Matt Borland
|
08dcd6ea4a
|
Add float distance for __float128
|
2023-07-06 16:58:58 +02:00 |
|
Matt Borland
|
b69b9fd5db
|
Add explicit specialization for __float128
|
2023-07-06 16:58:13 +02:00 |
|
Matt Borland
|
6ed41ab8fa
|
Fix __float128 specialization
|
2023-07-06 16:57:50 +02:00 |
|
Matt Borland
|
65abeb8f8e
|
Add __float128 conversion operator
|
2023-07-06 16:57:27 +02:00 |
|
Matt Borland
|
ae45b11c92
|
Expand types and restore complete testing
|
2023-07-06 16:38:54 +02:00 |
|
Matt Borland
|
3b16386e92
|
Use simpler float distance calculation
|
2023-07-06 16:01:20 +02:00 |
|
Matt Borland
|
30b45d5554
|
Add bit size guards
|
2023-07-06 15:57:17 +02:00 |
|
Matt Borland
|
31f51d14d0
|
Specialize long double testing
|
2023-07-06 15:26:01 +02:00 |
|
Matt Borland
|
1a4671024d
|
Improve rounding to 1 ULP
|
2023-07-06 15:07:45 +02:00 |
|
Matt Borland
|
31fa9edadd
|
Use single mantissa
|
2023-07-06 14:17:29 +02:00 |
|
Matt Borland
|
196fc6b9e1
|
Use lookup tables for big and small values
|
2023-07-06 13:37:10 +02:00 |
|
Matt Borland
|
5286001323
|
Add hex formatting
|
2023-07-06 13:36:59 +02:00 |
|
Matt Borland
|
4afef63dad
|
Only use fast path
|
2023-07-06 10:52:20 +02:00 |
|
Matt Borland
|
074464ce3e
|
Change 80 bit mask
|
2023-07-05 17:01:18 +02:00 |
|
Matt Borland
|
3f675c1a7a
|
Add binary output
|
2023-07-05 16:46:39 +02:00 |
|
Matt Borland
|
31852daf55
|
Add second mantissa calculation path
|
2023-07-05 15:35:33 +02:00 |
|
Matt Borland
|
140e8d5f44
|
Fix normalizing the significand
|
2023-07-05 15:04:40 +02:00 |
|
Matt Borland
|
5a3a9b3999
|
Increase precision of masking operations
[ci skip]
|
2023-06-30 17:40:12 +02:00 |
|
Matt Borland
|
dc6ce63e01
|
Fix 256-bit high bit calculation
|
2023-06-30 17:27:52 +02:00 |
|
Matt Borland
|
5808c27206
|
Add operator&
|
2023-06-30 17:23:57 +02:00 |
|
Matt Borland
|
fd61faaa51
|
Add conversion operators
|
2023-06-30 17:15:43 +02:00 |
|
Matt Borland
|
0f1970706d
|
Add numeric_limits specialization
|
2023-06-30 17:03:58 +02:00 |
|
Matt Borland
|
cfc9382469
|
Add 64-bit comparisons
|
2023-06-30 16:58:39 +02:00 |
|
Matt Borland
|
bfaf185ac3
|
Add 64-bit compatible operators/ and %
|
2023-06-30 16:20:27 +02:00 |
|
Matt Borland
|
7d18408c0c
|
Fix precision during calculation of z
|
2023-06-30 16:15:19 +02:00 |
|
Matt Borland
|
8243c2d387
|
Add sub operator
|
2023-06-30 16:08:42 +02:00 |
|
Matt Borland
|
112d3ece6b
|
Add operator|
|
2023-06-30 16:00:25 +02:00 |
|
Matt Borland
|
3d26f26c9b
|
Add operators >>= and <<=
|
2023-06-30 15:45:10 +02:00 |
|
Matt Borland
|
7ed86677ab
|
Add 256 bit comp operators
|
2023-06-30 15:03:01 +02:00 |
|
Matt Borland
|
1c9573344c
|
Improve final calculation
|
2023-06-30 14:42:40 +02:00 |
|
Matt Borland
|
38752a1e10
|
Fix calculation of u and p
|
2023-06-30 13:40:58 +02:00 |
|
Matt Borland
|
e0adbd0855
|
Add portable shifting for significant bit
[ci skip]
|
2023-06-29 15:05:32 +02:00 |
|
Matt Borland
|
b26cd77e9c
|
Improve portability of masking operation
|
2023-06-29 14:25:27 +02:00 |
|
Matt Borland
|
b1cbcab893
|
Fix calculation in step 5
|
2023-06-29 13:50:30 +02:00 |
|
Matt Borland
|
3331f04f7b
|
Change max and min powers
|
2023-06-29 13:49:33 +02:00 |
|
Matt Borland
|
a4a3c64f5a
|
Improve handling of out of range values
|
2023-06-29 13:37:12 +02:00 |
|
Matt Borland
|
487b70c554
|
Rework algorithm
|
2023-06-29 11:42:26 +02:00 |
|
Matt Borland
|
fe0726d5c5
|
Add left shift for u256
|
2023-06-28 18:59:58 +02:00 |
|
Matt Borland
|
f1c9a7d866
|
Add diagnostic for use of BOOST_UNLIKELY fallback routine
|
2023-06-27 16:24:46 +02:00 |
|
Matt Borland
|
8220960f53
|
Fix requirements for testing
|
2023-06-26 17:05:20 +02:00 |
|
Matt Borland
|
7522d6306e
|
Add random value test
|
2023-06-26 16:51:55 +02:00 |
|
Matt Borland
|
8e9d2fd59a
|
Add __float128 overload
|
2023-06-26 15:44:10 +02:00 |
|
Matt Borland
|
c68e1ab080
|
Use runtime calculation for fallback path
|
2023-06-26 14:51:09 +02:00 |
|
Matt Borland
|
f12a5f81b9
|
Restore number of tests
|
2023-06-26 14:47:40 +02:00 |
|
Matt Borland
|
d2d6bd59d4
|
Add runtime computation of power of 5
|
2023-06-26 14:37:55 +02:00 |
|
Matt Borland
|
29873873ab
|
Add 512 bit multiplication
|
2023-06-26 14:16:08 +02:00 |
|
Matt Borland
|
6bb0df05db
|
Add 256bit operator+
|
2023-06-26 14:15:32 +02:00 |
|
Matt Borland
|
f8bf05af18
|
Move 128-bit long double testing
|
2023-06-26 10:47:05 +02:00 |
|
Matt Borland
|
1ece924e99
|
Remove GCC workaround for 128-bit systems
|
2023-06-26 09:57:23 +02:00 |
|
Matt Borland
|
79be2b5b1f
|
Remove broken platform workaround
|
2023-06-26 09:42:11 +02:00 |
|