bpo-47009: Let PRECALL_NO_KW_LIST_APPEND do its own POP_TOP

[sweeneyde]

Most code won't do y = L.append(x) or whatnot, so PRECALL_NO_KW_LIST_APPEND is almost always followed by POP_TOP. We can verify at specialization time.

This saves a Py_INCREF(Py_None), a SET_TOP(Py_None), and POP_TOP's Py_DECREF(POP()); DISPATCH();.

Some microbenchmarks:

frompyperfimportRunner, perf_counterdefbench_append(loops, length): src=list(map(float, range(length))) arr= [] t0=perf_counter() foriinrange(loops): arr.clear() forxinsrc: arr.append(x) returnperf_counter() -t0defbench_append_less_gc(loops, length): src=list(map(float, range(length))) out= [None] *loopst0=perf_counter() foriinrange(loops): arr= [] forxinsrc: arr.append(x) out[i] =arrreturnperf_counter() -t0runner=Runner() fornin [100, 1_000, 10_000, 100_000]: runner.bench_time_func(f"append {n}", bench_append, n, inner_loops=n) runner.bench_time_func(f"append-less-gc {n}", bench_append_less_gc, n, inner_loops=n)

From GCC, --enable-optimizations, --with-lto:

- append 100000: 14.9 ns +- 0.3 ns -> 13.3 ns +- 0.4 ns: 1.12x faster - append 10000: 15.1 ns +- 0.3 ns -> 13.6 ns +- 0.5 ns: 1.11x faster - append-less-gc 100000: 16.4 ns +- 0.5 ns -> 14.9 ns +- 0.4 ns: 1.10x faster - append 1000: 15.6 ns +- 0.3 ns -> 14.2 ns +- 0.3 ns: 1.09x faster - append 100: 18.9 ns +- 0.6 ns -> 17.3 ns +- 0.6 ns: 1.09x faster - append-less-gc 100: 27.4 ns +- 1.1 ns -> 25.2 ns +- 1.2 ns: 1.09x faster - append-less-gc 10000: 19.2 ns +- 0.3 ns -> 17.8 ns +- 0.2 ns: 1.08x faster - append-less-gc 1000: 22.0 ns +- 0.6 ns -> 20.8 ns +- 0.3 ns: 1.06x faster Geometric mean: 1.09x faster 

https://bugs.python.org/issue47009

[markshannon]

Looks good. I'm a bit wary of specialized superinstructions, but this seems solid.
I can imagine cases where list.append() wouldn't be followed by a POP_TOP, but they are contrived and highly unlikely.

[tiran]

The assert is failing on s390x Fedora buildbot https://buildbot.python.org/all/#/builders/232/builds/524

_bootstrap_python: Python/ceval.c:5045: _PyEval_EvalFrameDefault: Assertion `next_instr[-1] == POP_TOP' failed. make: *** [Makefile:1204: Python/frozen_modules/io.h] Aborted (core dumped) 

[markshannon]

Strange. The bytecode is exactly the same on all platforms.