Am 13.07.2013 12:35, schrieb bearophile:
The resulting X86 asm:
__D4test12addParticlesFNaNbxS4test8ParticlexS4test8ParticleZS4test8Particle:
pushl %ebp
movl %esp, %ebp
andl $-16, %esp
subl $16, %esp
movaps 40(%ebp), %xmm0
movaps 56(%ebp), %xmm1
addps 8(%ebp), %xmm0
addps 24(%ebp), %xmm1
movups %xmm1, 16(%eax)
movups %xmm0, (%eax)
I think it would be more important if dmd would actually use the XMM
registers correctly for computations. As you can see from the
disassembly dmd generates code that always adds/moves from/to memory and
does not stay within the registers at all.
http://d.puremagic.com/issues/show_bug.cgi?id=10226
Until dmd uses the XMM registers correctly it doesn't make much sense to
add a special calling convetion for this purpose.
Kind Regards
Benjamin Thaut
