Commit Graph

21 Commits

Author SHA1 Message Date
AN0DIN 83db2efec3 shader: pack v_cvt_pkrtz_f16_f32 with the native half conversion (GTA V) (#1520)
* shader: pack v_cvt_pkrtz_f16_f32 with the native half conversion

The round-toward-zero pack was emulated with integer operations: about 20 to 30 operations
per half. GTA V uses it in 613 of its 656 pixel shaders, about 16 times in each G-buffer
pixel shader.

The pair is now packed with PackHalf2x16, unpacked again, and each half whose magnitude
exceeds its input's takes one unit off. Half signs and magnitudes are separate, so that is
the next value toward zero, whatever rounding the native conversion uses as long as it
returns one of the two neighbours. Same bits as before for every finite input and infinity;
a NaN stays a NaN. Checked against the old code on 2 million values on the CPU, and on an
Apple M5 GPU through Metal (400 000 values, subnormals included, safe and fast math).

GTA V on an Apple M5 (MoltenVK), 1080p, fixed script camera, three places, A B B A between
launches: -1.7 to -2.8 ms per frame (+4 to +6 %).

Scope: generic
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* shader: remove half-pack commentary and unused shift helper

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: nmzik <Nmzik@mail.ru>
2026-10-16 02:56:35 +02:00
nmzik 7b9997baea shader: fold exact float class masks into native NaN predicates 2026-10-09 02:54:21 +02:00
nmzik ee4dd90649 shader: emit native integer operations and remove signedness casts 2026-10-09 02:54:21 +02:00
nmzik 8abf99c56e shader: emit bounded SPIR-V operands without temporary vectors 2026-10-09 01:58:39 +02:00
nmzik 5b826c6e4a shader: emit native 64-bit operations where compatible 2026-10-05 20:41:03 +04:00
nmzik 77d7722173 shader: disable explicit MAD denormal checks for performance testing
Comment out the five explicit flush checks and record the substantial 3DMiniGolf menu FPS drop on NVIDIA RTX 5080 Laptop GPU. Preserve separate multiply/add rounding and both NoContraction decorations.

The matched runs used silent logging, no debugger, no Vulkan validation, and the cd9fc6ff revert on both revisions. Native denormal behavior remains GPU-dependent.
2026-10-05 18:39:31 +04:00
nmzik d8be245c22 shader: separate legacy FP32 MAD rounding from fused FMA
Co-authored-by: shadowbeat070 <70843316+shadowbeat070@users.noreply.github.com>
2026-10-05 04:11:48 +04:00
nmzik 8274e517b2 shader: add FP64 min, max and rounding
Co-authored-by: Çağatay Güçlü <33038942+defektu@users.noreply.github.com>
2026-10-05 02:24:51 +04:00
Senaxx 0d0b215fab shader: port 64-bit ALU and atomic opcodes 2026-10-02 00:51:15 +04:00
gohansharma79-svg 0a5774be7b shader: support FP64 conversions used by FighterZ (#884)
shader: support float and unsigned widening to FP64

Co-authored-by: nmzik <Nmzik@mail.ru>
2026-09-28 17:39:31 +04:00
nmzik 16b83a034c Shader: execute RDNA2 FP64 shadow arithmetic and log denorm limits
Shader: disable FP64 denorm preservation controls for now
2026-09-26 15:15:35 +04:00
nmzik 1208d22547 Fix Windows Clang SPIR-V serialization with typed operands 2026-09-12 13:08:23 +02:00
nmzik 20e57994cf shader: reuse native emitters and simplify memory dispatch 2026-09-12 13:42:15 +04:00
nmzik 59ef88c959 shader: generate SPIR-V dispatch and marshal handler arguments 2026-09-12 13:42:15 +04:00
nmzik 1d1f00bdef shader: use Khronos definitions and simplify compiler ownership 2026-09-12 13:42:15 +04:00
nmzik 9f49ac8249 shader: reuse SPIR-V helpers and memory resources 2026-09-08 21:26:59 +04:00
nmzik 85cf965d54 refactor(shader): centralize typed resource metadata 2026-08-21 03:32:47 +02:00
nmzik 6e9f162bd4 refactor(shader): compact paired 64-bit lowering 2026-08-21 03:32:47 +02:00
nmzik 5383972e96 refactor(shader): preserve native wide buffer operations 2026-08-21 03:32:47 +02:00
nmzik d34e30d892 refactor(shader): emit SPIR-V declarations on demand 2026-08-21 03:32:47 +02:00
nmzik 01275fb211 shader_recompiler: rename/clean up typed SSA SPIR-V emitter 2026-08-18 03:53:45 +02:00