#13025: Type family reduction irregularity (change from 7.10.3 to 8.0.1) -------------------------------------+------------------------------------- Reporter: acowley | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.1 Resolution: | Keywords: TypeFamilies Operating System: Unknown/Multiple | Architecture: Type of failure: Runtime | Unknown/Multiple performance bug | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by RyanGlScott): There does appear to be a difference in the `bench-builder-all` benchmark suite. The full results are available at https://gist.github.com/RyanGlScott/386da12bc71fd2fadae2362b257c41bc. The most noticeable differences come in the `vinyl` and `vinyl-plus` benchmarks. On GHC HEAD, I get: {{{ benchmarking vinyl time 106.9 µs (106.7 µs .. 107.0 µs) 1.000 R² (1.000 R² .. 1.000 R²) mean 106.7 µs (106.5 µs .. 106.9 µs) std dev 653.4 ns (506.9 ns .. 847.7 ns) benchmarking vinyl-lens time 83.98 µs (83.97 µs .. 84.00 µs) 1.000 R² (1.000 R² .. 1.000 R²) mean 83.97 µs (83.96 µs .. 83.98 µs) std dev 40.09 ns (33.48 ns .. 47.87 ns) }}} After reverting 1722fa106e10e63160bb2322e2ccb830fd5b9ab3, you get: {{{ benchmarking vinyl time 62.27 µs (62.25 µs .. 62.28 µs) 1.000 R² (1.000 R² .. 1.000 R²) mean 62.28 µs (62.26 µs .. 62.30 µs) std dev 58.65 ns (47.95 ns .. 71.78 ns) benchmarking vinyl-lens time 62.07 µs (62.01 µs .. 62.13 µs) 1.000 R² (1.000 R² .. 1.000 R²) mean 62.02 µs (61.98 µs .. 62.05 µs) std dev 122.0 ns (106.2 ns .. 144.8 ns) }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13025#comment:8 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler