On Mon, May 7, 2012 at 10:00 AM, Jakub Jelinek wrote:
>> > Shouldn't that be done only for generic tuning? If somebody uses
>> > -mtune=native, then emitting rep; bsf is overkill, the code is intended
>> > to be run on a CPU without (or with TARGET_BMI with) tzcnt insn support.
>>
>> Yes, this i
On Mon, May 07, 2012 at 09:36:38AM +0200, Uros Bizjak wrote:
> On Mon, May 7, 2012 at 7:43 AM, Jakub Jelinek wrote:
>
> >> Index: i386.md
> >> ===
> >> --- i386.md (revision 187217)
> >> +++ i386.md (working copy)
> >> @@ -12112,
On Mon, May 7, 2012 at 7:43 AM, Jakub Jelinek wrote:
>> Index: i386.md
>> ===
>> --- i386.md (revision 187217)
>> +++ i386.md (working copy)
>> @@ -12112,9 +12112,22 @@
>> (set (match_operand:SWI48 0 "register_operand" "=r")
On Mon, May 07, 2012 at 01:04:33AM +0200, Uros Bizjak wrote:
> Index: i386.md
> ===
> --- i386.md (revision 187217)
> +++ i386.md (working copy)
> @@ -12112,9 +12112,22 @@
> (set (match_operand:SWI48 0 "register_operand" "=r")
On Mon, Apr 30, 2012 at 10:09 AM, Uros Bizjak wrote:
> On Fri, Apr 27, 2012 at 3:30 PM, Paolo Bonzini wrote:
>> tzcnt is encoded as "rep;bsf" and unlike lzcnt is a drop-in replacement
>> if we don't care about the flags (it has the same semantics for non-zero
>> values).
>>
>> Since bsf is usuall
On Fri, Apr 27, 2012 at 3:30 PM, Paolo Bonzini wrote:
> tzcnt is encoded as "rep;bsf" and unlike lzcnt is a drop-in replacement
> if we don't care about the flags (it has the same semantics for non-zero
> values).
>
> Since bsf is usually slower, just emit tzcnt unconditionally. However,
> write
tzcnt is encoded as "rep;bsf" and unlike lzcnt is a drop-in replacement
if we don't care about the flags (it has the same semantics for non-zero
values).
Since bsf is usually slower, just emit tzcnt unconditionally. However,
write it as rep;bsf unless -mbmi is in use, to cater for old assemblers.
