On Sat, May 12, 2012 at 11:36 AM, Maciej W. Rozycki
wrote:
> On Sun, 6 May 2012, Andrew Pinski wrote:
>
>> > For the record: MIPS processors that implement CLZ/CLO (for some reason
>> > CTZ/CTO haven't been added to the architecture, but these operations can
>> > be cheaply transformed into CLZ/C
On Sun, 6 May 2012, Andrew Pinski wrote:
> > For the record: MIPS processors that implement CLZ/CLO (for some reason
> > CTZ/CTO haven't been added to the architecture, but these operations can
> > be cheaply transformed into CLZ/CLO) generally have a dedicated unit that
> > causes no pipeline st
On Sat, May 5, 2012 at 11:52 PM, Maciej W. Rozycki wrote:
> For the record: MIPS processors that implement CLZ/CLO (for some reason
> CTZ/CTO haven't been added to the architecture, but these operations can
> be cheaply transformed into CLZ/CLO) generally have a dedicated unit that
> causes no pi
On Fri, 27 Apr 2012, Paolo Bonzini wrote:
> > What about cost considerations? We only seem to have the general
> > "branches are expensive" metric - but ctz/clz may be prohibitely expensive
> > themselves, no?
>
> Yeah, that's a general problem with this kind of tricks. In general
> however clz
On 05/03/2012 08:23 AM, Paolo Bonzini wrote:
Yeah, not as I posted... of course config/mips/mips.h needs change, but
do you disagree that all targets_do_ have the same value at zero for
the optab as they have for RTL?
I don't agree without a port-by-port review, no.
If, as you suggest, we ge
Il 03/05/2012 17:10, Richard Henderson ha scritto:
>>>
>>>
>>> ... except that no target sets the macros to 2, and all of them could
>>> (as far as I could see). Looks like the code trumps the documentation;
>>> how does this look?
>
> "No target sets to 2"? You mean like mips? You forgot to lo
On 05/03/2012 03:23 AM, Paolo Bonzini wrote:
- ctzsi4 expands to bitrev + clz:SI, so it will return 32 too. ctz:SI
will never appear at all in the RTX stream, so you can define its value
at zero to be whatever you want. Hence the correct valid of
CTZ_DEFINED_VALUE_AT_ZERO is also 2.
That's for
On Fri, Apr 27, 2012 at 2:43 PM, Paolo Bonzini wrote:
Interesting, first time I hear about this...
... except that no target sets the macros to 2, and all of them could
(as far as I could see). Looks like the code trumps the documentation;
how does this look?
"No target sets to 2"? You mean
Il 03/05/2012 11:57, Richard Guenther ha scritto:
> Hmm, I'll let target maintainers have a look at this - it's non-obvious
> for arm (the only case I looked at) at least. My guess is that the
> difference was to allow __builtin_ctz to be "optimized" and CTZ
> to be always well-defined (which inci
On Fri, Apr 27, 2012 at 2:43 PM, Paolo Bonzini wrote:
> Il 27/04/2012 13:16, Richard Guenther ha scritto:
>> In optabs.c we compare the CTZ_DEFINED_VALUE_AT_ZERO against two,
>> is != 0 really what you want here? The docs suggest to me
>> that as you are using the optab below you should compare a
Il 27/04/2012 13:16, Richard Guenther ha scritto:
> In optabs.c we compare the CTZ_DEFINED_VALUE_AT_ZERO against two,
> is != 0 really what you want here? The docs suggest to me
> that as you are using the optab below you should compare against two, too.
Interesting, first time I hear about this.
On Fri, 27 Apr 2012, Paolo Bonzini wrote:
> If the value at zero is outside the range [0, GET_MODE_BITSIZE (mode)),
> "A != 0" and "A == 0" can be compiled to clz/ctz followed by a subtraction
> or one's complement (only for A != 0) and a shift. This trick can be
> used effectively on PPC (though
> Index: config/rs6000/rs6000.md
> ===
> --- config/rs6000/rs6000.md (revisione 186859)
> +++ config/rs6000/rs6000.md (copia locale)
> @@ -2129,7 +2129,7 @@
> (define_expand "abssi2"
>[(set (match_operand:SI 0 "gpc_reg_operan
If the value at zero is outside the range [0, GET_MODE_BITSIZE (mode)),
"A != 0" and "A == 0" can be compiled to clz/ctz followed by a subtraction
or one's complement (only for A != 0) and a shift. This trick can be
used effectively on PPC (though there are other problems in the machine
descriptio
14 matches
Mail list logo
