July 28, 2013
http://d.puremagic.com/issues/show_bug.cgi?id=6829



--- Comment #40 from bearophile_hugs@eml.cc 2013-07-28 10:19:00 PDT ---
(In reply to comment #13)

> (Perhaps here there's material for a small enhancement request for LLVM.)

Here it is, in the LLVM Bugzilla: http://llvm.org/bugs/show_bug.cgi?id=16726

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
August 13, 2013
http://d.puremagic.com/issues/show_bug.cgi?id=6829



--- Comment #41 from hsteoh@quickfur.ath.cx 2013-08-13 09:42:34 PDT ---
Related: issue #1116.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
August 13, 2013
http://d.puremagic.com/issues/show_bug.cgi?id=6829


Temtaime <temtaime@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |temtaime@gmail.com


--- Comment #42 from Temtaime <temtaime@gmail.com> 2013-08-13 10:47:52 PDT ---
Guys, please, fix the bugs, not improve performance of DMD's backend. For that purpose there are LDC and GDC.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
August 13, 2013
http://d.puremagic.com/issues/show_bug.cgi?id=6829



--- Comment #43 from bearophile_hugs@eml.cc 2013-08-13 11:20:41 PDT ---
(In reply to comment #42)
> Guys, please, fix the bugs, not improve performance of DMD's backend. For that purpose there are LDC and GDC.

Walter is free to do what he likes with his time, including enhancing the DMD back-end. And currently LDC2 back-end is not optimizing the rotations well enough. I now think that having two rotation intrinsics is a good solution.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
August 13, 2013
http://d.puremagic.com/issues/show_bug.cgi?id=6829



--- Comment #44 from Iain Buclaw <ibuclaw@ubuntu.com> 2013-08-13 12:05:05 PDT ---
(In reply to comment #42)
> Guys, please, fix the bugs, not improve performance of DMD's backend. For that purpose there are LDC and GDC.

Intrinsics benefits GDC code generation also...

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
September 09, 2013
http://d.puremagic.com/issues/show_bug.cgi?id=6829


kai@redstar.de changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |kai@redstar.de


--- Comment #45 from kai@redstar.de 2013-09-09 10:58:01 PDT ---
(In reply to comment #43)
> And currently LDC2 back-end is not optimizing the rotations well enough.

I had a closer look to this. The root cause is the uint nBits declaration. Because of this the x value is promoted to uint if T is ubyte or ushort (according to TDPL, 2.3.3). In this case LLVM is not smart enough to deduce that the zero extension is not necessary. If both types are ubyte or ushort, then the rolb/rolw and rorb/rorw instructions are generated.

I try to make LLVM smarter. :-)

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
November 02, 2013
http://d.puremagic.com/issues/show_bug.cgi?id=6829


Martin Nowak <code@dawg.eu> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |luis@luismarques.eu


--- Comment #46 from Martin Nowak <code@dawg.eu> 2013-11-01 19:06:39 PDT ---
*** Issue 1116 has been marked as a duplicate of this issue. ***

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
1 2 3 4 5
Next ›   Last »