This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug rtl-optimization/87985] New: Compile-time and memory hog w/ -O1 -ftree-slp-vectorize
- From: "asolokha at gmx dot com" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Mon, 12 Nov 2018 11:00:52 +0000
- Subject: [Bug rtl-optimization/87985] New: Compile-time and memory hog w/ -O1 -ftree-slp-vectorize
- Auto-submitted: auto-generated
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87985
Bug ID: 87985
Summary: Compile-time and memory hog w/ -O1
-ftree-slp-vectorize
Product: gcc
Version: unknown
Status: UNCONFIRMED
Keywords: compile-time-hog, memory-hog
Severity: normal
Priority: P3
Component: rtl-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: asolokha at gmx dot com
Target Milestone: ---
All gcc versions from the current trunk down to 4.5 branch (where
-ftree-slp-vectorize was introduced) take indefinite time and memory compiling
gcc/testsuite/gcc.dg/pr85180.c at any -O level (except -Og) w/
-ftree-slp-vectorize.
% timeout 10 gcc-9.0.0-alpha20181111 -O1 -ftree-slp-vectorize -c
gcc/testsuite/gcc.dg/pr85180.c
zsh: exit 124 timeout 10 gcc-9.0.0-alpha20181111 -O1 -ftree-slp-vectorize -c
perf top:
Overhead Shared Objec Symbol
56.63% cc1 [.] operand_equal_p
7.21% cc1 [.] tree_operand_check
6.65% cc1 [.] tree_strip_nop_conversions
5.14% cc1 [.] element_precision
3.86% cc1 [.] fold_binary_loc
2.42% cc1 [.] split_constant_offset_1
2.12% cc1 [.] split_tree
2.02% cc1 [.] wide_int_to_tree_1
1.79% cc1 [.] generic_simplify_PLUS_EXPR
1.44% cc1 [.] split_constant_offset
1.08% cc1 [.] build2
0.93% cc1 [.] ggc_internal_alloc
0.83% cc1 [.] integer_zerop
0.83% cc1 [.] build_int_cst
0.55% cc1 [.] tree_swap_operands_p
0.50% cc1 [.] canonize
0.50% cc1 [.] commutative_tree_code
It's even not necessary to leave ten Zs, as the issue is reproducible w/ as
"low" as two Ys.