This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Irix6.4 regression test egcs-1.0.3a vs egcs-19980803
- To: egcs at cygnus dot com, law at cygnus dot com
- Subject: Irix6.4 regression test egcs-1.0.3a vs egcs-19980803
- From: tprince at cat dot e-mail dot com
- Date: Thu, 13 Aug 1998 17:36:23 EDT
I was unable to get the testing template going; however, both
versions were able to complete the compiler sections of the
current version of gcc/testsuite. As the following comparison
shows, there is just one apparent regression from 1.0.3a to
19980803. Both versions were build in -n32 mode on Irix6.4
Livermore Kernel tests also show progress. All kernels show
expected double precision accuracy, and 7 of the 24 run at the
same speed as with SGI's best compiler. Performance
differences are due mainly to the lack of implementation in egcs
of the floating point multiply-add sequence instructions. 116
Mflops double precision geometric mean is not bad for egcs.
I was able to get the gcc-2.8.1/g77-0.5.23 test going, thanks to
Oliva's script, and will send those comparisons later today.
Irix6.4 -n32 mode, apparent regressions from egcs-1.0.3a to egcs-19980803
> FAIL: gcc.dg/980626-1.c (test for excess errors)
53,54c19,20
< # of expected passes 6316
< # of unexpected failures 862
---
> # of expected passes 7485
> # of unexpected failures 11
56d21
< # of untested testcases 318
58c23
< /wrk1/aerotherm/dawes/src/gnu/egcs-1.0.3a/i64/gcc/xgcc version egcs-2.90.29 980515 (egcs-1.0.3 release)
---
> /wrk1/aerotherm/dawes/src/gnu/egcs-19980803/i64/gcc/xgcc version egcs-2.91.53 19980803 (gcc2 ss-980609 experimental)
< # of expected passes 3689
< # of unexpected failures 396
< # of unexpected successes 3
< # of expected failures 219
---
> # of expected passes 4218
> # of unexpected successes 1 (not tested properly for egcs-1.0.3a)
> # of expected failures 88
471c37
< # of expected passes 277
< # of unexpected failures 107
< # of untested testcases 31
< Couldn't determine version of /wrk1/aerotherm/dawes/src/gnu/egcs-1.0.3a/i64/gcc/g77: Reading specs from /wrk1/aerotherm/dawes/src/gnu/
egcs-1.0.3a/i64/gcc/specs
< gcc version egcs-2.90.29 980515 (egcs-1.0.3 release)
< /wrk1/aerotherm/dawes/src/gnu/egcs-1.0.3a/i64/gcc/ld -call_shared -no_unresolved -init __do_global_ctors -fini __do_global_dtors -_SY
STYPE_SVR4 -n32 /usr/lib32/mips3/crt1.o -L/usr/lib32/mips3 -L/usr/lib32 /wrk1/aerotherm/dawes/src/gnu/egcs-1.0.3a/i64/gc
c/crtbegin.o -L/wrk1/aerotherm/dawes/src/gnu/egcs-1.0.3a/i64/gcc -L/homee/d28930/bin/103/lib/gcc-lib/mips-sgi-irix6.4/egcs-2.90.29 -L/ho
mee/d28930/bin/103/mips-sgi-irix6.4/lib -L/homee/d28930/bin/103/lib -lf2c -lm -dont_warn_unused -lgcc -warn_unused -dont
_warn_unused -lc -warn_unused -dont_warn_unused -lgcc -warn_unused /wrk1/aerotherm/dawes/src/gnu/egcs-1.0.3a/i64/gcc/crtend.o /usr/lib32
/mips3/crtn.o
---
> # of expected passes 415
> /wrk1/aerotherm/dawes/src/gnu/egcs-19980803/i64/gcc/g77 g77 version egcs-2.91.53 19980803 (gcc2 ss-980609 experimental) (from FSF-g77 v
ersion 0.5.24-19980725)
>
To: INTERNET - IBMMAIL