condition DCE of partially dead calls
Xinliang David Li
davidxl@google.com
Fri May 30 17:44:00 GMT 2008
Ralf Wildenhues wrote:
> Hello David,
>
> a couple of trivial typo nits:
>
> * Xinliang David Li wrote on Fri, May 30, 2008 at 09:11:40AM CEST:
>> Index: tree-call-cdce.c
>> ===================================================================
>> --- tree-call-cdce.c (revision 0)
>> +++ tree-call-cdce.c (revision 0)
>> @@ -0,0 +1,889 @@
>
>> + With this change, call to log (x) is effectively eliminated, as
>> + in majority of the cases, log won't be called with x out of
>> + range. The branch is totally predicatable, so the branch cost
>> + is low.
>
> s/predicatable/predictable/ ?
>
This was changed from predicatible orginally :)
Thanks for catching the typos.
David
>> +/* A helper function to generate condition
>> + code for the y argument in call pow (some_const, y).
>> + See candidate selection in check_pow. Since the
>> + candidates's base values have a limited range,
>
> s/candidates's/candidates'/
>
>> + the guarded code generated for y are simple:
>
>> + See C99 standard, 7.12.7.4:2, for description of pow (x, y).
>> + The precice condition for domain errors are complex. In this
>
> s/precice/precise/
>
>> + implementation, a simplified (but conservative) valid domain
>> + for x and y are used: x is positive to avoid dom errors, while
>> + y is smaller than a upper bound (depending on x) to avoid range
>> + errors. Runtime code is generated to check x (if not constant)
>
>> +/* Propability of the branch (to the call) is taken. */
>
> s/Propability/Probability/
>
>> +#define ERR_PROB 0.01
>
>> + /* Now it is time to insert the first condtional expression
>
> s/condtional/conditional/
>
>> + into bi_call_bb and split this bb so that bi_call is
>> + shrink-wrapped. */
>
> Cheers,
> Ralf
More information about the Gcc-patches
mailing list