René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Optimize int add/sub for wide exact ints","articleBody":"CPython has a fast path for compact integers in binary add/sub, but wide exact ints still go through the generic long arithmetic path even when both operands fit in int64_t.\n\nThis issue proposes adding a separate fast path for exact PyLong operands that fit in signed 64-bit integers, while preserving the existing compact-int path.\n\nSuggested implementation:\n- Keep the current compact-int specialization unchanged.\n- Add a separate wide-int path for exact ints that fit in int64_t.\n- Preserve current behavior for overflow, subclasses, and other non-exact-int cases.\n\nMotivation:\n- Improve performance for wide integer add/sub without affecting the common compact-int hot path.\n- Avoid adding new opcodes in the compact-int path.\n- Fit within the current interpreter and specialization structure.\n\nBenchmark evidence:\n- I prototyped this locally with a benchmark covering compact and wide add/sub cases.\n- Wide cases improved substantially, while compact cases remained effectively flat.\n- Representative interpreter-only results with JIT disabled:\n  - add_wide: about 25% faster\n  - sub_wide: about 35% faster\n  - add_compact/sub_compact: effectively unchanged\n\nBenchmark script used locally:\n```python\n\"\"\"Microbenchmark compact vs wide int add/sub with pyperf.\n\nUse this with PYTHON_JIT=0 and -S if you want a stable interpreter-only run:\n\n    PYTHON_JIT=0 ./python.exe -S Tools/scripts/bench_wide_int_pyperf.py\n\"\"\"\n\nfrom __future__ import annotations\n\nimport pyperf\n\n\ndef bench_add_compact() -> int:\n    a = 1\n    b = 2\n    return a + b\n\n\ndef bench_add_wide() -> int:\n    a = 10_000_000_000\n    b = 1\n    return a + b\n\n\ndef bench_sub_compact() -> int:\n    a = 1\n    b = 2\n    return a - b\n\n\ndef bench_sub_wide() -> int:\n    a = 10_000_000_000\n    b = 1\n    return a - b\n\n\ndef main() -> None:\n    runner = pyperf.Runner()\n    runner.bench_func(\"add_compact\", bench_add_compact)\n    runner.bench_func(\"add_wide\", bench_add_wide)\n    runner.bench_func(\"sub_compact\", bench_sub_compact)\n    runner.bench_func(\"sub_wide\", bench_sub_wide)\n\n\nif __name__ == \"__main__\":\n    main()\n```\n\n\n\u003c!-- gh-linked-prs -->\n### Linked PRs\n* gh-151290\n* gh-159688\n\u003c!-- /gh-linked-prs -->\n","author":{"url":"https://github.com/KRRT7","@type":"Person","name":"KRRT7"},"datePublished":"2026-06-10T22:17:08Z","url":"https://github.com/python/cpython/issues/151289"}