[patch] lno branch merge - vectorizer patch #3
Dorit Naishlos
DORIT@il.ibm.com
Mon Sep 20 17:25:00 GMT 2004
Committed the attached fix to the new vectorizer testcases
vect-[74,79,80].c which caused a failure on x86.
dorit
(See attached file: testpatch.Sept20)
Dorit
Naishlos/Haifa/IBM To: Mark Mitchell <mark@codesourcery.com>
@IBMIL cc: Ayal Zaks/Haifa/IBM@IBMIL, gcc-patches@gcc.gnu.org, Ira
Sent by: Rosen/Haifa/IBM@IBMIL
gcc-patches-owner@ Subject: Re: [patch] lno branch merge - vectorizer patch #3
gcc.gnu.org
20/09/2004 10:20
Committed, along with new testcases vect-[62-69].c that were left out of
the patch, and a fix to vect-13.c that caused a failure on ix86 (attached).
(See attached file: patch.Sept20)
dorit
Mark Mitchell
<mark@codesourcery To: Ira
Rosen/Haifa/IBM@IBMIL
.com> cc:
gcc-patches@gcc.gnu.org, Dorit
Naishlos/Haifa/IBM@IBMIL, Ayal
Sent by: Zaks/Haifa/IBM@IBMIL
gcc-patches-owner@ Subject: Re: [patch] lno
branch merge - vectorizer patch #3
gcc.gnu.org
15/09/2004 20:54
Ira Rosen wrote:
>Hi,
>
>This patch adds vectorization support for additional forms of data
>references along with relevant new test-cases.
>We now support multidimensional arrays, pointer accesses with initial
value
>that is not restricted to an SSE_NAME (e.g: p=&a[16]-4B), and array
>accesses with initial value that is not restricted to INTEGER_CST (e.g:
>a[i+off]). Bootstrapped and passed the vectorizer testcases ans SPEC on
>ppc-darwin.
>
I think you should also run the entire testsuite with the vectorizer
hard-wired on. Have you done that?
+ #define VECT_SMODULO(x,y) ((x) % (y) < 0 ? ((x) % (y) + (y)) : (x) % (y))
This needs a comment.
! if (!evolution_function_is_constant_p (access_fn))
! {
! if (!vect_is_simple_iv_evolution (loop_containing_stmt (DR_STMT
(dr))->num,
! access_fn, &init, &step, true))
! {
! if (vect_debug_details (NULL))
That should be:
if (!e_f_i_c_p (access_fn)
&& !v_i_s_i_e (...)
&& vec_debug_details (NULL))
not three nested ifs. And vec_debug_details should probaby go at the
outside since it will usually be false, avoiding computations implied by
the other functions.
Patch OK with those changes.
--
Mark Mitchell
CodeSourcery, LLC
(916) 791-8304
mark@codesourcery.com
#### patch.Sept20 has been removed from this note on September 20, 2004 by
Dorit Naishlos
-------------- next part --------------
A non-text attachment was scrubbed...
Name: testpatch.Sept20
Type: application/octet-stream
Size: 4954 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20040920/b81940f8/attachment.obj>
More information about the Gcc-patches
mailing list