andrewthad
commented
4 years ago Uh-oh. This is correct, but the performance is awful. The problem is that, rather than using a few registers like I would expect, GHC uses all the registers and then spills onto the stack. Here's the cmm:
[reserved#_entry() // [R2]
{ info_tbls: [(c1aKA,
label: reserved#_info
rep: HeapRep static { Fun {arity: 1 fun_type: ArgSpec 4} }
srt: Nothing)]
stack_info: arg_space: 8 updfr_space: Just 8
}
{offset
c1aKA:
(_c1aKF::I64) = call MO_Ctz W64(R2 & 4294967040 ^ 3405803776);
(_c1aKS::I64) = call MO_Ctz W64(R2 & 4294967040 ^ 3325256704);
(_c1aL5::I64) = call MO_Ctz W64(R2 & 4294967040 ^ 3227017984);
(_c1aLi::I64) = call MO_Ctz W64(R2 & 4294967040 ^ 3221225984);
(_c1aLv::I64) = call MO_Ctz W64(R2 & 4294967040 ^ 3221225472);
(_c1aLI::I64) = call MO_Ctz W64(R2 & 4294901760 ^ 3232235520);
(_c1aLV::I64) = call MO_Ctz W64(R2 & 4294901760 ^ 2851995648);
(_c1aM8::I64) = call MO_Ctz W64(R2 & 4294836224 ^ 3323068416);
(_c1aMl::I64) = call MO_Ctz W64(R2 & 4293918720 ^ 2886729728);
(_c1aMy::I64) = call MO_Ctz W64(R2 & 4290772992 ^ 1681915904);
(_c1aML::I64) = call MO_Ctz W64(R2 & 4278190080 ^ 2130706432);
(_c1aMV::I64) = call MO_Ctz W64(R2 & 4278190080);
(_c1aN8::I64) = call MO_Ctz W64(R2 & 4278190080 ^ 167772160);
(_c1aNl::I64) = call MO_Ctz W64(R2 & 4026531840 ^ 4026531840);
(_c1aNy::I64) = call MO_Ctz W64(R2 & 4026531840 ^ 3758096384);
R1 = _c1aNy::I64 | _c1aNl::I64 | _c1aN8::I64 | _c1aMV::I64 | _c1aML::I64 | _c1aMy::I64 | _c1aMl::I64 | _c1aM8::I64 | _c1aLV::I64 | _c1aLI::I64 | _c1aLv::I64 | _c1aLi::I64 | _c1aL5::I64 | _c1aKS::I64 | _c1aKF::I64 >> 5;
call (P64[Sp])(R1) args: 8, res: 0, upd: 8;
}
},
section ""data" . reserved#_closure" {
reserved#_closure:
const reserved#_info;
}]And here is the assembly, truncated after the stack spilling becomes clear:
_c1aKA:
movl $3405803776,%eax
movl $4294967040,%ebx
movq %r14,%rcx
andq %rbx,%rcx
xorq %rax,%rcx
bsf %rcx,%rax
movl $64,%ebx
cmovne %rax,%rbx
movl $3325256704,%eax
movl $4294967040,%ecx
movq %r14,%rdx
andq %rcx,%rdx
xorq %rax,%rdx
bsf %rdx,%rax
movl $64,%ecx
cmovne %rax,%rcx
movl $3227017984,%eax
movl $4294967040,%edx
movq %r14,%rsi
andq %rdx,%rsi
xorq %rax,%rsi
bsf %rsi,%rax
movl $64,%edx
cmovne %rax,%rdx
movl $3221225984,%eax
movl $4294967040,%esi
movq %r14,%rdi
andq %rsi,%rdi
xorq %rax,%rdi
bsf %rdi,%rax
movl $64,%esi
cmovne %rax,%rsi
movl $3221225472,%eax
movl $4294967040,%edi
movq %r14,%r8
andq %rdi,%r8
xorq %rax,%r8
bsf %r8,%rax
movl $64,%edi
cmovne %rax,%rdi
movl $3232235520,%eax
movl $4294901760,%r8d
movq %r14,%r9
andq %r8,%r9
xorq %rax,%r9
bsf %r9,%rax
movl $64,%r8d
cmovne %rax,%r8
movl $2851995648,%eax
movl $4294901760,%r9d
movq %r14,%r10
andq %r9,%r10
xorq %rax,%r10
bsf %r10,%rax
movl $64,%r9d
cmovne %rax,%r9
movl $3323068416,%eax
movl $4294836224,%r10d
movq %r14,%r11
andq %r10,%r11
xorq %rax,%r11
bsf %r11,%rax
movl $64,%r10d
cmovne %rax,%r10
movl $2886729728,%eax
movl $4293918720,%r11d
movq %rbx,64(%rsp)
movq %r14,%rbx
andq %r11,%rbx
xorq %rax,%rbx
bsf %rbx,%rax
movl $64,%ebx
cmovne %rax,%rbx
movl $4290772992,%eax
movq %r14,%r11
andq %rax,%r11
xorq $1681915904,%r11
bsf %r11,%rax
movl $64,%r11d
cmovne %rax,%r11
movl $4278190080,%eax
movq %rcx,72(%rsp)
movq %r14,%rcx
andq %rax,%rcx
xorq $2130706432,%rcx
bsf %rcx,%rax
movl $64,%ecx
cmovne %rax,%rcx
movl $4278190080,%eax
movq %rdx,80(%rsp)
movq %r14,%rdx
andq %rax,%rdx
bsf %rdx,%rax
movl $64,%edx
cmovne %rax,%rdx
movl $4278190080,%eax
movq %rsi,88(%rsp)
movq %r14,%rsi
andq %rax,%rsi
xorq $167772160,%rsi
bsf %rsi,%rax
movl $64,%esi
cmovne %rax,%rsi
movl $4026531840,%eax
movq %rdi,96(%rsp)
movl $4026531840,%edi
movq %r8,104(%rsp)
movq %r14,%r8
andq %rdi,%r8
...Darn.
chessai
Add tests and benchmarks reserved range testing, remove check for 255.255.255.255 since it is redundant.
This will resolve https://github.com/andrewthad/haskell-ip/issues/60 when completed.