Vsevolod Livinskij
|
735e6a8ab3
|
Saturation arithmetic mul and div for int8/int16/int32 and div for int64 was added
|
2014-02-18 02:07:13 +04:00 |
|
Vsevolod Livinskij
|
f5508db24f
|
Saturation arithmetic (sub and add) was added for int32/int64.
|
2014-02-17 18:55:40 +04:00 |
|
Dmitry Babokin
|
04fda2fcbe
|
Merge pull request #739 from ifilippov/fluky
Fix for fluky problem 'argument out of range'
|
2014-02-13 16:13:32 +03:00 |
|
Ilia Filippov
|
e7b3a1c822
|
fix for fluky problem 'argument out of range'
|
2014-02-13 16:47:33 +04:00 |
|
Dmitry Babokin
|
ccbcd0a80d
|
Merge pull request #738 from ifilippov/err
Set 3.4 version of LLVM to alloy.py, perf.py correction
|
2014-02-13 14:24:31 +03:00 |
|
Ilia Filippov
|
54b991cfcb
|
set 3.4 version of LLVM to alloy.py, perf.py correction
|
2014-02-13 15:14:04 +04:00 |
|
Dmitry Babokin
|
b719019b26
|
Merge pull request #737 from ifilippov/err
Adding patch for LLVM 3.4 for bug #712
|
2014-02-13 13:33:31 +03:00 |
|
Ilia Filippov
|
cc81cd3215
|
adding patch for LLVM 3.4 for bug #712
|
2014-02-13 12:57:05 +04:00 |
|
Dmitry Babokin
|
e8039cd822
|
Merge pull request #673 from Vsevolod-Livinskij/master
Saturation arithmetic.
|
2014-02-11 16:40:40 +03:00 |
|
Vsevolod
|
a3c77e6dc6
|
Merge pull request #1 from dbabokin/Vsevolod-Livinskij-master
Fix for generic-1
|
2014-02-11 16:35:41 +03:00 |
|
Dmitry Babokin
|
ea0a514e03
|
Fix for generic-1
|
2014-02-11 15:33:23 +04:00 |
|
evghenii
|
193bba77b0
|
accuracy fix
|
2014-02-11 11:49:03 +01:00 |
|
Evghenii
|
f0779f95a3
|
added double precision tests
|
2014-02-11 11:40:40 +01:00 |
|
Vsevolod Livinskij
|
65d947e449
|
Else branch with error report was added
|
2014-02-10 15:18:48 +04:00 |
|
Vsevolod Livinskij
|
cef5b2eb04
|
Some changes in saturation arithmetic
|
2014-02-10 12:40:53 +04:00 |
|
Vsevolod Livinskij
|
1c1614d207
|
Some errors in comments and code were fixed
|
2014-02-09 21:39:42 +04:00 |
|
evghenii
|
8490efe0ad
|
fix for knc.h. Due to a bug in ICC (tested with 13.1.3 & 14.0.1) ,the resulting .cpp file fails to compile
|
2014-02-07 16:00:21 +01:00 |
|
evghenii
|
438cee4e21
|
added support for double precision/native transendentals/trigonometry
|
2014-02-07 15:43:42 +01:00 |
|
Evghenii
|
70a9b286e5
|
added support for native and double precision trigonometry/transendentals
|
2014-02-07 15:28:39 +01:00 |
|
Evghenii
|
81aa19a8f0
|
added use of native_transendentals, need to add IR
|
2014-02-07 11:49:24 +01:00 |
|
Evghenii
|
668645fcda
|
first commit
|
2014-02-07 11:05:36 +01:00 |
|
Evghenii
|
14e76108cb
|
optimization for _all
|
2014-02-06 14:24:50 +01:00 |
|
Evghenii
|
9ecb4f4ac8
|
added tunnings for aobench
|
2014-02-06 10:13:18 +01:00 |
|
Evghenii
|
9e1ab7c6b6
|
allow to add ISPC_FLAGS
|
2014-02-06 10:13:01 +01:00 |
|
Evghenii
|
8ffa84f875
|
added some #ifdef .. #endif for control flow tests
|
2014-02-06 10:12:31 +01:00 |
|
Evghenii
|
c8e92feb14
|
added additional optimizaotion passes for PTX target
|
2014-02-06 10:11:58 +01:00 |
|
Evghenii
|
c23dd8a951
|
fixed __puts_nvptx
|
2014-02-05 17:48:04 +01:00 |
|
Evghenii
|
7b2ceba128
|
added "internal" for helper functions to avoid them being exported to PTX
|
2014-02-05 17:02:05 +01:00 |
|
Dmitry Babokin
|
2570385770
|
Merge pull request #730 from egaburov/double_math
Added double precision support for reciprocals: rsqrt rcp
|
2014-02-05 17:57:39 +03:00 |
|
Evghenii
|
aeb2f01a15
|
some performance fix. it works, but have no idea why. checkpoint
|
2014-02-05 15:36:06 +01:00 |
|
evghenii
|
c59cff396d
|
added {rsqrt,rcp}d support for knc.h. test-147.ispc & test-148.ispc pass.
|
2014-02-05 13:55:38 +01:00 |
|
evghenii
|
ecc9c88ff8
|
fix packed_store_active2 for knc-i1x8.h
|
2014-02-05 13:52:24 +01:00 |
|
Evghenii
|
eb01ffd4e6
|
first commit for {rsqrt,rcp}d knc support. going to test on other node now
|
2014-02-05 13:43:07 +01:00 |
|
Evghenii
|
f225b558ec
|
added {rsqrt,rcp}d support for sse4.h
|
2014-02-05 13:42:45 +01:00 |
|
Evghenii
|
688d9c9a82
|
added support for rsqrtd/rcpd for generic-*.h
|
2014-02-05 13:20:44 +01:00 |
|
evghenii
|
09e8381ec7
|
change {rsqrt,rcp}_double to {rsqrt,rcp}d_decl
|
2014-02-05 13:05:04 +01:00 |
|
evghenii
|
732a315a4b
|
removed __declspec(safe) duplicate
|
2014-02-05 13:04:45 +01:00 |
|
Evghenii
|
686c1d676d
|
improvements
|
2014-02-05 12:04:36 +01:00 |
|
Evghenii
|
048da693c5
|
fix sqrt
|
2014-02-05 10:52:08 +01:00 |
|
Dmitry Babokin
|
9a3b949687
|
Merge pull request #734 from dbabokin/run_test_compiler
Update list of accepted system compilers in run_test.py
|
2014-02-05 12:40:07 +03:00 |
|
Dmitry Babokin
|
40186d3813
|
Update list of accepted system compilers in run_test.py
|
2014-02-05 13:39:28 +04:00 |
|
Dmitry Babokin
|
66c986ba13
|
Merge pull request #733 from jbrodman/master
Modify alloy.py to put dbg llvm builds in different folders. Disallow initializing void * with ptr to const. (#731)
|
2014-02-05 11:32:25 +03:00 |
|
jbrodman
|
98cfc17843
|
Fix bug with printing due to uneven handling of bool types
|
2014-02-04 08:12:02 -08:00 |
|
Evghenii
|
d3a6693eef
|
adding __have_native_{rsqrtd,rcpd} to select between native support for double precision reciprocals and using slower but safe version in stdlib
|
2014-02-04 16:29:23 +01:00 |
|
Evghenii
|
fe98fe8cdc
|
added fast approximate rcp(double) accurate to 15 digits
|
2014-02-04 15:23:34 +01:00 |
|
Evghenii
|
eb1a495a7a
|
added support for fast approximate rsqrt(double). Provide 16 digit accurancy but is over 3x faster than 1/sqrt(double)
|
2014-02-04 14:44:54 +01:00 |
|
jbrodman
|
720975dff4
|
Disallow initializing void * with ptr to const.
|
2014-02-04 03:36:19 -08:00 |
|
jbrodman
|
4ee0e6996a
|
Merge branch 'master' of https://github.com/ispc/ispc
|
2014-02-04 02:48:41 -08:00 |
|
jbrodman
|
47bdca1041
|
Modify alloy.py to put dbg llvm builds in different folders.
|
2014-02-04 02:46:07 -08:00 |
|
Evghenii
|
c2ed214a74
|
added declaretion for movmsk_ptx
|
2014-02-03 08:57:27 +01:00 |
|