H. Peter Anvin
2ce0274303
Use a 32-bit floating-point limb size; support 8-bit float ...
Use a 32-bit limb size ("like a digit, but bigger") for floating-point
conversion. This cuts the number of multiplications per constant by a
factor of four.
This means supporting fractional-limb-sized numbers, so while we're at
it, add support for 8-bit floating point numbers (apparently used in
graphics and in audio compression applications.)
2007-10-29 20:20:12 -07:00
2007-09-13 18:13:20 -07:00
2002-11-07 22:45:46 +00:00
2002-09-12 16:34:06 +00:00
2007-10-28 23:23:24 -07:00
2002-04-30 20:52:49 +00:00
2002-04-30 20:52:08 +00:00
2007-04-12 02:40:54 +00:00
2002-04-30 20:52:49 +00:00
2002-04-30 21:02:01 +00:00
2002-05-27 02:19:43 +00:00
2002-05-27 23:19:35 +00:00
2002-04-30 21:08:11 +00:00
2002-04-30 20:52:08 +00:00
2007-10-11 20:32:33 -07:00
2002-04-30 20:52:49 +00:00
2002-05-17 04:51:10 +00:00
2007-10-11 20:32:33 -07:00
2002-05-22 23:01:06 +00:00
2002-05-22 20:48:01 +00:00
2007-10-29 20:20:12 -07:00
2007-10-16 15:46:04 -07:00
2007-10-29 20:20:12 -07:00
2007-10-24 15:29:51 -07:00
2007-10-08 12:12:23 -07:00
2007-10-22 16:53:48 -07:00
2007-09-17 17:25:27 -07:00
2002-05-07 04:31:28 +00:00
2002-04-30 20:52:08 +00:00
2002-04-30 20:51:32 +00:00
2002-04-30 20:52:08 +00:00
2007-09-22 22:35:28 -07:00
2007-09-25 14:26:03 -07:00
2003-08-06 07:13:59 +00:00
2003-09-08 22:59:45 +00:00
2007-10-19 13:16:51 -07:00
2007-09-02 16:37:03 +00:00
2007-09-25 15:44:40 -07:00
2002-05-24 14:33:35 +00:00
2003-09-25 11:43:28 +00:00
2003-09-08 22:59:45 +00:00
2002-04-30 20:52:49 +00:00
2007-10-11 20:32:33 -07:00
2002-04-30 20:52:08 +00:00
2007-09-02 06:23:29 +00:00
2007-09-17 13:19:25 -07:00
2007-10-26 21:38:02 -07:00
2007-09-28 17:17:20 -07:00
2007-10-19 14:42:29 -07:00
2007-09-22 21:50:03 -07:00
2002-09-12 04:33:58 +00:00
2007-10-15 17:48:43 -07:00
2003-09-10 23:33:38 +00:00
2007-10-19 14:19:52 -07:00
2007-10-05 17:44:16 -07:00
2ce0274303