tested with the converted asm code on cuda1, sage,box1,2,3 (k10,penryn,i7,k8,k7) with options (various random choices) of enable-assert , enable-fat, enable-alloca=debug,build=none ,enable-cxx the usual configure , make ,make check
On Tuesday 14 April 2009 13:18:16 Bill Hart wrote: > Here is the latest test matrix: > > iras 4.1.2 ncmk +cmk / 4.3.3 ncmk +cmk > eno - 4.3.0 ncmk fcmk +cmk ;parallel build is VERY slow > cleo 4.1.2 ncmk +cmk / 4.3.3 ncmk +cmk > cicero 4.3.0 ncmk fcmk +cmk > varro A 4.0.1 ncmk +cmk / 4.3.3 ncmk +cmk ;broken parallel build > fulvia 4.3.3 ncmk fcmk +cmk / S ncm*k fc** +cmk ;no parallel build > with Sun cc > menas 4.2.1 ncmk fcmk +cmk / 4.3.3 ncmk fcmk +cmk ;broken parallel > build > cato Pathscale ncmk +cmk > mark 4.3.3 ncmk +cmk / S ncmk +cmk ;parallel make not available > sage.math ncmk fcmk +cmk > cuda1.tarbox.org ncmk fcmk +cmk > > * addmul_2.asm won't assemble with Sun assembler > ** build failure due to Sun assembler (@GOT) > > n = normal build > f = fat build > + = cxx build > c = configure > m = make > k = make check > > So we only have two issues: > > 1) The Sun --enable-fat assembler issue with @GOT that we have had in > previous versions - not terribly important as it only affects Sun cc > on x86_64 and only fat binary builds > > 2) addmul_2.asm does not build with the x86_64 Sun assembler on > fulvia. To get around this issue I simply deleted the file. Obviously > as soon as Brian has converted it to yasm format, this issue will > disappear as the Sun assembler will no longer be used. > > We should run try tests on sage.math (penryn), cuda1 (k10) and > host-57-183 (k8) to verify everything is fine after the new assembly > code was added. > > Hopefully by the end of the day we'll have a release candidate. > > If anyone else would like to test the mpir-1.1 branch from svn: > > svn co http://modular.math.jmu.edu/svn/mpir/mpir/mpir-1.1 mpir-1.1 > > we'd be much obliged. It would help if you can provide the output of > > config.guess > gcc -v > > for both test pass and failure reports. Note 1.1 will definitely fail > to build on Sun X86_64 with Sun cc, however a build on that system > with gcc ought to pass. > > Benchmarks can now be obtained by doing make bench in the bench > directory. > > Bill. > > > > --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---
