Ooh a big-endian user. Guess I'll have to write those routines. For now you can comment out the offending #error but don't use kenlm's trie implementation (the default probing hash table is fine). It looks like the switchable endianness on PPC is a choice made by the operating system and you're running one that chooses big).
On 11/27/10 09:30, Korzec, Sanne wrote: > Hi, > > I've recently succesfully installed srilm on my ppc64 system. > > Now I am compiling moses and get the following error in > moses/kenlm/util/bit_packing.hh > > make output: > > Making all in kenlm > make[2]: Entering directory `/gpfs/h07/skorzec/moses/kenlm' > g++ -DHAVE_CONFIG_H -I. -I.. -W -Wall -ffor-scope > -D_FILE_OFFSET_BITS=64 -D_LARGE_FILES -DTRACE_ENABLE=1 > -I/home/skorzec/srilm/include -I/home/skorzec/moses/kenlm -g -O2 -MT > trie.o -MD -MP -MF .deps/trie.Tpo -c -o trie.o `test -f 'lm/trie.cc' || > echo './'`lm/trie.cc > In file included from lm/trie.cc:3: > ./util/bit_packing.hh:16:2: error: #error The bit aligned storage > functions assume little endian architecture > make[2]: *** [trie.o] Error 1 > make[2]: Leaving directory `/gpfs/h07/skorzec/moses/kenlm' > make[1]: *** [all-recursive] Error 1 > make[1]: Leaving directory `/gpfs/h07/skorzec/moses' > make: *** [all] Error 2 > > I am not sure how little endian is set on the ppc64 system, is there > some way to set this specifically? > > the bit_packing.hh suggests it is written with 64-bit architecture in > mind, why then do I get this error. Or why is little endian not set on > my architecture? > > thanks for any help. > > Sanne > > > > _______________________________________________ > Moses-support mailing list > Moses-support@mit.edu > http://mailman.mit.edu/mailman/listinfo/moses-support _______________________________________________ Moses-support mailing list Moses-support@mit.edu http://mailman.mit.edu/mailman/listinfo/moses-support
