On 19/02/2019 12:45, Richard Biener wrote:
The following limits mpfr operations to the maximum exponent range as
determined by available float modes. This avoids excessive work
for evaluating functions like ctan for large arguments.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk.
Richard.
2019-02-19 Richard Biener <rguent...@suse.de>
PR middle-end/88074
* toplev.c (do_compile): Initialize mpfr's exponent range
based on available float modes.
* gcc.dg/pr88074.c: New testcase.
This patch causes a number of test regressions for (at least) arm and
amdgcn. I've run a bisect and confirmed this is the first commit where
it doesn't work.
spawn -ignore SIGHUP amdgcn-unknown-amdhsa-gcc
.../gcc/testsuite/gcc.dg/torture/builtin-math-7.c
-fno-diagnostics-show-caret -fno-diagnostics-show-line-numbers
-fdiagnostics-color=never -O0 -lm -o ./builtin-math-7.exe^M
ld: error: undefined symbol: link_error^M
>>> referenced by /tmp/ccKynVNn.o:(main)^M
^M
ld: error: undefined symbol: link_error^M
>>> referenced by /tmp/ccKynVNn.o:(main)^M
collect2: error: ld returned 1 exit status^M
compiler exited with status 1
output is:
ld: error: undefined symbol: link_error^M
>>> referenced by /tmp/ccKynVNn.o:(main)^M
^M
ld: error: undefined symbol: link_error^M
>>> referenced by /tmp/ccKynVNn.o:(main)^M
collect2: error: ld returned 1 exit status^M
FAIL: gcc.dg/torture/builtin-math-7.c -O0 (test for excess errors)
I see the same failures posted here:
https://gcc.gnu.org/ml/gcc-testresults/2019-02/msg02389.html
Andrew
