Maximum Performance TinyCore Packages

Flags

• -Ofast
  Enables all -O3 optimizations plus unsafe math and standard-breaking transforms (e.g., assumes no NaNs or signed overflow).

• -fmerge-all-constants
  Deduplicates identical constants (arrays, strings, etc.) across translation units to reduce size and improve cache efficiency.

• -fno-semantic-interposition
  Assumes symbols are not replaced at runtime, allowing cross-module inlining and better optimization of function calls and globals.

• -ftree-vectorize
  Enables vectorization at the GIMPLE tree level for both loops and straight-line code, using SIMD instructions where possible.

• -fipa-pta
  Performs whole-program pointer analysis across functions to improve alias detection and unlock more aggressive optimization.

• -funroll-loops
  Fully unrolls loops with constant bounds to remove loop control overhead and increase ILP (instruction-level parallelism).

• -floop-nest-optimize
  Enables Pluto-based loop restructuring using ISL to improve cache locality and expose parallelism. Experimental.

---

Warning

These flags prioritize performance over portability and standards compliance.

Package Optimization Summary

Package	Version	Notes
MicroPython	1.25.0
Wine (Staging)	10.12	-Ofast broken: -O3 x86-64-v4 broken: Ignored
OpenTTD	Main	x86-64-v4 broken: Ignored
PaleMoon	33.8.0r2	x86-64-v4, -Ofast, alll broken

Supported CPU Architectures

The binaries are currently compiled and available as artifacts for the following processors:

Level	Enabled SIMD extensions	Since
x86-64	MMX, SSE, SSE2	AMD64, K8, Prescott
x86-64-v2	SSE3, SSE4	Nehalem, Silvermont, Bulldozer, Jaguar, Nano, Eden "C"
x86-64-v3	AVX, AVX2, FMA	Haswell, Gracemont, Excavator, QEMU 7.2+
x86-64-v4	AVX512	Skylake-X, Zen 4

Aditional CPU Architectures

Feel free to send a pr with your cpu type

Arch	Features
alderlake	MOVBE, MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AES, PREFETCHW, PCLMUL, RDRND, XSAVE, XSAVEC, XSAVES, XSAVEOPT, FSGSBASE, PTWRITE, RDPID, SGX, GFNI-SSE, CLWB, MOVDIRI, MOVDIR64B, WAITPKG, ADCX, AVX, AVX2, BMI, BMI2, F16C, FMA, LZCNT, PCONFIG, PKU, VAES, VPCLMULQDQ, SERIALIZE, HRESET, KL, WIDEKL, AVX-VNNI, UINTR, VXIFMA, AVXVNNIINT8, AVXNECONVERT and CMPCCXADD