Re: [PATCH] rs6000: Revamp rotate-and-mask and insert

2015-07-20 Thread David Edelsohn
On Sun, Jul 12, 2015 at 1:08 PM, Segher Boessenkool wrote: > This rewrites all the rotate-and-mask and insert patterns. > > The goals are to have simpler, shorter, less error-prone code (with much > fewer machine description patterns), as well as to get better optimised > machine code. > > All "ma

Re: [PATCH] rs6000: Revamp rotate-and-mask and insert

2015-07-15 Thread Segher Boessenkool
On Wed, Jul 15, 2015 at 12:10:51PM +0930, Alan Modra wrote: > There are one or two regressions related to a TODO that Segher added. > The following produces poorer code than mainline. Yes, that's why it says TODO :-) > Also, rs6000.md patterns uses SImode for the rotate/shift count. > Segher has

Re: [PATCH] rs6000: Revamp rotate-and-mask and insert

2015-07-14 Thread Alan Modra
On Sun, Jul 12, 2015 at 04:18:31PM -0400, David Edelsohn wrote: > On Sun, Jul 12, 2015 at 1:08 PM, Segher Boessenkool > wrote: > > This rewrites all the rotate-and-mask and insert patterns. > > This is great! I'm glad that you completed this feature. Compared to mainline the results do look goo

Re: [PATCH] rs6000: Revamp rotate-and-mask and insert

2015-07-12 Thread David Edelsohn
On Sun, Jul 12, 2015 at 1:08 PM, Segher Boessenkool wrote: > This rewrites all the rotate-and-mask and insert patterns. > > The goals are to have simpler, shorter, less error-prone code (with much > fewer machine description patterns), as well as to get better optimised > machine code. > > All "ma

[PATCH] rs6000: Revamp rotate-and-mask and insert

2015-07-12 Thread Segher Boessenkool
This rewrites all the rotate-and-mask and insert patterns. The goals are to have simpler, shorter, less error-prone code (with much fewer machine description patterns), as well as to get better optimised machine code. All "mask only" insns are now handled by a single pattern; all rotate (or shift