FWIW, here's the inner loop of the accumulating parameter factorial
compiled with yesterday's HEAD on x86_64, via C:
Fac_zdwfac_info:
movq %rsi, %rax
testq %rsi, %rsi
jne .L4
movq %rdi, %rbx
jmp *(%rbp)
.L4:
leaq -1(%rsi), %rsi
imulq %rax, %rdi
jmp Fac_zdwfac_info
It's not perfect, but note the absence of memory operations.
The NCG version is similar, but has a couple of extra reg-to-reg moves
(we need to beef up the NCG optimisations a bit).
Cheers,
Simon
_______________________________________________
Glasgow-haskell-users mailing list
Glasgow-haskell-users@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
