Port upstream float and denormal peepholes (U1) by RINNECODER · Pull Request #477 · RINNECODER/AnyPS5

RINNECODER · 2026-10-13T23:33:09Z

Ports the float and denormal peepholes of unit U1 from upstream open PRs. Each PR is its own commit with "Ported from upstream PR #N" in the body. Refs upstream-pr-batch-1.

Ported (11 PRs, 12 commits, in rebase order)

perf(shader): pick the f32 NaN result with one quiet and one fallback select boykopovar/AnyPS5#4287 pick the f32 NaN result with one quiet and one fallback select
perf(shader): drop the output flush of f32 products boykopovar/AnyPS5#4539 drop the output flush of f32 products
perf(shader): treat f32 min, max, med3, floor and ceil results as denormal-free boykopovar/AnyPS5#4490 treat f32 min, max, med3, floor and ceil results as denormal-free (conflict with perf(shader): drop the output flush of f32 products boykopovar/AnyPS5#4539 in the DenormalFlushEliminator switch resolved by keeping both case groups)
perf(shader): treat f32 fract, sin and cos results as denormal-free boykopovar/AnyPS5#4737 treat f32 fract, sin and cos results as denormal-free
perf(shader): drop flushes of f32 results that cannot be denormal boykopovar/AnyPS5#5086 drop flushes of f32 results that cannot be denormal (trunc, round-even, sqrt, rsqrt, sign-bit and exponent-bit masks; both case groups and both tests kept)
fix(shader): take the fused result when an f16 fma's sum is exact boykopovar/AnyPS5#4293 take the fused result when an f16 fma's sum is exact (identical to fix(shader): take the fused result when an f16 fma's sum is exact boykopovar/AnyPS5#5082, one taken)
perf(shader): take the f32 mul, fma and add result fixups only when needed boykopovar/AnyPS5#5400 take the f32 mul, fma and add result fixups only when needed (SelectFlags now has notDenormal and rarelyTrue)
fix(shader): read lane 63 of half-wave reductions whose keys exclude helper lanes boykopovar/AnyPS5#5189 read lane 63 of half-wave reductions whose keys exclude helper lanes (+ its search-order test commit)
perf(shader): find the ds_permute_b32 source lane with ballots boykopovar/AnyPS5#5334 find the ds_permute_b32 source lane with ballots (uses OpGroupNonUniformBallot / FindUMsb, both already handled by the Metal bridge)
perf(shader): mark runtime metadata buffers read-only boykopovar/AnyPS5#5297 mark runtime metadata buffers read-only
perf(shader): fence a CAS-loop LDS atomic once boykopovar/AnyPS5#5399 fence a CAS-loop LDS atomic once

Metal check: every denormal-free proof only removes a flush where the value is already non-denormal by construction (min/max/med3 select an operand, floor/ceil/trunc/round-even give integers, sqrt/rsqrt cannot reach the denormal range), so it holds with float16DenormPreserve false and Metal fast-math flushing. No peephole relies on f16 denormals.

Skipped

fix(shader): take the fused result when an f16 fma's sum is exact boykopovar/AnyPS5#5082: identical to fix(shader): take the fused result when an f16 fma's sum is exact boykopovar/AnyPS5#4293, taken once.
perf(shader): round f16 interpolation and mix results in f32 when exact boykopovar/AnyPS5#4171 (round f16 interpolation and mix results in f32 when exact): conflicts with main, which has no f64 on Metal and already returns a single fused f32 Interpolate path before the exact wide-sum code. The PR only reshapes that exact path, so there is nothing to gain here.

Cache version

ShaderDiskCache::FormatVersion 32 -> 33 in one last commit, since emitted shader code changes. Other units that also bump it should rebase onto the highest value rather than add another bump. The request-serializer version is untouched.

Evidence

Build: x86_64 host config for agc_denormal_flush_elimination_tests; x86_64 -DANYPS5_ROSETTA_LIBS=ON for rosetta_libs rosetta_engine rosetta_launcher_tests metal_utility_tests rosetta_direct_graphics_tests, no compile errors. shader_recompiler_agc_driver compiles all touched backend files.
ctest: agc_denormal_flush_elimination 1/1 (includes the new perf(shader): drop the output flush of f32 products boykopovar/AnyPS5#4539, perf(shader): treat f32 min, max, med3, floor and ceil results as denormal-free boykopovar/AnyPS5#4490, perf(shader): treat f32 fract, sin and cos results as denormal-free boykopovar/AnyPS5#4737, perf(shader): drop flushes of f32 results that cannot be denormal boykopovar/AnyPS5#5086 cases); metal_utility_tests, rosetta_direct_graphics, rosetta_launcher 4/4.
Not run: the AGC execution tests added or touched by fix(shader): take the fused result when an f16 fma's sum is exact boykopovar/AnyPS5#4293, fix(shader): read lane 63 of half-wave reductions whose keys exclude helper lanes boykopovar/AnyPS5#5189, perf(shader): find the ds_permute_b32 source lane with ballots boykopovar/AnyPS5#5334, perf(shader): take the f32 mul, fma and add result fixups only when needed boykopovar/AnyPS5#5400 (ShaderMemory, DsPermuteLanes, F32DenormalFlush). The agc_shader_memory_tests target does not build on the macOS host config on main either (pthread_getattr_np in GuestMemory.cpp), and the execution tests need a Vulkan device.
Game smoke, 300 s wall, through the eboot lock, on AC power, scripted menu input, Rosetta direct run, private shader cache copy, same recipe for both runs:
- this branch: exit 0 at 300.5 s, 3223 flips, 0 skipped draws or dispatches, 0 shader disk cache load failures, 0 conversion failures, no fault; the shader cache showed 0 hits and 93+ misses by the end, so the changed shaders were compiled fresh. Reached the post-boot menus; final captures stable at 535 distinct colours.
- main (same build dir, the 13 changed source files reverted): exit 0 at 301.4 s, 2169 flips, 0 skipped, no fault; final captures stable at 497 distinct colours.
- Neither run reached the 3D map in 300 s (the recorded main baseline reaches it at about 296 s of a 360 s run), so fps on the map is not recorded. Flip counts are noisy because other agents shared the machine.
No upstream bug found beyond the Metal adaptation, so no upstream-comments draft.

🤖 Generated with Claude Code

https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

… select nanResultF32 gives the f32 add, sub, subrev, mul and fma forms, the legacy mul, mad and fma forms, v_div_fmas_f32 and v_fma_mix_f32 with f32 sources the AMD NaN result: the first NaN source, quieted in IEEE mode, else the default NaN 0xffc00000 when the result is NaN, else the result. It started from the fallback and then, for each source, tested it for NaN with an integer and and compare, quieted it with its own or, and selected it. The addend's test was also combined with the inverse of the invalid-product flag. The chain now carries only source bits. The addend starts it, replaced by 0 when the product is 0 * inf. Each earlier source that FPIsNan32 finds NaN takes its place. One quiet and one select then choose between that source, when it is NaN, and the fallback. The candidate is NaN exactly when some eligible source is, and it is then the first one, so every result is the same. agc_driver_f32_arithmetic_nan already checks the source order and the 0 * inf product with a NaN addend for every form. Ported from upstream PR boykopovar#4287. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

In a FLOAT_MODE that flushes f32 denormals, writeOperand flushes every f32 result, and DenormalFlushEliminator removes the flush only when the value can't be a denormal. It had no rule for products, so every v_mul_f32 kept its output flush, although the tiny-product flush before it already leaves no denormal. The eliminator now proves a product denormal-free from three facts: - The tiny-product flush of a multiply (round to nearest even, no addend) returns a signed zero or a product at least 2^-126 in magnitude, so its select is flagged notDenormal. An fma isn't flagged, since its addend can cancel the result into a denormal. - The NaN select picks its first arm only when that arm is a NaN, so it is as denormal-free as its fallback. - A product by a constant that is ±0 or at least 1 in magnitude, of a denormal-free factor, isn't a denormal. A dependent chain of 2048 v_mul_f32 on an RTX 4060 runs 9% faster. v_mul_legacy_f32 loses the same flush. Results are unchanged. The f32 denormal flush test gets rows where a nonzero addend cancels a normal product into an exact denormal, which the fma output flush must turn into a signed zero. No row covered that before. Ported from upstream PR boykopovar#4539. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

…ormal-free In the float modes that flush f32 denormals, writeOperand flushes the result of v_min/max/min3/max3/med3_f32 and v_floor/ceil_f32. DenormalFlushEliminator drops a flush only when it can show the value isn't a denormal, and it had no rule for these opcodes, so their result flush stayed even when every operand had already been flushed. FPMin32, FPMax32, FPMinTri32, FPMaxTri32 and FPMedTri32 return one of their operands' bits, or the and or or of two zeros, so the result is denormal-free when every operand is. FPFloor32 and FPCeil32 return an integer, a zero, an infinity or a NaN whatever their operand. In IEEE mode the min and max results and the floor and ceil results go through selects with a quieted operand, which the eliminator can't follow, so their flush stays there. agc_denormal_flush_elimination covers the five min and max forms with flushed and with unknown operands, floor and ceil of an unknown value, and a max carried around a loop. Ported from upstream PR boykopovar#4490. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

In the float modes that flush f32 denormals, writeOperand flushes the result of v_fract_f32, v_sin_f32 and v_cos_f32 after flushing their input. DenormalFlushEliminator had no rule for FPFract32, FPSin or FPCos, so the result flush stayed. None of the three returns a denormal for an input that isn't one. EmitFPFract32 takes fract(|x|), which is |x| below 1 and otherwise a multiple of the input's ulp, and 1 - fract, which is at least 2^-24. EmitFPSin and EmitFPCos reduce the input exactly with x - roundEven(x), so the cycle and its distance from a zero of the function are 0 or at least 2^-126. Near a zero, TrigLinearNearZero returns the distance times about 6.28. Elsewhere the result is within 4 ulps of a series value of at least about 1.5e-3. In IEEE mode floatUnary quiets a NaN input with an or, which the eliminator can't follow, so the flush stays there. agc_denormal_flush_elimination covers the three opcodes with a flushed and with an unknown input. Ported from upstream PR boykopovar#4737. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

DenormalFlushEliminator removes f32 denormal flushes on values it can prove are not denormal. Recognises FPTrunc32, FPRoundEven32, square roots (FPSqrt), reciprocal square roots (FPRecipSqrt32), and bitwise operations that set exponent bits (BitwiseOr32) or mask to signed zeros (BitwiseAnd32). Adds unit tests in DenormalFlushElimination.cpp. Ported from upstream PR boykopovar#5086. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

fmaF16RoundedToOdd adds the f32 product and the addend separately and uses TwoSum to round to odd when the sum is inexact. When the error is zero it returned the separate sum. On MoltenVK (Apple M3 Pro) the separate multiply and add produce +0 where the IEEE result is -0: agc_driver_dpp16_operands got 0x0000 instead of 0x8000 in the high half of lane 1's row_shr v_pk_fmac_f16. The exact case now returns an f32 fma of the same operands. The product of two f16 values is exact in f32 (11-bit significands, exponents well inside the f32 range), so when TwoSum's error is zero the separate sum and the fused result are the same IEEE value, including the sign of zero; the change only matters on devices whose separate multiply and add do not keep it. The inexact path is unchanged. Ported from upstream PR boykopovar#4293. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

…eeded v_mul_f32, v_fma_f32, v_add_f32 and v_sub_f32 compute two fixups on every result: the AMD NaN result (the first NaN operand, quieted, or the default NaN) and, for products, the flush of results that are tiny after rounding to 24 bits. Both change the result only when it is a NaN or its magnitude is at most 2^-126. They were straight-line selects, so a v_mul_f32 took 16.5 SASS instructions on an RTX 4060 where the multiply is one. The translator now wraps the fixed result in a select that takes it only when the host result is NaN (add and sub) or NaN or at most 2^-126 in magnitude (mul and fma, round to nearest even). That select is flagged as rarely true, and the SPIR-V backend moves the instructions used only by its rare arm into a branch. The rare arm is the same IR as before, so results are unchanged. Loop headers, continue targets and values shared between the two halves of a wave64 program keep plain selects. A dependent chain of 2048 v_mul_f32 runs 2.2 times as fast, v_fma_f32 1.85 times and v_add_f32 1.13 times; on all-NaN inputs v_mul_f32 is 6% slower. Static SASS grows, since both paths are compiled. Ported from upstream PR boykopovar#5400. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

…helper lanes On a 32-lane host, a wave64 pixel program's v_readlane_b32 of lane 63 is emitted only when its half-wave scan's keys are proven zero for lanes that did not enter. The proof walk accepted the not-helper test (IEqual32 of HelperInvocation and 0) only inside ballots, not as a lane bit, and it searched conjunctions depth first: a key mask like exec & ~vcc inside a loop spent the 512-step budget on the loop's exec phis before it reached the helper test. Accept the helper test as a lane bit, and look for a false or helper conjunct before walking an operand. Resident Evil Village's wave-wide OR reductions (DPP row shifts, v_permlanex16, v_readlane of lanes 31 and 63) failed this way. Ported from upstream PR boykopovar#5189. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

The proof that a half-wave scan's keys are zero for lanes that did not enter walks a conjunction depth first under a 512-step budget. Narrowing exec with 600 compares before the scan puts the helper-lane test at the bottom of a 600-deep LogicalAnd chain in the keys: walking the first operand first spends the whole budget before reaching it, and the v_readlane_b32 of lane 63 is refused. Looking for a false or helper conjunct before walking an operand finds it, so lane 31 becomes a subgroup OR on a 32-lane host. Resident Evil Village's wave64 pixel programs key these scans by a loop's exec ahead of the helper test. Ported from upstream PR boykopovar#5189. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

ds_permute_b32 sends each active lane's data to lane ((addr + offset) >> 2) & 31 of its 32-lane group; the highest sending lane wins and a lane nobody writes to gets 0. It was emitted as one shuffle of the data, the address and the EXEC bit from each of the 32 source lanes: 96 shuffles and 144.5 SASS instructions per instruction on an RTX 4060 when the address changes between instructions. Each receiving lane now finds its writers directly: the ballot of the active lanes, intersected with the ballot of the lanes whose target has bit b set, or its complement, for each of the 5 target bits. The highest set bit is the winning lane, and one shuffle fetches its data. Results are unchanged. A chain of 256 dependent ds_permute_b32 runs 4.7 times faster. New test agc_driver_ds_permute_lanes compares wave32 and wave64 programs with random targets, heavy collisions and run-time EXEC masks against a model of the rule. Ported from upstream PR boykopovar#5334. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

Ported from upstream PR boykopovar#5297. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

SharedAtomicUpdate emitted a workgroup acquire-release fence after AtomicUpdate, which already ends its compare-and-swap loop with the same fence for LDS, so ds_rsub, ds_add_f32, ds_cmpst, ds_mskor, ds_wrap and the LDS float min/max fenced twice. The second fence is now emitted only for GDS, where the loop's own fence is device-scoped on uniform memory and differs from it. Ported from upstream PR boykopovar#5399. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

…le ports The ported float, denormal and wave-lane changes alter the emitted shader code, so cached entries from before them must not be reused. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

chatgpt-codex-connector · 2026-10-13T23:33:18Z

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-13T23:35:33.866162Z	`134d43d`	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

coderabbitai · 2026-10-13T23:33:20Z

Important

🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: defaults
Review profile: CHILL
Plan: Advanced
Run ID: 626d934c-f175-4ce0-80b0-486285c82b8c

Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

_{Comment @coderabbitai help to get the list of available commands.}

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 134d43d2ca

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-10-13T23:35:31Z

+        special = &ir.Emit(IrOpcode::FPIsNan32, IrType::U1, {plain});
    }
-    return &ir.BitCastF32(*bits);
+    return &ir.Emit(IrOpcode::SelectF32, IrType::F32, {special, fixed, plain}, SelectFlags{true});


Set rarelyTrue instead of notDenormal

SelectFlags{true} initializes the first field, notDenormal, while leaving rarelyTrue false. When output-denormal flushing is enabled and an add/subtract or canceling FMA produces a subnormal, writeOperand adds a flush, but DenormalFlushEliminator now trusts this flag and removes that required flush, returning the subnormal instead of signed zero; the intended rare-arm branching is also never enabled. Initialize this select with SelectFlags{false, true}.

Useful? React with 👍 / 👎.

karanvora2599 and others added 13 commits October 13, 2026 18:32

perf(shader): mark runtime metadata buffers read-only

d28fb65

Ported from upstream PR boykopovar#5297. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DgW4n3qYgi5jAAohEmXRQj

RINNECODER added the upstream-sync Upstream boykopovar/AnyPS5 intake label Oct 13, 2026

chatgpt-codex-connector Bot reviewed Oct 13, 2026

View reviewed changes

RINNECODER merged commit 1bfe83e into main Oct 13, 2026
6 checks passed

RINNECODER deleted the upr1/float-peepholes branch October 13, 2026 23:36

Navigation Menu

Repository navigation

Port upstream float and denormal peepholes (U1) - #477

Port upstream float and denormal peepholes (U1)#477
RINNECODER merged 13 commits into
mainfrom
upr1/float-peepholes

RINNECODER commented Oct 13, 2026

Uh oh!

chatgpt-codex-connector Bot commented Oct 13, 2026 •

edited

Loading

Uh oh!

coderabbitai Bot commented Oct 13, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Uh oh!

chatgpt-codex-connector Bot Oct 13, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

Conversation

RINNECODER commented Oct 13, 2026

Ported (11 PRs, 12 commits, in rebase order)

Skipped

Cache version

Evidence

Uh oh!

chatgpt-codex-connector Bot commented Oct 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codex Review Summary

Uh oh!

coderabbitai Bot commented Oct 13, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Oct 13, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

chatgpt-codex-connector Bot commented Oct 13, 2026 •

edited

Loading