| Commit message (Expand) | Author | Age | Files | Lines |
... | |
* | Use a temporary array instead of the arg array y[] for calling | bde | 2008-02-25 | 2 | -8/+8 |
* | Change __ieee754_rem_pio2f() to return double instead of float so that | bde | 2008-02-25 | 5 | -27/+24 |
* | Fix some off-by-1 errors. | bde | 2008-02-25 | 3 | -6/+11 |
* | Let PowerPC world optionally build with -msoft-float. For FPU-less PowerPC | raj | 2008-02-24 | 1 | -0/+5 |
* | Optimize the 9pi/2 < |x| <= 2**19pi/2 case some more by avoiding an | bde | 2008-02-23 | 2 | -11/+8 |
* | Avoid using FP-to-integer conversion for !(amd64 || i386) too. Use the | bde | 2008-02-22 | 2 | -6/+4 |
* | Remove the "quick check no cancellation" optimization for | bde | 2008-02-22 | 1 | -12/+1 |
* | Optimize the 9pi/2 < |x| <= 2**19pi/2 case on amd64 and i386 by avoiding | bde | 2008-02-22 | 2 | -0/+18 |
* | Add an irint() function in inline asm for amd64 and i386. irint() is | bde | 2008-02-22 | 1 | -0/+30 |
* | Optimize the conversion to bits a little (by about 11 cycles or 16% | bde | 2008-02-22 | 1 | -5/+13 |
* | Optimize the fixup for +-0 by using better classification for this case | bde | 2008-02-22 | 1 | -2/+4 |
* | Fix rintl() on signaling NaNs and unsupported formats. | bde | 2008-02-22 | 1 | -5/+3 |
* | s/rcsid/__FBSDID/ | das | 2008-02-22 | 79 | -241/+158 |
* | Remove an unused variable. | das | 2008-02-22 | 3 | -3/+3 |
* | Eliminate some warnings. | das | 2008-02-22 | 3 | -9/+3 |
* | Merge cosmetic changes from e_rem_pio2.c 1.10 (convert to __FBSDID(); | bde | 2008-02-19 | 1 | -6/+4 |
* | Optimize for 3pi/4 <= |x| <= 9pi/4 in much the same way as for | bde | 2008-02-19 | 1 | -18/+56 |
* | Rearrange the polynomial evaluation for better parallelism. This | bde | 2008-02-19 | 2 | -9/+9 |
* | Document return values better. | das | 2008-02-18 | 1 | -3/+34 |
* | Add tgammaf() as a simple wrapper around tgamma(). | das | 2008-02-18 | 5 | -9/+63 |
* | 2 long double constants were missing L suffixes. This helped break tanl() | bde | 2008-02-18 | 1 | -2/+2 |
* | Fix a typo which broke k_tanl.c on !(amd64 || i386). | bde | 2008-02-18 | 1 | -1/+1 |
* | Inline __ieee754__rem_pio2(). With gcc4-2, this gives an average | bde | 2008-02-18 | 4 | -13/+19 |
* | Use volatile hacks to make sure exp() generates an underflow | das | 2008-02-17 | 1 | -1/+2 |
* | Hook up sinl(), cosl(), and tanl() to the build. | das | 2008-02-17 | 2 | -7/+12 |
* | Add implementations of sinl(), cosl(), and tanl(). | das | 2008-02-17 | 7 | -0/+367 |
* | Documentation for sinl(), cosl(), and tanl(). | das | 2008-02-17 | 3 | -35/+46 |
* | Add kernel functions for 128-bit long doubles. These could be improved | das | 2008-02-17 | 3 | -0/+239 |
* | Add kernel functions for 80-bit long doubles. Many thanks to Steve and | das | 2008-02-17 | 3 | -0/+264 |
* | Add more pi for long doubles. Also, avoid storing multiple copies | das | 2008-02-17 | 4 | -50/+154 |
* | Sigh, the weak reference for ceill(), floorl() and truncl() was in | bde | 2008-02-15 | 3 | -6/+10 |
* | Oops, the weak reference for ceill(), floorl() and truncl() was in the | bde | 2008-02-14 | 6 | -12/+12 |
* | Use the expression fabs(x+0.0)+fabs(y+0.0) instad of a+b (where a is | bde | 2008-02-14 | 2 | -8/+8 |
* | Fix the hi+lo decomposition for 2/(3ln2). The decomposition needs to | bde | 2008-02-14 | 1 | -2/+2 |
* | Use the expression (x+0.0)-(y+0.0) instead of x+y when mixing NaN arg(s). | bde | 2008-02-14 | 2 | -10/+8 |
* | s_ceill.c | bde | 2008-02-13 | 3 | -9/+6 |
* | On arches where long double is the same as double, alias ceil(), floor() | bde | 2008-02-13 | 4 | -6/+19 |
* | Fix the C version of ceill(x) for -1 < x <= -0 in all rounding modes. | bde | 2008-02-13 | 1 | -1/+1 |
* | Fix exp2*(x) on signaling NaNs by returning x+x as usual. | bde | 2008-02-13 | 4 | -4/+4 |
* | Rearrange the polynomial evaluation for better parallelism. This is | bde | 2008-02-13 | 1 | -3/+4 |
* | Use hardware remainder on amd64 since it is 5 to 10 times faster than | bde | 2008-02-13 | 3 | -1/+78 |
* | Fix remainder() and remainderf() in round-towards-minus-infinity mode | bde | 2008-02-12 | 2 | -8/+8 |
* | Use double precision for z and thus for the entire calculation of | bde | 2008-02-11 | 1 | -3/+4 |
* | As usual, use a minimax polynomial that is specialized for float | bde | 2008-02-09 | 1 | -7/+8 |
* | Fix a comment about coefficients and expand a related one. | bde | 2008-02-09 | 1 | -2/+2 |
* | Fix truncl() when the result should be -0.0L. When the result is +-0.0L, | bde | 2008-02-08 | 1 | -1/+2 |
* | Oops, fix the fix in rev.1.10. logb() and logbf() were broken on | bde | 2008-02-08 | 1 | -5/+4 |
* | Use a better method of scaling by 2**k. Instead of adding to the | bde | 2008-02-07 | 2 | -26/+16 |
* | Use a better method of scaling by 2**k. Instead of adding to the | bde | 2008-02-07 | 6 | -55/+57 |
* | As for the float trig functions and logf, use a minimax polynomial | bde | 2008-02-06 | 1 | -6/+7 |