Archive: computation

Up to parent directory

Knuth's 33 constants

knuth33-1000.pdf 13Dec02 14:57:07EST 81K Brent's paper on computing the constants to 1000 digits.

knuth33-1000.txt 11Jan03 09:06:46EST 43K Brent's output, not in the order listed in the paper.

knuth33-50.mma 11Jan03 09:06:54EST 2.9K Mathematica program to compute the constants.

knuth33-50.txt 11Jan03 09:07:00EST 4.8K List for 50 significant digits.

a few more constants

trigconst-50.mma 11Jan03 09:07:55EST 2.7K

trigconst-50.txt 11Jan03 09:08:07EST 4.3K

Also see the constants at the numbers, constants, and computation site.

floating point

macharpr.f 11Jan03 09:07:43EST 13K A printing version of William. J. Cody's Fortran code to determine floating point parameters. Has to be edited for single or double precision.

macharspr.f
machardpr.f 11Jan03 09:07:49EST 11Jan03 09:07:36EST 13K 13K Single and double precision versions of the above, compile and run.

float-params.c
float-params.txt 25May09 08:42:46EDT 25May09 09:25:57EDT 3.4K 3.0K C program to print out the values of the floating point parameters defined in float.h under C99. They are not always viewable in the file itself, especially in gcc. The .txt file has the values for Mac OS X and Linux.

machar.fs
machar.log 09Mar05 08:44:50EST 09Mar05 08:59:06EST 15K 1.1K A port of MACHAR to ANS Forth, plus its change log.

celefunt/ [an error occurred while processing this directive] A copy of the netlib Fortran source for W. J. Cody's celefunt complex function accuracy tests, also split into double precision programs with a Makefile.

celefunt.tar.gz 20Jan05 14:59:16EST 23K Archive file containing the above directory.

setjmp-fpmode.html 19Jun05 11:35:01EDT 5.3K A survey we conducted which shows that several systems do and several do not save and restore the C99 floating-point rounding control mode in setjmp()/longjmp(). At the end is a short polemic, where we advocate that the save/restore should not be automatic.

fe-handling-example.c 01Oct09 08:09:10EDT 8.9K Example C99 code for nondefault IEEE 754 exception handling for Mac OS X ppc/intel and GNU/Linux. It implements the GNU/Linux functions feenableexcept(), fedisableexcept(), and fegetexcept() for OS X.

Up to top of archive


`knuth33-1000.pdf`	`13Dec02 14:57:07EST`	`81K`	Brent's paper on computing the constants to 1000 digits.
`knuth33-1000.txt`	`11Jan03 09:06:46EST`	`43K`	Brent's output, not in the order listed in the paper.
`knuth33-50.mma`	`11Jan03 09:06:54EST`	`2.9K`	Mathematica program to compute the constants.
`knuth33-50.txt`	`11Jan03 09:07:00EST`	`4.8K`	List for 50 significant digits.


`trigconst-50.mma`	`11Jan03 09:07:55EST`	`2.7K`
`trigconst-50.txt`	`11Jan03 09:08:07EST`	`4.3K`


`macharpr.f`	`11Jan03 09:07:43EST`	`13K`	A printing version of William. J. Cody's Fortran code to determine floating point parameters. Has to be edited for single or double precision.
`macharspr.f` `machardpr.f`	`11Jan03 09:07:49EST 11Jan03 09:07:36EST`	`13K 13K`	Single and double precision versions of the above, compile and run.
`float-params.c` `float-params.txt`	`25May09 08:42:46EDT 25May09 09:25:57EDT`	`3.4K 3.0K`	C program to print out the values of the floating point parameters defined in `float.h` under C99. They are not always viewable in the file itself, especially in gcc. The `.txt` file has the values for Mac OS X and Linux.
`machar.fs` `machar.log`	`09Mar05 08:44:50EST 09Mar05 08:59:06EST`	`15K 1.1K`	A port of MACHAR to ANS Forth, plus its change log.
`celefunt/`	`[an error occurred while processing this directive]`		A copy of the netlib Fortran source for W. J. Cody's celefunt complex function accuracy tests, also split into double precision programs with a Makefile.
`celefunt.tar.gz`	`20Jan05 14:59:16EST`	`23K`	Archive file containing the above directory.
`setjmp-fpmode.html`	`19Jun05 11:35:01EDT`	`5.3K`	A survey we conducted which shows that several systems do and several do not save and restore the C99 floating-point rounding control mode in `setjmp()/longjmp()`. At the end is a short polemic, where we advocate that the save/restore should not be automatic.
`fe-handling-example.c`	`01Oct09 08:09:10EDT`	`8.9K`	Example C99 code for nondefault IEEE 754 exception handling for Mac OS X ppc/intel and GNU/Linux. It implements the GNU/Linux functions `feenableexcept()`, `fedisableexcept()`, and `fegetexcept()` for OS X.