https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
Richard Biener changed:
What|Removed |Added
Assignee|unassigned at gcc dot gnu.org |rsandifo at gcc dot
gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
Richard Biener changed:
What|Removed |Added
Priority|P3 |P1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
rsandifo at gcc dot gnu.org changed:
What|Removed |Added
CC||rsandifo at gcc dot
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
--- Comment #14 from CVS Commits ---
The master branch has been updated by Richard Sandiford :
https://gcc.gnu.org/g:79ae13067f6afabcbae4784ada07dcbb7f00953e
commit r12-6594-g79ae13067f6afabcbae4784ada07dcbb7f00953e
Author: Richard Sandiford
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
--- Comment #13 from Tobias Burnus ---
(In reply to Tobias Burnus from comment #12)
> It looks as if it is the same as in comment 8. If so, it does not seem to
> work. Cf. comment 7.
I meant: cf. comment 11.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
--- Comment #12 from Tobias Burnus ---
(In reply to rdapp from comment #10)
> Created attachment 52192 [details]
> Proposed patch
>
> Could you try the proposed patch? Bootstraps cleanly for me and no
> regressions on Power or x86.
It looks
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
--- Comment #11 from Tobias Burnus ---
(In reply to rdapp from comment #8)
> The following should help:
> @@ -1170,6 +1170,9 @@ vect_verify_loop_lens (loop_vec_info loop_vinfo)
> + LOOP_VINFO_PARTIAL_LOAD_STORE_BIAS (loop_vinfo) =
> +
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
--- Comment #10 from rdapp at linux dot ibm.com ---
Created attachment 52192
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52192=edit
Proposed patch
Could you try the proposed patch? Bootstraps cleanly for me and no regressions
on Power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
ktkachov at gcc dot gnu.org changed:
What|Removed |Added
Status|UNCONFIRMED |NEW
Last reconfirmed|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
--- Comment #8 from rdapp at linux dot ibm.com ---
I think you're right. In one of the last iterations of the patch I moved
+ LOOP_VINFO_PARTIAL_LOAD_STORE_BIAS (loop_vinfo) = partial_load_bias;
after the unsupported check. It is now only
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
--- Comment #7 from Tobias Burnus ---
(In reply to rdapp from comment #4)
> This code path should only be active when the backend has len_load/len_store
> patterns.
> signed char partial_load_store_bias;
> is new in the "middle" of
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
Andrew Stubbs changed:
What|Removed |Added
CC||ams at gcc dot gnu.org
--- Comment #6
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
--- Comment #5 from Tobias Burnus ---
(In reply to Richard Biener from comment #2)
> OTOH, does amdgcn really require the load bias stuff?
According to Andrew, it doesn't - it uses maskstore.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
rdapp at linux dot ibm.com changed:
What|Removed |Added
CC||rdapp at linux dot ibm.com
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
--- Comment #3 from Tobias Burnus ---
Input file is:
https://sourceware.org/git/?p=newlib-cygwin.git;a=history;f=newlib/libc/stdlib/ecvtbuf.c;hb=HEAD
Compiled with amdgcn-amdhsa-gcc -g -O2 -ffunction-sections
I tried to reduce it – which
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
Richard Biener changed:
What|Removed |Added
CC||rdapp at gcc dot gnu.org
Target
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104026
--- Comment #1 from Tobias Burnus ---
Breakpoint 4, vect_set_loop_controls_directly (...)
at src/gcc-mainline/gcc/tree-vect-loop-manip.c:672
673 gassign *minus = gimple_build_assign (adjusted_len, PLUS_EXPR,
674
17 matches
Mail list logo