forked from lix-project/lix
81e7c40264
We now parse function applications as a vector of arguments rather
than as a chain of binary applications, e.g. 'substring 1 2 "foo"' is
parsed as
ExprCall { .fun = <substring>, .args = [ <1>, <2>, <"foo"> ] }
rather than
ExprApp (ExprApp (ExprApp <substring> <1>) <2>) <"foo">
This allows primops to be called immediately (if enough arguments are
supplied) without having to allocate intermediate tPrimOpApp values.
On
$ nix-instantiate --dry-run '<nixpkgs/nixos/release-combined.nix>' -A nixos.tests.simple.x86_64-linux
this gives a substantial performance improvement:
user CPU time: median = 0.9209 mean = 0.9218 stddev = 0.0073 min = 0.9086 max = 0.9340 [rejected, p=0.00000, Δ=-0.21433±0.00677]
elapsed time: median = 1.0585 mean = 1.0584 stddev = 0.0024 min = 1.0523 max = 1.0623 [rejected, p=0.00000, Δ=-0.20594±0.00236]
because it reduces the number of tPrimOpApp allocations from 551990 to
42534 (i.e. only small minority of primop calls are partially
applied) which in turn reduces time spent in the garbage collector.
|
||
|---|---|---|
| .. | ||
| flake | ||
| primops | ||
| attr-path.cc | ||
| attr-path.hh | ||
| attr-set.cc | ||
| attr-set.hh | ||
| common-eval-args.cc | ||
| common-eval-args.hh | ||
| eval-cache.cc | ||
| eval-cache.hh | ||
| eval-inline.hh | ||
| eval.cc | ||
| eval.hh | ||
| fetchurl.nix | ||
| function-trace.cc | ||
| function-trace.hh | ||
| get-drvs.cc | ||
| get-drvs.hh | ||
| imported-drv-to-derivation.nix | ||
| json-to-value.cc | ||
| json-to-value.hh | ||
| lexer.l | ||
| local.mk | ||
| nix-expr.pc.in | ||
| nixexpr.cc | ||
| nixexpr.hh | ||
| parser.y | ||
| primops.cc | ||
| primops.hh | ||
| symbol-table.hh | ||
| value-to-json.cc | ||
| value-to-json.hh | ||
| value-to-xml.cc | ||
| value-to-xml.hh | ||
| value.hh | ||