This is an automated email from the git hooks/post-receive script. ppm-guest pushed a commit to annotated tag v0.36 in repository libmath-prime-util-perl.
commit 18302e13dba55e3d6ae189b8c6f0fe8258ff718d Author: Dana Jacobsen <d...@acm.org> Date: Tue Dec 24 21:21:31 2013 -0800 Small comment changes --- Changes | 3 ++- lehmer.c | 2 +- lmo.c | 8 ++++---- 3 files changed, 7 insertions(+), 6 deletions(-) diff --git a/Changes b/Changes index bafbab2..9d2d570 100644 --- a/Changes +++ b/Changes @@ -33,7 +33,8 @@ Revision history for Perl module Math::Prime::Util lehmer.c. Really only affects LMOS and Legendre, but it's significant for them. Thanks to Kim Walisch for questioning my earlier decision. - - Rewrite sieve composite map. Segment sieving is faster. + - Rewrite sieve composite map. Segment sieving is faster. It's a little + faster than primegen for me, but still slower than primesieve and yafu. 0.35 2013-12-08 diff --git a/lehmer.c b/lehmer.c index 40cbf14..c4d0a59 100644 --- a/lehmer.c +++ b/lehmer.c @@ -33,7 +33,7 @@ * * Times and memory use for prime_count(10^15) on a Haswell 4770K, asterisk * indicates parallel operation. The standalone versions of my code use - * Kim Walisch's excellent primesieve, which faster than my sieve. + * Kim Walisch's excellent primesieve, which is faster than my sieve. * His Lehmer/Meissel/Legendre seem a bit slower in serial, but * parallelize much better. * diff --git a/lmo.c b/lmo.c index 518876f..43993a6 100644 --- a/lmo.c +++ b/lmo.c @@ -86,10 +86,10 @@ typedef UV sword_t; #define SWORD_MASKBIT(bits) (UVCONST(1) << ((bits) % SWORD_BITS)) #define SWORD_CLEAR(s,bits) s[bits/SWORD_BITS] &= ~SWORD_MASKBIT(bits) -/* GCC 3.4-4.1 has broken 64-bit popcount. 4.2+ can generate awful code when - * it doesn't have asm. 4.8 seems to generate similar code to the below. - * Hence, only use the builtin when we have asm. - * Compile with -march=native to get a big speedup on newer processors. */ +/* GCC 3.4 - 4.1 has broken 64-bit popcount. + * GCC 4.2+ can generate awful code when it doesn't have asm (GCC bug 36041). + * When the asm is present (e.g. compile with -march=native on a platform that + * has them, like Nahelem+), then it is almost as fast as the direct asm. */ #if SWORD_BITS == 64 #if defined(__POPCNT__) && defined(__GNUC__) && (__GNUC__> 4 || (__GNUC__== 4 && __GNUC_MINOR__> 1)) #define bitcount(b) __builtin_popcountll(b) -- Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/pkg-perl/packages/libmath-prime-util-perl.git _______________________________________________ Pkg-perl-cvs-commits mailing list Pkg-perl-cvs-commits@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-perl-cvs-commits