On Tue, Jul 14, 2015 at 3:06 AM, Andrew Pinski pins...@gmail.com wrote:
On Tue, Jul 14, 2015 at 1:18 AM, Kyrill Tkachov kyrylo.tkac...@arm.com
wrote:
Hi Segher,
On 14/07/15 01:38, Segher Boessenkool wrote:
On Mon, Jul 13, 2015 at 10:48:19AM +0100, Kyrill Tkachov wrote:
For the testcase
On 14/07/15 11:06, Andrew Pinski wrote:
On Tue, Jul 14, 2015 at 1:18 AM, Kyrill Tkachov kyrylo.tkac...@arm.com wrote:
Hi Segher,
On 14/07/15 01:38, Segher Boessenkool wrote:
On Mon, Jul 13, 2015 at 10:48:19AM +0100, Kyrill Tkachov wrote:
For the testcase in the patch we were generating an
On Jul 13, 2015, at 5:48 PM, Kyrill Tkachov kyrylo.tkac...@arm.com wrote:
Hi all,
For the testcase in the patch we were generating an extra neg instruction:
cmp w0, wzr
csneg w0, w0, w0, ge
neg w0, w0
ret
instead of the optimal:
cmp
Hi Segher,
On 14/07/15 01:38, Segher Boessenkool wrote:
On Mon, Jul 13, 2015 at 10:48:19AM +0100, Kyrill Tkachov wrote:
For the testcase in the patch we were generating an extra neg instruction:
cmp w0, wzr
csneg w0, w0, w0, ge
neg w0, w0
ret
On Tue, Jul 14, 2015 at 1:18 AM, Kyrill Tkachov kyrylo.tkac...@arm.com wrote:
Hi Segher,
On 14/07/15 01:38, Segher Boessenkool wrote:
On Mon, Jul 13, 2015 at 10:48:19AM +0100, Kyrill Tkachov wrote:
For the testcase in the patch we were generating an extra neg
instruction:
cmp
On Tue, Jul 14, 2015 at 3:13 AM, Kyrill Tkachov kyrylo.tkac...@arm.com wrote:
On 14/07/15 11:06, Andrew Pinski wrote:
On Tue, Jul 14, 2015 at 1:18 AM, Kyrill Tkachov kyrylo.tkac...@arm.com
wrote:
Hi Segher,
On 14/07/15 01:38, Segher Boessenkool wrote:
On Mon, Jul 13, 2015 at 10:48:19AM
On Tue, Jul 14, 2015 at 07:04:13PM +0800, pins...@gmail.com wrote:
Combine knows how to use define_split without being an insn.
Combine uses define_split in very different circumstances than it uses
define_insn. In this case, define_split will only do anything if the
nabs is combined from
On Tue, Jul 14, 2015 at 09:18:06AM +0100, Kyrill Tkachov wrote:
Before combine, you have two insns, a negation and an abs, so that is
not so very strange :-)
Oh, hrm, my aarch64 cross was three months old, and this now changed.
Or I messed up. Sorry for the noise.
It does look like the
On 14/07/15 11:40, Andrew Pinski wrote:
On Tue, Jul 14, 2015 at 3:13 AM, Kyrill Tkachov kyrylo.tkac...@arm.com wrote:
On 14/07/15 11:06, Andrew Pinski wrote:
On Tue, Jul 14, 2015 at 1:18 AM, Kyrill Tkachov kyrylo.tkac...@arm.com
wrote:
Hi Segher,
On 14/07/15 01:38, Segher Boessenkool wrote:
On 14/07/15 13:55, Segher Boessenkool wrote:
On Tue, Jul 14, 2015 at 09:18:06AM +0100, Kyrill Tkachov wrote:
Before combine, you have two insns, a negation and an abs, so that is
not so very strange :-)
Oh, hrm, my aarch64 cross was three months old, and this now changed.
Or I messed up.
Hi all,
For the testcase in the patch we were generating an extra neg instruction:
cmp w0, wzr
csneg w0, w0, w0, ge
neg w0, w0
ret
instead of the optimal:
cmp w0, wzr
csneg w0, w0, w0, lt
ret
The reason is that combine
On Mon, Jul 13, 2015 at 10:48:19AM +0100, Kyrill Tkachov wrote:
For the testcase in the patch we were generating an extra neg instruction:
cmp w0, wzr
csneg w0, w0, w0, ge
neg w0, w0
ret
instead of the optimal:
cmp w0, wzr
12 matches
Mail list logo