--- Additional Comments From kazu at cs dot umass dot edu 2004-11-16 17:48
---
Subject: Re: Micro-optimize tree_code_class
Hi Steven,
> How about making the classes bitmasks?
Yes, I came across this idea. There are a lot of predicates that can
simply use bitmasks.
Kazu Hir
--- Additional Comments From kazu at cs dot umass dot edu 2004-11-16 17:08
---
The binary search converges between 2004-10-27 17:40GMT and 2004-10-27 17:50GMT,
so this patch is likely to have caused the segfault.
2004-10-27 Daniel Berlin <[EMAIL PROTECTED]>
Fix P
--- Additional Comments From kazu at cs dot umass dot edu 2004-11-16 15:06
---
Reduced a little more:
/* ./cc1 -O1 pr18519.c */
int baz (int *);
void
foo (int *uc)
{
int *invalidp
Severity: enhancement
Priority: P2
Component: tree-optimization
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: kazu at cs dot umass dot edu
CC: gcc-bugs at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18487
--- Additional Comments From kazu at cs dot umass dot edu 2004-11-14 17:30
---
*** Bug 18475 has been marked as a duplicate of this bug. ***
--
What|Removed |Added
Bug 18478
--- Additional Comments From kazu at cs dot umass dot edu 2004-11-14 17:30
---
The underlying problem is the same as that of PR 18478.
Basically, SWITCH_EXPR is incorrectly duplicated when a basic block
is duplicated.
*** This bug has been marked as a duplicate of 18478
--- Additional Comments From kazu at cs dot umass dot edu 2004-11-14 17:26
---
See PR 18478 for patch.
--
What|Removed |Added
Keywords
--- Additional Comments From kazu at cs dot umass dot edu 2004-11-14 15:47
---
I am doing a binary search right now, but it's probably Jeff's patch that
causes this bug.
With his patch, GCC shares some CASE_LABEL_EXPR within one SWITCH_EXPR.
When a basic block is copied, I d
ED
Keywords: ice-on-valid-code
Severity: critical
Priority: P2
Component: tree-optimization
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: kazu at cs dot umass dot edu
CC: gcc-bugs at gcc dot gnu dot org
GCC build triplet: i6
--
What|Removed |Added
CC||dpatel at apple dot com
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18472
t gnu dot org
ReportedBy: kazu at cs dot umass dot edu
CC: gcc-bugs at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18472
--- Additional Comments From kazu at cs dot umass dot edu 2004-11-03 16:51 ---
Jeff,
Oops, you're right!
Kazu Hirata
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15524
--- Additional Comments From kazu at cs dot umass dot edu 2004-11-03 16:33 ---
Richard,
My patches to expand_case does not change its asymptotic behavior.
Jeff,
You might want to use SBITMAP instead of BITMAP in your patch to
expand_case because the bitmap you construct will be dense
--- Additional Comments From kazu at cs dot umass dot edu 2004-11-01 19:34 ---
Subject: Re: [4.0 Regression] jump threading
on trees is slow with switch statements with large # of cases
Hi Steven,
> OK, then can you see if this hack helps...?
+ /* Step 4: Update the case la
ReportedBy: kazu at cs dot umass dot edu
CC: gcc-bugs at gcc dot gnu dot org
GCC target triplet: i686-pc-linux-gnu
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18146
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-23 20:48 ---
OK. Now mainline is a lot faster than 3.3 on this testcase.
3.3 takes more than 20 seconds
mainlines takes less than 3 seconds
Thanks Andrew!
--
What|Removed |Added
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-21 18:18 ---
The quadratic behavior of thread_jumps has been solved, but still
mainline is slower than 3.3 on this testcase.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17966
NCONFIRMED
Keywords: missed-optimization
Severity: enhancement
Priority: P2
Component: middle-end
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: kazu at cs dot umass dot edu
CC: gcc-bugs at gcc dot gnu dot org
GCC target triple
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-17 13:06 ---
Confirmed.
--
What|Removed |Added
Status|UNCONFIRMED |NEW
Component: middle-end
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: kazu at cs dot umass dot edu
CC: gcc-bugs at gcc dot gnu dot org
GCC target triplet: i686-pc-linux-gnu
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18038
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-16 15:49 ---
A patch here:
http://gcc.gnu.org/ml/gcc-patches/2004-10/msg01333.html
--
What|Removed |Added
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-16 15:49 ---
A patch here:
http://gcc.gnu.org/ml/gcc-patches/2004-10/msg01333.html
--
What|Removed |Added
gnu dot org
ReportedBy: kazu at cs dot umass dot edu
CC: gcc-bugs at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18031
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-16 03:50 ---
>From IRC:
kazu: Looks like PR 18008 only has a single AND with -march=pentiumpro,
but two ANDs with -march=i386
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18008
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-16 03:46 ---
All the interesting events happen in store_field.
When store_field expands "() (unsigned char) v",
it generates an AND with 3 because () is an anonymous type
of two-bit integer.
When it e
words: missed-optimization
Severity: enhancement
Priority: P2
Component: middle-end
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: kazu at cs dot umass dot edu
CC: gcc-bugs at gcc dot gnu dot org
GCC target triplet: i686-pc-lin
--
What|Removed |Added
Summary|a twiddleing of a |an xor of a single bit
||bitfield is inefficient
http://gcc.gnu
gnu dot org
ReportedBy: kazu at cs dot umass dot edu
CC: gcc-bugs at gcc dot gnu dot org
GCC target triplet: i686-pc-linux-gnu
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18029
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-15 21:18 ---
This is a regression from 3.3.3, which does not generate two consecutive andl.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18008
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-15 16:55 ---
There are at least two strange things.
1. We emit two ands in the first place.
2. The combiner does not seem to be working.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18008
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-15 16:42 ---
No, I don't think this is a duplicate of mine.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18008
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-13 15:38 ---
Andrew,
Your algorithm would still present a quadratic behavior in the
following situation.
bb0<- a block with COND_EXPR only
| \
| \
| bb1 <- a block with COND_EXP
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-13 14:24 ---
Andrew,
Regarding your patch, if we have exactly one basic block immediately preceding
EXIT_BLOCK_PTR, and that block happens to be a forwarder block, your algorithm
terminates without doing any further work
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-13 13:29 ---
I should note that I used -O2 when I came up with "44%" in the initial
description of this bug.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17966
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-13 12:35 ---
Andrew, I did it in the hard way. :-)
void
foo (int a)
{
int b;
goto b1;
b4: return;
b3: if (a) { b = a; return; } else goto b4;
b2: if (a) { b = a; return; } else goto b3;
b1: if (a) { b = a
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-13 12:11 ---
In this particular testcase, going through BBs in backwards does help
by cutting down the running time by about half, but I wonder if we can
create an identical testcase except that the order of BBs are
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-13 03:20 ---
tree_redirect_edge_and_branch needs O(n) time to redirect an edge of
a SWITCH_EXPR in the worst case, where n is the number of case labels
in a SWITHC_EXPR.
Therefore, redirecting O(n) edges costs O(n^2) in
c behavior in thread_jumps
Product: gcc
Version: 4.0.0
Status: UNCONFIRMED
Keywords: compile-time-hog
Severity: enhancement
Priority: P2
Component: tree-optimization
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: ka
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-12 14:11 ---
A patch posted at:
http://gcc.gnu.org/ml/gcc-patches/2004-10/msg00986.html
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17931
iority: P2
Component: rtl-optimization
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: kazu at cs dot umass dot edu
CC: gcc-bugs at gcc dot gnu dot org
GCC target triplet: i686-pc-linux-gnu
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17935
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-11 15:32 ---
I'll be testing a patch shortly.
--
What|Removed |Added
AssignedTo|unassigned a
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-11 14:47 ---
The combiner does try to combine andl and testb,
but the suggested combined insn is rejected by combine_validate_cost.
The cost of "andl $1, %eax" is 4.
The cost of "testb %al, %al" i
--- Additional Comments From kazu at cs dot umass dot edu 2004-10-11 13:54 ---
Actually, we don't need "movl $1, %eax" at the end, either.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17931
NFIRMED
Keywords: missed-optimization
Severity: enhancement
Priority: P2
Component: rtl-optimization
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: kazu at cs dot umass dot edu
CC: gcc-bugs at gcc dot gnu dot org
GCC target
301 - 344 of 344 matches
Mail list logo