https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96633
Bug ID: 96633
Summary: missed optimization?
Product: gcc
Version: 10.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: nathan at gcc dot gnu.org
Target Milestone: ---
Matt Godbolt's https://queue.acm.org/detail.cfm?id=3372264
has an example of optimizing on amd64:
bool isWhitespace(char c)
{
return c == ' '
|| c == '\r'
|| c == '\n'
|| c == '\t';
}
GCC generates:
xorl %eax, %eax
cmpb $32, %dil
ja .L1
movabsq $4294977024, %rax
movl %edi, %ecx
shrq %cl, %rax
andl $1, %eax
.L1:
ret
following an amazing comment on the ML, I've determined the following is abot
12% faster (and shorter too):
movabsq $4294977024, %rax
movl %edi, %ecx
shrq %cl, %rax
shr $6, %ecx
andl $1, %eax
shrq %cl, %rax
ret
We're dealing with chars in the range [-128,128), and x86's shift operator only
considers the bottom 6 bits.
