Commit Graph

4 Commits

Author SHA1 Message Date
6c171d4906 sx: step 14 — inline JIT primitives (-69% fib, -62% loop, -50% sum on bench_vm)
The bytecode compiler emitted OP_CALL_PRIM (52) for every primitive call, even
for arithmetic and comparison hot-paths. The VM had specialized opcodes
(OP_ADD, OP_SUB, OP_EQ, etc.) defined but unused.

- lib/compiler.sx (compile-call): emit specialized 1-byte opcode when the
  primitive name + arity matches one of {+, -, *, /, =, <, >, cons, not, len,
  first, rest}. Falls back to CALL_PRIM otherwise. fib bytecode: 50 → 38 bytes.
- hosts/ocaml/lib/sx_compiler.ml: mirror change in the auto-generated OCaml
  compiler so SXBC export from mcp_tree uses the same emission.
- hosts/ocaml/lib/sx_vm.ml: extend OP_ADD/SUB/MUL/DIV to handle Integer+Integer
  (not just Number+Number). Inline OP_EQ via Sx_runtime._fast_eq. Inline
  OP_LT/GT mixed-numeric comparisons. Avoids Hashtbl lookup on the fallback
  path for the common integer cases that dominate tight loops.
- hosts/ocaml/bin/bench_vm.ml: VM-only benchmark — loads compiler.sx via CEK,
  JIT-compiles each fn, measures Sx_vm.call_closure throughput.

Median improvements (best of 3 runs of 9-min, bench_vm.exe):
  fib(22)         107.87ms →  33.13ms   -69%
  loop(200000)    429.64ms → 161.16ms   -62%
  sum-to(50000)    72.85ms →  36.74ms   -50%
  count-lt(20000)  28.44ms →  17.58ms   -38%
  count-eq(20000)  37.23ms →  15.46ms   -58%

Tests: 4550/4550 OCaml passing (unchanged). Zero regressions.

Last step in the sx-improvements roadmap — all 14 steps complete.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 02:38:47 +00:00
397d0f39c0 Re-bootstrap compiler + render after IO registry and Step 5 changes
Compiler (lib/compiler.sx):
- Fix emit-op return type: 8 definition form cases (defstyle,
  defhandler, defpage, etc.) and the perform case now return nil
  explicitly via (do (emit-op em N) nil) instead of bare emit-op
  which transpiled to unit-returning OCaml.
- compile_match PREAMBLE: return Nil instead of unit (was ignore).
- Added init wrapper to PREAMBLE (needed by compile-module).
- All 41 compiler functions re-transpiled cleanly.

Render (bootstrap_render.py): re-transpiled, no changes.

Runtime: restored keyword_p predicate (needed by defio-parse-kwargs!
in the transpiled evaluator).

2608/2608 tests passing.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-03 21:26:20 +00:00
408eca1cb0 Set arity in native compiler and bytecode modules
Sx_compiler.compile/compile_module now emit arity (local slot count) in
the bytecode dict. MCP sx_build_bytecode serializes arity into .sxbc.json.
sx-platform.js passes arity through to K.loadModule(). Without this, the
VM allocated only 16 local slots per module frame.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-30 14:31:55 +00:00
a24efc1a00 Bootstrap SX bytecode compiler to native OCaml
Transpile lib/compiler.sx → hosts/ocaml/lib/sx_compiler.ml (42 functions).
The bytecode compiler now runs as native OCaml instead of interpreted SX,
eliminating the 24s JIT warm-up for compiler functions.

- bootstrap_compiler.py: transpiler script (like bootstrap.py for evaluator)
- sx_compiler.ml: 39KB native compiler (compile, compile-module, etc.)
- Bind compile/compile-module as native functions in setup_core_operations
- Add mutable_list to sx_runtime.ml (used by compiler pool)
- Add native parse function (wraps Sx_parser.parse_all)
- compile-match delegated via ref (uses letrec, transpiler can't handle)
- Compile all 23 bytecode modules successfully (was 0/23 due to WASM overflow)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-30 10:30:53 +00:00