depth, 1 for tfe). This caused problem when expand_vector() used the dmask to generate the final results, because the value for the sparse component was read from the wrong index. So after the call to emit_mimg() dmask needs to be adjusted because the components will be stored in order, so if mask is 0x11 the tfe value would be stored at invalid index=5 (while it should be at index=1). This fixes KHR-GL46.sparse_texture_clamp_tests.SparseTextureClampLookupResidency_texture_2d_depth_component16 and KHR-GL46.sparse_texture2_tests.SparseTexture2Lookup_texture_2d_depth_component16 with ACO. Reviewed-by: Marek Olšák <marek.olsak@amd.com> Reviewed-by: Rhys Perry <pendingchaos02@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35206>	2025-06-11 12:11:28 +00:00
Georg Lehmann	f36ac8434c	aco: add a readme entry for v_pk_cvt_u8_f32 ... Reviewed-by: Rhys Perry <pendingchaos02@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35391>	2025-06-10 07:32:05 +00:00
Georg Lehmann	94c191e6d9	aco: remove p_v_cvt_pk_u8_f32 ... Now unused. Reviewed-by: Rhys Perry <pendingchaos02@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35391>	2025-06-10 07:32:04 +00:00
Georg Lehmann	d95e90ab5f	aco: do not use v_cvt_pk_u8_f32 for f2u8 ... The ISA docs don't mention this, but instead of always truncating like other integer conversions, this opcode actually uses the single precision rounding mode. We could continue to use the opcode and set the rounding mode to rtz in lower_to_hw_instrs, but I think I should just concede that f2u8 isn't worth the effort. Fixes: 9bb10b58 ("aco: use v_cvt_pk_u8_f32 for f2u8") Reviewed-by: Rhys Perry <pendingchaos02@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35391>	2025-06-10 07:32:04 +00:00
Natalie Vock	a28515f096	aco/opt: Rename loop header phis ... Fossil stats on top of !35269: Totals from 133 (0.16% of 81077) affected shaders: Instrs: 4328456 -> 4327891 (-0.01%) CodeSize: 22890004 -> 22887732 (-0.01%); split: -0.01%, +0.00% Latency: 28406452 -> 28404732 (-0.01%) InvThroughput: 5361458 -> 5361153 (-0.01%) Copies: 376788 -> 376222 (-0.15%) VALU: 2429210 -> 2428645 (-0.02%) VOPD: 57 -> 56 (-1.75%) Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35270>	2025-06-09 14:36:44 +00:00
Rhys Perry	00dd0d0dd1	aco: update VALUReadSGPRHazard comment ... Signed-off-by: Rhys Perry <pendingchaos02@gmail.com> Reviewed-by: Georg Lehmann <dadschoorse@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35387>	2025-06-09 10:12:25 +00:00
Rhys Perry	a714a19e16	aco/gfx12: fix VALUReadSGPRHazard with carry-out ... fossil-db (gfx1201): Totals from 370 (0.46% of 79653) affected shaders: Instrs: 3933639 -> 3935914 (+0.06%) CodeSize: 20743448 -> 20752068 (+0.04%); split: -0.00%, +0.04% Latency: 26261246 -> 26261921 (+0.00%); split: -0.00%, +0.00% InvThroughput: 5363675 -> 5363760 (+0.00%); split: -0.00%, +0.00% Signed-off-by: Rhys Perry <pendingchaos02@gmail.com> Fixes: 65f95ae74e ("aco/insert_NOPs: implement VALU -> VALU case for VALUReadSGPRHazard on GFX12") Reviewed-by: Georg Lehmann <dadschoorse@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35387>	2025-06-09 10:12:25 +00:00