Re: [PATCH] Avoid false vector mask conversion

2015-12-10 Thread Richard Biener
On Fri, Dec 4, 2015 at 4:00 PM, Ilya Enkovich wrote: > On 02 Dec 16:27, Richard Biener wrote: >> On Wed, Dec 2, 2015 at 4:24 PM, Ilya Enkovich wrote: >> > >> > The problem is that conversion is supposed to be handled by >> > vectorizable_conversion, >> > but it fails to because it is not actually

Re: [PATCH] Avoid false vector mask conversion

2015-12-04 Thread Ilya Enkovich
On 02 Dec 16:27, Richard Biener wrote: > On Wed, Dec 2, 2015 at 4:24 PM, Ilya Enkovich wrote: > > > > The problem is that conversion is supposed to be handled by > > vectorizable_conversion, > > but it fails to because it is not actually a conversion. I suppose it > > may be handled > > in vectori

Re: [PATCH] Avoid false vector mask conversion

2015-12-02 Thread Richard Biener
On Wed, Dec 2, 2015 at 4:24 PM, Ilya Enkovich wrote: > 2015-12-02 17:52 GMT+03:00 Richard Biener : >> On Thu, Nov 12, 2015 at 5:08 PM, Ilya Enkovich >> wrote: >>> Hi, >>> >>> When we use LTO for fortran we may have a mix 32bit and 1bit scalar >>> booleans. It means we may have conversion of one

Re: [PATCH] Avoid false vector mask conversion

2015-12-02 Thread Ilya Enkovich
2015-12-02 17:52 GMT+03:00 Richard Biener : > On Thu, Nov 12, 2015 at 5:08 PM, Ilya Enkovich wrote: >> Hi, >> >> When we use LTO for fortran we may have a mix 32bit and 1bit scalar >> booleans. It means we may have conversion of one scalar type to another >> which confuses vectorizer because val

Re: [PATCH] Avoid false vector mask conversion

2015-12-02 Thread Richard Biener
On Thu, Nov 12, 2015 at 5:08 PM, Ilya Enkovich wrote: > Hi, > > When we use LTO for fortran we may have a mix 32bit and 1bit scalar booleans. > It means we may have conversion of one scalar type to another which confuses > vectorizer because values with different scalar boolean type may get the

Re: [PATCH] Avoid false vector mask conversion

2015-12-02 Thread Ilya Enkovich
Ping 2015-11-23 16:05 GMT+03:00 Ilya Enkovich : > Ping > > 2015-11-13 16:17 GMT+03:00 Ilya Enkovich : >> 2015-11-13 13:03 GMT+03:00 Richard Biener : >>> On Thu, Nov 12, 2015 at 5:08 PM, Ilya Enkovich >>> wrote: Hi, When we use LTO for fortran we may have a mix 32bit and 1bit scala

Re: [PATCH] Avoid false vector mask conversion

2015-11-23 Thread Ilya Enkovich
Ping 2015-11-13 16:17 GMT+03:00 Ilya Enkovich : > 2015-11-13 13:03 GMT+03:00 Richard Biener : >> On Thu, Nov 12, 2015 at 5:08 PM, Ilya Enkovich >> wrote: >>> Hi, >>> >>> When we use LTO for fortran we may have a mix 32bit and 1bit scalar >>> booleans. It means we may have conversion of one scal

Re: [PATCH] Avoid false vector mask conversion

2015-11-13 Thread Ilya Enkovich
2015-11-13 13:03 GMT+03:00 Richard Biener : > On Thu, Nov 12, 2015 at 5:08 PM, Ilya Enkovich wrote: >> Hi, >> >> When we use LTO for fortran we may have a mix 32bit and 1bit scalar >> booleans. It means we may have conversion of one scalar type to another >> which confuses vectorizer because val

Re: [PATCH] Avoid false vector mask conversion

2015-11-13 Thread Richard Biener
On Thu, Nov 12, 2015 at 5:08 PM, Ilya Enkovich wrote: > Hi, > > When we use LTO for fortran we may have a mix 32bit and 1bit scalar booleans. > It means we may have conversion of one scalar type to another which confuses > vectorizer because values with different scalar boolean type may get the

[PATCH] Avoid false vector mask conversion

2015-11-12 Thread Ilya Enkovich
Hi, When we use LTO for fortran we may have a mix 32bit and 1bit scalar booleans. It means we may have conversion of one scalar type to another which confuses vectorizer because values with different scalar boolean type may get the same vectype. This patch transforms such conversions into comp