Commit graph

427 commits

Author SHA1 Message Date
Eladash 5e110f2844 SPU LLVM: Try to precompile filler-sapce between functions 2023-09-02 12:31:11 +03:00
Eladash a626ccfcad SPU LLVM: Initial precompilation of tail-calls 2023-09-02 12:31:11 +03:00
Eladash a9810ccb72 SPU LLVM: Another fix for Game Collection's precompilation 2023-09-02 12:31:11 +03:00
Eladash e851c044b5 SPU: Function discovery fix
Do not detect branch to next.
2023-09-02 12:31:11 +03:00
Eladash f9f2657c98 SPU LLVM: Optimize spu_idisable 2023-09-01 18:08:15 +03:00
Eladash ee9477dc21 SPU: support pure SPU code precompilation discovery 2023-08-30 08:45:29 +03:00
Eladash 37212a632c SPU: Refactor function discovery 2023-08-30 08:45:29 +03:00
Elad Ashkenazi 105c5759f3 Add SPU Precompilation to Create PPU Cache 2023-08-30 08:45:29 +03:00
Elad Ashkenazi 3d2229ca05 SPU LLVM Precompilation Fixup 2023-08-28 13:33:43 +03:00
Eladash b5faf5800b SPU LLVM Precompilation
Implement function SPU function discovery in images or random SPU code
2023-08-28 09:03:56 +03:00
Malcolm Jestadt 290ff5b839 Zero register optimization for AVX-512-VBMI
- Take advantage of the fact that AVX instructions zero the upper 128 bits for a nice optimization when one input vector is zeroed
2023-08-28 05:09:30 +03:00
Eladash a001e6ef09 Progress Dialog: Fix race on PPU compilation status 2023-08-22 05:40:53 +03:00
Malcolm Jestadt f2e782f5dd SPU LLVM: Inline timer reads for WrDec and RdDec
- Uses RDTSC to emulate the spu decrementer
2023-08-13 00:16:35 +03:00
Malcolm Jestadt 512f0a814c SPU LLVM: Fix for AVX-512 CFLTU path
- vcvvtps2udq doesn't turn negative numbers into 0, fix by using signed integer max with 0 instead of vrangeps
2023-08-12 02:55:08 +03:00
Eladash 2a0278fbb1 Fixup SPU/PPU Cache Abortion 2023-08-06 21:37:10 +03:00
Megamouse 343ba8733b Merge xfloat options 2023-08-06 09:30:53 +03:00
Ivan Chikish d34287b2cc Linux: use futex_waitv syscall for atomic waiting
In order to make this possible, some unnecessary features were removed.
2023-08-02 21:46:06 +03:00
Whatcookie fd6829f757
SPU LLVM: AVX-512 optimization for CFLTU (#14384)
- Takes advantage of vrangeps and the new float to uint instructions from AVX-512
- Down from 6 to 3 instructions

TODO: Somehow ensure that this is what llvm outputs using CreateFPToUI?
2023-07-29 09:01:01 +03:00
Whatcookie 4ecb06c901
SPU LLVM: Optimize common SFI+ROTQMBY pattern 2023-07-28 10:26:40 +03:00
Elad Ashkenazi 9265ff53d0
Include spu.log inside RPCS3.log when SPU Debug is true 2023-07-27 19:15:32 +03:00
Eladash 50dad6801b SPU LLVM: Use get_known_bits() in SHUFB 2023-07-18 22:27:45 +03:00
Malcolm Jestadt ee7475a9d4 SPU LLVM: Handle SHUFB special cases with a lookup table
- Needs 3 instructions to handle the special cases, since x86 lacks an 8 bit simd shift instruction
2023-07-18 22:27:45 +03:00
oltolm 0c94606fcf
Make compile with msvc, clang and gcc on Windows 2023-07-11 21:40:30 +03:00
Eladash 482dd0e8f8 SPU: Remove wrong clamp in MFC_Size
Just crashes real MFC.
2023-07-09 13:33:03 +03:00
RipleyTom cbb1b1f28e Fix spu_fm 2023-05-19 18:26:42 +03:00
RipleyTom f11770b88b
Better accuracy for FREST/FRSQEST (#13863) 2023-05-15 17:20:47 +01:00
RipleyTom 5c0113ce59 Deterministic FREST and FRSQEST 2023-05-06 12:59:34 +03:00
Ivan Chikish bb8e43f16c SPU LLVM: fixup custom LICM pass 2023-04-22 03:07:06 +03:00
Ivan Chikish 1041284384 SPU LLVM: sink stores deeper in custom LICM pass 2023-04-21 18:11:59 +03:00
Ivan Chikish 183bea3b98 SPU LLVM: upgrade custom DSE pass 2023-04-20 11:12:31 +03:00
Ivan Chikish 39d17a94c6 SPU LLVM: make savestates unsavable inside the code 2023-04-18 12:19:15 +03:00
Ivan Chikish 8153e5338f SPU LLVM: optimize register stores
Custom implementation of DSE+LICM
2023-04-18 12:19:15 +03:00
Ivan Chikish 44b3709d1d SPU LLVM: use volatile stores for PC update 2023-04-15 12:40:59 +03:00
Ivan Chikish ba29f0ccd1 SPU LLVM: use atomic loads in read channel count 2023-04-14 13:36:04 +03:00
Ivan Chikish 3473e19508 SPU LLVM: fix savestate safety guards
Volatile was removed since it prevented optimizations.
2023-04-14 07:26:30 +03:00
RipleyTom d35fecbeea Forces deterministic FP operations when online 2023-04-12 15:31:36 +03:00
Ivan Chikish 06b0e35fb9 Update to LLVM 16.0.1
Fix Zen4+ AVX-512 detection
2023-04-11 12:13:09 +03:00
oltolm 6fbca1acfd remove unnecessary pointer bitcasts 2023-04-09 12:45:18 +03:00
Ivan Chikish fb88e1c1c9 Update to LLVM 16.0.0, switch to upstream LLVM 2023-04-06 10:19:31 +03:00
oltolm cf5346c263 use new LLVM API in SPURecompiler 2023-03-12 10:11:06 +03:00
Ivan Chikish 776b3b5efa SPU LLVM: fix regression from #13500
Fixes #13526
2023-03-11 19:48:55 +03:00
oltolm 520524285a
llvm: update code to new API (#13500)
* llvm: update code to new API

* llvm: remove OLDLLVM define
2023-03-11 01:57:21 +03:00
Malcolm Jestadt 813f7b50c1 SPU LLVM: Minor SUMB AVX-512 path optimization
- Tweak shuffle to allow LLVM to emit a cheap blend instruction instead of the expensive VPERMI2W instruction
2023-01-27 13:06:48 +03:00
Eladash 2a00a88e2a SPU LLVM: don't force-enter process_mfc_cmd() because it's slower 2022-10-04 16:28:34 +03:00
Malcolm Jestadt d8897c585d PPU/SPU LLVM: Allow Zen4 cpus to use VPERMI2B/VPERMT2B instead of the vperm2b256to128 path
- Zen4 based cpus can process VPERM2B in a single uop, unlike intel where it is 3 uops.
2022-10-01 15:38:29 +03:00
Nekotekina 6ff6a4989a Implement at32() util
Works like .at() but uses source location for "exception".
2022-09-26 18:04:15 +03:00
Nekotekina b49a1f27eb Warning fixes 2022-09-17 16:35:02 +03:00
Nekotekina 5985f0eefa BufferUtils: cleanup regarding ARM64 2022-09-07 17:59:07 +03:00
sguo35 a0d48c588a spu/arm64: clean up assembly code generation
Clean up asmjit usage so we don't unnecessarily allocate memory
anymore for SPURecompiler functions.
2022-09-07 17:33:01 +03:00
Eladash ee1384341e rsx: Implement atomic vertex upload (with Strict Rendering Mode) 2022-09-01 20:09:28 +03:00