On Wednesday 26 May 2010 22:02:44 Bill Hart wrote: > You can set minimum values in tuneup.c itself. >
it's already set to 3 , just wonder why it not doing what it should . I looked at this before , but clearly gave up. lazy boy.... > On 26 May 2010 22:03, Jason Moxham <[email protected]> wrote: > > PASS: t-redc_basecase > > PASS: t-divebyBm1of > > mulhigh_n.c:106: GNU MP assertion failed: n >= 3 > > /bin/bash: line 1: 16923 Abort (core dumped) ${dir}$tst > > FAIL: t-mullowhigh > > PASS: t-mullow_basecase > > PASS: t-neg > > PASS: t-mulmod_2expp1 > > PASS: t-mulmod_2expm1 > > PASS: t-tdiv_q > > PASS: t-sb_divappr_q > > PASS: t-dc_divappr_q_n > > PASS: t-inv_divappr_q_n > > PASS: t-invert > > PASS: t-sb_div_q > > PASS: t-sb_div_qr > > PASS: t-dc_div_q > > PASS: t-dc_div_qr > > PASS: t-dc_divappr_q > > PASS: t-dc_div_qr_n > > PASS: t-inv_divappr_q > > PASS: t-inv_div_q > > PASS: t-inv_div_qr > > PASS: t-inv_div_qr_n > > PASS: t-tdiv_qr > > PASS: t-sb_bdiv_q > > PASS: t-sb_bdiv_qr > > PASS: t-dc_bdiv_q_n > > PASS: t-dc_bdiv_qr_n > > PASS: t-dc_bdiv_qr > > PASS: t-dc_bdiv_q > > PASS: t-gcdext > > PASS: st_fat > > PASS: st_instrument > > ============================================================= > > 1 of 48 tests failed > > Please report to http://groups.google.co.uk/group/mpir-devel/ > > ============================================================= > > make[4]: *** [check-TESTS] Error 1 > > make[4]: Leaving directory `/scratch/jason/32/t2/tests/mpn' > > make[3]: *** [check-am] Error 2 > > make[3]: Leaving directory `/scratch/jason/32/t2/tests/mpn' > > make[2]: *** [check-recursive] Error 1 > > make[2]: Leaving directory `/scratch/jason/32/t2/tests' > > make[1]: *** [check-recursive] Error 1 > > make[1]: Leaving directory `/scratch/jason/32/t2' > > make: *** [check] Error 2 > > t2 > > > > PASSED CC=gcc CXX=g++ configure=ABI=32 > > PASSED CC=gcc CXX=g++ configure=ABI=32 --enable-cxx --enable-gmpcompat > > FAILED CC=gcc CXX=g++ configure=ABI=32 --enable-cxx --enable-gmpcompat -- > > enable-assert --enable-alloca=debug > > > > We have seen this error before , it's just the tuning on t2 32bit > > doesn't exclude that trivial case , we just manually fixed the param > > file last time , mulhigh is not used internally at the mo , and it's not > > documented so it wont effect any code , i'll fix it manually now , and > > try and work out why the tuning for it is broken (on sparc only) > > > > I think we should have a automatic mechanism for checking that the > > thresholds that mpir builds with satisfy all the restrictions even when > > we dont build with --enable-assert , something like > > > > CHECK_ASSERT(threshold_kara >=2 ) in the mul_n.c > > which are empty macros for building mpir > > > > and a script(python?) that extracts all out into one big c file , > > compiles and runs it at build time to see if it is satisfied > > (check_assert macro not empty) > > a script is an overkill , something simple like cat *.c | grep -e CHECK_ASSERT > check.c gcc -DCHECK_ASSERT(x)=etc... check.c as part of the prebuild ie with fac_ui.h etc I was pondering getting rid of the prebuild (among other things), to simplify the excessively complicated configure mechanism , or rather moving it into the predist phase (ie in make dist) . At the mo configure has to run on all systems , which means that we get odd errors on some systems that arn't tested much like bsd/solaris/mingw. If we push all the complicated configure bits into make dist then we only have to make them work on common systems(linux :) (OK windows if you want ) ( I joke , but modern , bug fixed systems , linux/cygwin/solaris?) . Note we dont have to push all the complicated configure mechanism out of the way , just the bits that arn't maintained by the GNU foundation , ie just the existing gmp extensions . Other advantages are we can upgrade easily to the latest gnulib,autotools etc . Enough prattle for now.... > > Jason > > > > On Wednesday 26 May 2010 21:21:35 Jason Moxham wrote: > >> I'm just testing the machines I guess are most likely to break things > >> > >> On t2 64bit > >> > >> PASS: t-dc_bdiv_qr > >> PASS: t-dc_bdiv_q > >> gcdext.c:1156: GNU MP assertion failed: an >= n > >> /bin/bash: line 1: 5208 Abort (core dumped) > >> ${dir}$tst FAIL: t-gcdext > >> PASS: st_fat > >> PASS: st_instrument > >> ============================================================= > >> 1 of 48 tests failed > >> Please report to http://groups.google.co.uk/group/mpir-devel/ > >> ============================================================= > >> make[4]: *** [check-TESTS] Error 1 > >> make[4]: Leaving directory `/scratch/jason/t2/tests/mpn' > >> make[3]: *** [check-am] Error 2 > >> make[3]: Leaving directory `/scratch/jason/t2/tests/mpn' > >> make[2]: *** [check-recursive] Error 1 > >> make[2]: Leaving directory `/scratch/jason/t2/tests' > >> make[1]: *** [check-recursive] Error 1 > >> make[1]: Leaving directory `/scratch/jason/t2' > >> make: *** [check] Error 2 > >> t2 > >> > >> PASSED CC=gcc CXX=g++ configure= > >> PASSED CC=gcc CXX=g++ configure= --enable-cxx --enable-gmpcompat > >> FAILED CC=gcc CXX=g++ configure= --enable-cxx --enable-gmpcompat > >> --enable- assert --enable-alloca=debug > >> > >> > >> the tuning has been done (on mark.skynet) so params should be correct > >> > >> On Wednesday 26 May 2010 20:29:44 Jeff Gilchrist wrote: > >> > On Wed, May 26, 2010 at 3:00 AM, Minh Nguyen <[email protected]> > > > > wrote: > >> > > MPIR 2.1.0-rc1 was released on 25th May 2010. > >> > > >> > Here are some testing results with: configure -enable-cxx; make; make > >> > check. > >> > > >> > - Linux 64bit - Core2 - icc (ICC) 11.0 20090318 > >> > All tests passed > >> > > >> > - Linux 64bit - Core2 - gcc version 4.1.2 20080704 > >> > All tests passed > >> > > >> > - Linux 64bit - Itanium2 - icc (ICC) 11.0 20090318 > >> > All tests passed > >> > > >> > - Linux 64bit - Itanium2 - gcc version 4.1.2 20070115 > >> > All tests passed > >> > > >> > - Linux 64bit - K8 - gcc version 3.4.6 20060404 > >> > All tests passed > >> > > >> > Jeff. > > > > -- > > You received this message because you are subscribed to the Google Groups > > "mpir-devel" group. To post to this group, send email to > > [email protected]. To unsubscribe from this group, send email > > to [email protected]. For more options, visit this > > group at http://groups.google.com/group/mpir-devel?hl=en. -- You received this message because you are subscribed to the Google Groups "mpir-devel" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/mpir-devel?hl=en.
