Re: [x86] Don't use builtins for unaligned load/store

2016-08-29 Thread Marc Glisse
On Mon, 29 Aug 2016, Uros Bizjak wrote: On Mon, Aug 29, 2016 at 5:22 PM, Marc Glisse wrote: On Mon, 29 Aug 2016, Kirill Yukhin wrote: On 29.08.2016 14:58, Marc Glisse wrote: this patch gets rid of a few more builtins (well, I actually kept them, since Ada users may still need them). I had

Re: [x86] Don't use builtins for unaligned load/store

2016-08-29 Thread Uros Bizjak
On Mon, Aug 29, 2016 at 5:22 PM, Marc Glisse wrote: > On Mon, 29 Aug 2016, Kirill Yukhin wrote: > >> On 29.08.2016 14:58, Marc Glisse wrote: >>> >>> this patch gets rid of a few more builtins (well, I actually kept them, >>> since Ada users may still need them). I had to tweak the flags for >>> pr

Re: [x86] Don't use builtins for unaligned load/store

2016-08-29 Thread Marc Glisse
On Mon, 29 Aug 2016, Kirill Yukhin wrote: On 29.08.2016 14:58, Marc Glisse wrote: this patch gets rid of a few more builtins (well, I actually kept them, since Ada users may still need them). I had to tweak the flags for pr59539-2.c, otherwise the compiler thinks it is more efficient to split

Re: [x86] Don't use builtins for unaligned load/store

2016-08-29 Thread Kirill Yukhin
Hello, On 29.08.2016 14:58, Marc Glisse wrote: Hello, this patch gets rid of a few more builtins (well, I actually kept them, since Ada users may still need them). I had to tweak the flags for pr59539-2.c, otherwise the compiler thinks it is more efficient to split the loads, reading 128 bits

[x86] Don't use builtins for unaligned load/store

2016-08-29 Thread Marc Glisse
Hello, this patch gets rid of a few more builtins (well, I actually kept them, since Ada users may still need them). I had to tweak the flags for pr59539-2.c, otherwise the compiler thinks it is more efficient to split the loads, reading 128 bits at a time. This still breaks one testcase: avx