Anton Mitrokhin
|
4d4a512f72
|
fixed warnings produced by 64 bit reduce functions without affecting the functionality
|
2014-12-11 17:58:51 +04:00 |
|
Anton Mitrokhin
|
91538b5366
|
fixed noisy warnings in icpc13 produced by unsupported 64 bit 'reduce' functions
|
2014-12-11 17:31:48 +04:00 |
|
Anton Mitrokhin
|
e4c79418c8
|
'__scatter64_i64' function fixed for icc 13
|
2014-12-11 14:13:30 +04:00 |
|
Anton Mitrokhin
|
dc2e174d4f
|
added compiler version guards, '__scatter64_i64' is not implemented
|
2014-12-11 13:13:16 +04:00 |
|
Vsevolod Livinskiy
|
44ee5737a7
|
Changes in cast_zext
|
2014-12-04 23:04:25 +04:00 |
|
Vsevolod Livinskiy
|
9022806fc0
|
Changes in cast
|
2014-12-04 19:53:23 +04:00 |
|
Vsevolod Livinskiy
|
91866396ef
|
Fix for half to float
|
2014-12-04 19:40:30 +04:00 |
|
Vsevolod Livinskiy
|
ff9380b248
|
Fix for count_leading_zeros
|
2014-12-04 19:30:30 +04:00 |
|
Vsevolod Livinskiy
|
b00575e95c
|
Fix for __cast_zext(__vec16_i64, __vec16_i1)
|
2014-12-04 18:59:40 +04:00 |
|
Vsevolod Livinskiy
|
df9f839ed4
|
Merge
|
2014-12-04 13:45:47 +04:00 |
|
Vsevolod Livinskiy
|
8d51e0620e
|
Fix for int64 comparison
|
2014-12-04 13:37:24 +04:00 |
|
Vsevolod Livinskiy
|
70adb5d283
|
Fix for unsigned cmp for int64
|
2014-12-04 13:36:19 +04:00 |
|
Vsevolod Livinskiy
|
7e63862dff
|
__sqrt_varying_double
|
2014-12-04 13:35:03 +04:00 |
|
Vsevolod Livinskiy
|
190d7957d4
|
Fix for __shuffle2_float
|
2014-12-04 13:35:03 +04:00 |
|
Vsevolod Livinskiy
|
307c825c4e
|
Fix for __shift_i32
|
2014-12-04 13:35:03 +04:00 |
|
Vsevolod Livinskiy
|
7e39ed115d
|
Fix for shuffle2_double
|
2014-12-04 13:35:03 +04:00 |
|
Vsevolod Livinskiy
|
52eddcdcfd
|
Some changes in max/min_varying_uint/int64
|
2014-12-04 13:35:03 +04:00 |
|
Vsevolod Livinskiy
|
0d9eceb668
|
Fix for prefetch
|
2014-12-04 13:35:02 +04:00 |
|
Vsevolod Livinskiy
|
7084432a0e
|
Fix for max/min_varying_uint/int64
|
2014-12-04 13:35:02 +04:00 |
|
Hans van den Bogert
|
d1b4ad6aca
|
Added cstdlib include
the atoi and other functions were not transitively included under clang/llvm3.5
|
2014-12-01 14:09:13 +01:00 |
|
Anton Mitrokhin
|
6f518dfcc9
|
fixed lshr (i64 ...) function implementation
|
2014-11-27 18:33:44 +04:00 |
|
Anton Mitrokhin
|
0a987ad06f
|
reduce_add_int8/16 implementation fixed
|
2014-11-27 17:41:21 +04:00 |
|
Anton Mitrokhin
|
296b057a0a
|
added debug helpers for knc-i1x16.h
|
2014-11-27 16:54:46 +04:00 |
|
Anton Mitrokhin
|
5f3128bbb2
|
cast_sext/sext, i64 (l)shl/r and mul
|
2014-11-27 16:53:25 +04:00 |
|
Anton Mitrokhin
|
7233e25907
|
removed unnecessary &255's (knc.h)
|
2014-11-14 17:17:24 +04:00 |
|
Anton Mitrokhin
|
615d36b236
|
Merge branch 'master' of https://github.com/ncos/ispc into knc-backend-merge
Conflicts:
examples/intrinsics/knc.h
|
2014-11-14 11:54:34 +04:00 |
|
Anton Mitrokhin
|
b21043c309
|
- fixed 'cast_trunc', 'cast_zext' and 'cast_sext' implementations (the previous was faulty)
- added '__select (i8/16)', '__equal_i8/16_and_mask', '__not_equal_i8/16_and_mask' functions
|
2014-11-13 22:41:55 +04:00 |
|
Vsevolod Livinskiy
|
37a4362417
|
Fix for __cast_uitofp
|
2014-11-13 20:21:02 +04:00 |
|
Vsevolod Livinskiy
|
8f68769af8
|
cast double/float to/from int64
|
2014-11-13 18:10:31 +04:00 |
|
Anton Mitrokhin
|
79fa1c3d4d
|
Added some more functions to knc.h:
- __scatter64_i64
- __scatter_base_offsets32_double
- __scatter_base_offsets32_i64
|
2014-11-13 17:18:47 +04:00 |
|
Vsevolod Livinskiy
|
6a2cb442ee
|
__reduce_add/min/max_int64
|
2014-11-13 14:21:10 +04:00 |
|
Vsevolod Livinskiy
|
6606d20a47
|
insert/extract_element_i8/16 and broadcast_i8/16
|
2014-11-13 13:48:01 +04:00 |
|
Vsevolod Livinskiy
|
67243f3550
|
Merge commit
|
2014-11-13 13:00:17 +04:00 |
|
Vsevolod Livinskiy
|
308746c7fb
|
Minor fixes to remove copy-paste
|
2014-11-13 12:45:16 +04:00 |
|
Vsevolod Livinskiy
|
28e9032b10
|
__not_equal_i8/16 was added
|
2014-11-13 12:44:41 +04:00 |
|
Vsevolod Livinskiy
|
44bcce3cc8
|
__smear_i8/16 was added
|
2014-11-13 12:44:41 +04:00 |
|
Anton Mitrokhin
|
4152a0f9ed
|
rewrite masked_load_i8/16 for it not to segfault
|
2014-11-09 22:06:10 +04:00 |
|
Anton Mitrokhin
|
d8f4635366
|
rewrittem __masked_store_i8/16 functions so they won't segfault
|
2014-11-09 20:01:58 +04:00 |
|
Anton Mitrokhin
|
e7717e58b5
|
- added 'cast_trunk' and 'cast_zext' sets of functions
- commented out fault [] overload for vec16_i64
- added loop based function macros for vec16_i8/16 data types
|
2014-11-09 11:59:19 +04:00 |
|
Vsevolod Livinskiy
|
83f3ee7cfa
|
__shuffle2_i8/16/32/64 was added
|
2014-11-09 00:21:20 +04:00 |
|
Vsevolod Livinskiy
|
1b8afb73ad
|
__rotate_i8/16/64 was added
|
2014-11-09 00:21:10 +04:00 |
|
Vsevolod Livinskiy
|
67cc62d619
|
Minor fixes to remove copy-paste
|
2014-11-09 00:21:08 +04:00 |
|
Vsevolod Livinskiy
|
12376e6a0c
|
__shuffle_i8/16 was added. __reduce* functions were fixed.
|
2014-11-09 00:21:04 +04:00 |
|
Vsevolod Livinskiy
|
9316dd65c0
|
__not_equal_i8/16 was added
|
2014-11-09 00:21:01 +04:00 |
|
Vsevolod Livinskiy
|
3b1445c660
|
__smear_i8/16 was added
|
2014-11-09 00:20:58 +04:00 |
|
Vsevolod Livinskiy
|
39b1e4a204
|
masked_load/store_i16 was added
|
2014-11-09 00:20:24 +04:00 |
|
Dmitry Babokin
|
df7d65b076
|
Merge pull request #891 from ncos/knc-backend-merge
Add missing _cast_uitofp/fptoui, setzero and min_varying functions
|
2014-11-06 09:03:54 -08:00 |
|
Anton Mitrokhin
|
15e1f86711
|
- Removed the debug code occasionally sneaked to the repo
- Put 12 knc runfails to the fail_db.txt (were invisible due to debug code)
- Several minor fixes not affecting functionality
|
2014-11-06 16:48:51 +04:00 |
|
Vsevolod Livinskiy
|
9a8782d122
|
Fix for load/store i16
|
2014-10-30 18:37:38 +04:00 |
|
Anton Mitrokhin
|
e2249365f3
|
setzero functions for i8 and i16
|
2014-10-30 18:36:39 +04:00 |
|