diff options
| author | Torbjorn Granlund <tege@gmplib.org> | 2009-11-28 12:57:53 +0100 |
|---|---|---|
| committer | Torbjorn Granlund <tege@gmplib.org> | 2009-11-28 12:57:53 +0100 |
| commit | 4fab69d02817734bc7e21e6c3f61bd3098b3b394 (patch) | |
| tree | 0a43f3246a49a0a04fba10900dcea1868c208578 /mpn/generic/dive_1.c | |
| parent | 2c639d5f1b78f47267eecfca5396a0eabc7ea5e9 (diff) | |
| download | gmp-4fab69d02817734bc7e21e6c3f61bd3098b3b394.tar.gz | |
Finish renaming BITS_PER_MP_LIMB to GMP_LIMB_BITS.
Diffstat (limited to 'mpn/generic/dive_1.c')
| -rw-r--r-- | mpn/generic/dive_1.c | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/mpn/generic/dive_1.c b/mpn/generic/dive_1.c index 27df57b80..f246b091c 100644 --- a/mpn/generic/dive_1.c +++ b/mpn/generic/dive_1.c @@ -30,7 +30,7 @@ along with the GNU MP Library. If not, see http://www.gnu.org/licenses/. */ /* Divide a={src,size} by d=divisor and store the quotient in q={dst,size}. q will only be correct if d divides a exactly. - A separate loop is used for shift==0 because n<<BITS_PER_MP_LIMB doesn't + A separate loop is used for shift==0 because n<<GMP_LIMB_BITS doesn't give zero on all CPUs (for instance it doesn't on the x86s). This separate loop might run faster too, helping odd divisors. @@ -50,7 +50,7 @@ along with the GNU MP Library. If not, see http://www.gnu.org/licenses/. */ faster on some CPUs and would mean just the shift==0 style loop would be needed. - If n<<BITS_PER_MP_LIMB gives zero on a particular CPU then the separate + If n<<GMP_LIMB_BITS gives zero on a particular CPU then the separate shift==0 loop is unnecessary, and could be eliminated if there's no great speed difference. |