Re: [RFC] Generalize formation of lane-reducing ops in loop reduction

2024-08-21 Thread Feng Xue OS
>> >> >> 1. Background >> >> >> >> For loop reduction of accumulating result of a widening operation, the >> >> preferred pattern is lane-reducing operation, if supported by target. >> >> Because >> >> this kind of operation need not preserve intermediate results of widening >> >> operation, and o

Re: [RFC] Generalize formation of lane-reducing ops in loop reduction

2024-08-08 Thread Richard Biener
On Sat, Aug 3, 2024 at 2:42 PM Feng Xue OS wrote: > > >> 1. Background > >> > >> For loop reduction of accumulating result of a widening operation, the > >> preferred pattern is lane-reducing operation, if supported by target. > >> Because > >> this kind of operation need not preserve intermediat

Re: [RFC] Generalize formation of lane-reducing ops in loop reduction

2024-08-03 Thread Feng Xue OS
>> 1. Background >> >> For loop reduction of accumulating result of a widening operation, the >> preferred pattern is lane-reducing operation, if supported by target. Because >> this kind of operation need not preserve intermediate results of widening >> operation, and only produces reduced amount

Re: [RFC] Generalize formation of lane-reducing ops in loop reduction

2024-07-26 Thread Richard Biener
On Sun, Jul 21, 2024 at 11:12 AM Feng Xue OS wrote: > > Hi, > > I composed some patches to generalize lane-reducing (dot-product is a > typical representative) pattern recognition, and prepared a RFC document so > as to help > review. The original intention was to make a complete solution for