1fd83c9 | Pranav Bhandarkar | 10 August 2018, 23:44:17 UTC | Spooky action inside CodeGen_LLVM.cpp | 10 August 2018, 23:44:17 UTC |
72cd870 | Steven Johnson | 09 August 2018, 01:30:01 UTC | WIP | 09 August 2018, 01:30:01 UTC |
ee46e13 | Steven Johnson | 09 August 2018, 01:25:39 UTC | WIP | 09 August 2018, 01:25:39 UTC |
a404d56 | Steven Johnson | 09 August 2018, 01:20:33 UTC | Merge branch 'master' into srj-hvx-codegen-bug | 09 August 2018, 01:20:33 UTC |
693d287 | Steven Johnson | 09 August 2018, 01:20:30 UTC | WIP | 09 August 2018, 01:20:30 UTC |
ea9c863 | Steven Johnson | 07 August 2018, 17:19:49 UTC | Merge pull request #3195 from halide/srj-uint64 call to halide_hexagon_run(): arg_sizes should be uint64*, not size_t* | 07 August 2018, 17:19:49 UTC |
9c53d76 | Steven Johnson | 07 August 2018, 17:19:37 UTC | Merge pull request #3196 from halide/srj-stable Use std::stable_sort in LoopCarry | 07 August 2018, 17:19:37 UTC |
ddb1619 | Steven Johnson | 06 August 2018, 21:39:18 UTC | Use std::stable_sort in LoopCarry This is really just for debugging hygiene: use of std::sort means that chains of same length can be in different orders in different C++ compilers; this doesn't appear to affect correctness of output, but having the sort be stable does make comparing IR from different compiler versions easier. | 06 August 2018, 21:39:18 UTC |
e21e61a | Steven Johnson | 06 August 2018, 17:43:08 UTC | call to halide_hexagon_run(): arg_sizes should be uint64*, not size_t* | 06 August 2018, 17:43:08 UTC |
1289ad3 | Steven Johnson | 03 August 2018, 22:27:18 UTC | Merge pull request #3190 from halide/srj-foreach2 Make Buffer::for_each_value const-correct | 03 August 2018, 22:27:18 UTC |
a7215a6 | Steven Johnson | 03 August 2018, 18:25:35 UTC | Make Buffer::for_each_value const-correct Declare it const, then add perfect-forwarding to some helper methods; this allows us to call it on `const Buffer<T>` (which are fine to mutate if T is not const) and also `Buffer<const T>` (which are now correctly prohibited from mutating the values via nonconst reference args to the lambda). | 03 August 2018, 18:25:35 UTC |
366f949 | Steven Johnson | 02 August 2018, 17:36:59 UTC | Merge pull request #3041 from matthiaskramm/ctlsz Make CodeGen_C aware of count_leading_zeros | 02 August 2018, 17:36:59 UTC |
fd47a65 | Steven Johnson | 02 August 2018, 17:16:35 UTC | Merge pull request #3118 from SanderVocke/disable_leakcheck Disable asan leak checking during code generation. | 02 August 2018, 17:16:35 UTC |
2101955 | Steven Johnson | 02 August 2018, 16:48:42 UTC | Merge pull request #3173 from halide/srj-copy-as Add some HalideBuffer convenience methods (Issue #3113) | 02 August 2018, 16:48:42 UTC |
2e7fb15 | Andrew Adams | 02 August 2018, 15:19:52 UTC | Merge pull request #3172 from halide/fix_warp_shuffles Fix some pathological behavior in warp shuffles | 02 August 2018, 15:19:52 UTC |
aa70aaa | Steven Johnson | 01 August 2018, 23:58:33 UTC | Merge pull request #3183 from halide/bump_llvm_version LLVM is bumping their version number to 8 | 01 August 2018, 23:58:33 UTC |
6c3d30b | Steven Johnson | 01 August 2018, 23:08:18 UTC | Change copy() to just return nonconst | 01 August 2018, 23:08:18 UTC |
e0316d1 | Steven Johnson | 01 August 2018, 22:43:58 UTC | Remove nonstatic make_with_shape_of() | 01 August 2018, 22:43:58 UTC |
3b88293 | Andrew Adams | 01 August 2018, 22:35:36 UTC | Add test generator to reproduce hexagon issue | 01 August 2018, 22:35:36 UTC |
d8917c4 | Andrew Adams | 01 August 2018, 22:17:50 UTC | Some build bots don't have the requisite cuda version | 01 August 2018, 22:17:50 UTC |
ac129fe | Andrew Adams | 01 August 2018, 22:00:29 UTC | LLVM is bumping their version number to 8 | 01 August 2018, 22:00:29 UTC |
2292112 | Andrew Adams | 01 August 2018, 20:27:18 UTC | Merge remote-tracking branch 'origin/master' into metaprogrammed_simplifier_rules | 01 August 2018, 20:27:18 UTC |
f39ea35 | Andrew Adams | 01 August 2018, 18:12:42 UTC | Merge remote-tracking branch 'origin/master' into fix_warp_shuffles | 01 August 2018, 18:12:42 UTC |
8c772dd | Andrew Adams | 01 August 2018, 17:26:01 UTC | Merge pull request #3165 from halide/quieter_test_correctness only report failures in test_correctness | 01 August 2018, 17:26:01 UTC |
0122fd7 | Sander Vocke | 01 August 2018, 11:13:40 UTC | Merge remote-tracking branch 'origin/master' into disable_leakcheck | 01 August 2018, 11:13:40 UTC |
92a16ea | Zalman Stern | 01 August 2018, 01:05:52 UTC | Merge pull request #3177 from halide/readme_update Update README.md to reflect a slightly more current version of target support. | 01 August 2018, 01:05:52 UTC |
a9de173 | Zalman Stern | 01 August 2018, 00:52:25 UTC | Merge pull request #3178 from halide/srj-warn Don't emit #pragma message in D3D12 backend | 01 August 2018, 00:52:25 UTC |
58d2c8a | Steven Johnson | 01 August 2018, 00:48:46 UTC | Don't emit #pragma message in D3D12 backend | 01 August 2018, 00:48:46 UTC |
4829e0a | Zalman Stern | 01 August 2018, 00:38:35 UTC | Typo. | 01 August 2018, 00:38:35 UTC |
b61de00 | Zalman Stern | 01 August 2018, 00:33:25 UTC | Update README.md to reflect a slightly more current version of target support. | 01 August 2018, 00:33:25 UTC |
5713da6 | Steven Johnson | 31 July 2018, 23:08:18 UTC | Merge branch 'metaprogrammed_simplifier_rules' of https://github.com/halide/Halide into metaprogrammed_simplifier_rules | 31 July 2018, 23:08:18 UTC |
dfcf9e4 | Steven Johnson | 31 July 2018, 23:08:16 UTC | Merge branch 'master' into metaprogrammed_simplifier_rules | 31 July 2018, 23:08:16 UTC |
fd5947a | Steven Johnson | 31 July 2018, 20:47:11 UTC | Fix comment | 31 July 2018, 20:47:11 UTC |
d8edcae | Dillon Sharlet | 31 July 2018, 19:51:15 UTC | Merge pull request #3166 from halide/faster_cascaded_filters Speed up cascaded filters test | 31 July 2018, 19:51:15 UTC |
39c5839 | Steven Johnson | 31 July 2018, 18:19:02 UTC | Add some HalideBuffer convenience methods (Issue #3113) - augment copy() to allow copying-to-a-different-type - add nonstatic variant of make_with_shape_of() - add reset() | 31 July 2018, 18:19:02 UTC |
e0e73f1 | Andrew Adams | 31 July 2018, 18:10:17 UTC | Fix some pathological behavior in warp shuffles | 31 July 2018, 18:10:17 UTC |
ddbf7e6 | Andrew Adams | 31 July 2018, 18:08:57 UTC | Merge remote-tracking branch 'origin/master' into quieter_test_correctness | 31 July 2018, 18:08:57 UTC |
51fd2c8 | Steven Johnson | 31 July 2018, 17:04:29 UTC | Merge pull request #3155 from suyogsarda/master Update image info API's for halide_buffer_t | 31 July 2018, 17:04:29 UTC |
92fe75a | Steven Johnson | 31 July 2018, 16:51:49 UTC | Merge pull request #3168 from halide/srj-buf Buffer move-construct ctor was missing assert_can_convert_from | 31 July 2018, 16:51:49 UTC |
a3ee5f7 | Steven Johnson | 31 July 2018, 00:56:30 UTC | Merge pull request #3158 from halide/srj-incl Add missing #include to HexagonAlignment.h | 31 July 2018, 00:56:30 UTC |
01cad4f | Steven Johnson | 31 July 2018, 00:37:46 UTC | Buffer move-construct ctor was missing assert_can_convert_from | 31 July 2018, 00:37:46 UTC |
db256bd | Zalman Stern | 30 July 2018, 23:40:17 UTC | Merge pull request #3160 from halide/openglcompute-location-fix Add explicit location layout specifier to uniforms. | 30 July 2018, 23:40:17 UTC |
30ab4c4 | Steven Johnson | 30 July 2018, 23:35:27 UTC | Merge branch 'master' into faster_cascaded_filters | 30 July 2018, 23:35:27 UTC |
0295c73 | Steven Johnson | 30 July 2018, 23:34:51 UTC | Merge branch 'master' into quieter_test_correctness | 30 July 2018, 23:34:51 UTC |
9652625 | Steven Johnson | 30 July 2018, 23:30:57 UTC | Merge pull request #3159 from halide/srj-d3d Fix D3D12 issues | 30 July 2018, 23:30:57 UTC |
d8ab812 | Steven Johnson | 30 July 2018, 20:02:54 UTC | Merge branch 'master' into srj-d3d | 30 July 2018, 20:02:54 UTC |
49136b3 | Steven Johnson | 30 July 2018, 18:11:39 UTC | Merge branch 'master' into openglcompute-location-fix | 30 July 2018, 18:11:39 UTC |
77ed646 | Steven Johnson | 30 July 2018, 18:09:19 UTC | Merge pull request #3081 from tbennun/master Makefile LLVM static library fix | 30 July 2018, 18:09:19 UTC |
a861bf9 | Andrew Adams | 30 July 2018, 17:25:47 UTC | Speed up cascaded filters test | 30 July 2018, 17:25:47 UTC |
dbca67d | Andrew Adams | 30 July 2018, 17:11:36 UTC | only failures in test_correctness | 30 July 2018, 17:11:36 UTC |
598cca3 | Andrew Adams | 30 July 2018, 16:52:28 UTC | Merge branch 'metaprogrammed_simplifier_rules' of https://github.com/halide/Halide into metaprogrammed_simplifier_rules | 30 July 2018, 16:52:28 UTC |
968aa7a | Andrew Adams | 30 July 2018, 16:52:23 UTC | New simplifier is better at nested if statements | 30 July 2018, 16:52:23 UTC |
ccb66a4 | Andrew Adams | 30 July 2018, 16:25:36 UTC | Merge branch 'master' into metaprogrammed_simplifier_rules | 30 July 2018, 16:25:36 UTC |
98f0b1a | Matthias Kramm | 14 June 2018, 01:17:54 UTC | always use inline implementation of clz | 30 July 2018, 05:46:50 UTC |
bf76726 | Matthias Kramm | 11 June 2018, 12:53:22 UTC | use "sizeof(long long) * 8" instead of "64" | 30 July 2018, 05:46:50 UTC |
9f5dc7c | Matthias Kramm | 09 June 2018, 18:08:52 UTC | add tests for 8/16/32/64 bit | 30 July 2018, 05:46:50 UTC |
a1c9683 | Matthias Kramm | 09 June 2018, 02:17:17 UTC | use clzll instead of clz, for 64 bit integers | 30 July 2018, 05:46:50 UTC |
f3a26ef | Matthias Kramm | 08 June 2018, 21:57:55 UTC | Make CodeGen_C aware of count_leading_zeros | 30 July 2018, 05:46:50 UTC |
d14d3a1 | Z Stern | 28 July 2018, 00:28:59 UTC | Add explicit location layout specifier to uniforms as GLSL does not guarantee they are in any particular order otherwise. | 28 July 2018, 00:28:59 UTC |
be66f83 | Steven Johnson | 27 July 2018, 21:38:42 UTC | Remove redundant 'return' to silence ClangTidy | 27 July 2018, 21:38:42 UTC |
866c3ce | Steven Johnson | 27 July 2018, 21:09:16 UTC | Fix D3D12 issues - avoid the use of fallthru-without-break in CodeGen_D3D12Compute_Dev.cpp (some compiler settings assume this is an error, and the conventional equivalent is fine here) - allow runtime linker to compile when WITH_D3D12 is not defined | 27 July 2018, 21:09:16 UTC |
6a7ff1b | Steven Johnson | 27 July 2018, 20:20:23 UTC | Add missing #include to HexagonAlignment.h | 27 July 2018, 20:20:23 UTC |
cd95b27 | Steven Johnson | 27 July 2018, 16:56:13 UTC | Merge pull request #3116 from halide/srj-features Rework runtime to allow more than 64 Target::Features (Issue #2911) | 27 July 2018, 16:56:13 UTC |
78fde6c | Suyog Sarda | 27 July 2018, 15:48:04 UTC | Update image info API's for halide_buffer_t Replace buffer_t with halide_buffer_t and other data structure. | 27 July 2018, 15:48:04 UTC |
14f594e | Tal Ben-Nun | 27 July 2018, 11:52:25 UTC | Merge pull request #2 from halide/master Sync | 27 July 2018, 11:52:25 UTC |
d734dc0 | Steven Johnson | 26 July 2018, 22:29:32 UTC | Merge branch 'master' into srj-features | 26 July 2018, 22:29:32 UTC |
d91647c | Dillon Sharlet | 26 July 2018, 16:15:07 UTC | Merge pull request #2913 from suyogsarda/master Add v65 instructions for HVX | 26 July 2018, 16:15:07 UTC |
b6b9d6b | Dillon Sharlet | 26 July 2018, 16:14:11 UTC | Merge pull request #3152 from halide/interleave-bug Interleave bug | 26 July 2018, 16:14:11 UTC |
efe17ef | Andrew Adams | 26 July 2018, 15:59:56 UTC | Merge pull request #3153 from jn80842/verification_pr add fold when rewriting constant expressions | 26 July 2018, 15:59:56 UTC |
b18d04a | jn80842 | 26 July 2018, 15:20:42 UTC | fixed broken merge | 26 July 2018, 15:20:42 UTC |
831b259 | jn80842 | 26 July 2018, 14:55:19 UTC | merge | 26 July 2018, 14:55:19 UTC |
457d333 | Marcos Slomp | 26 July 2018, 14:18:23 UTC | Merge pull request #2755 from halide/d3d12 Direct3D 12 (compute) GPU back-end | 26 July 2018, 14:18:23 UTC |
6664656 | Steven Johnson | 25 July 2018, 21:42:11 UTC | Merge pull request #3145 from halide/srj-py Python: don't delete intermediates, don't rely on HALIDE_PATH | 25 July 2018, 21:42:11 UTC |
565e2d9 | jn80842 | 25 July 2018, 21:40:01 UTC | add fold when rewriting constant expressions | 25 July 2018, 21:40:01 UTC |
0f3f455 | Marcos Slomp | 25 July 2018, 20:57:14 UTC | Merge remote-tracking branch 'remotes/upstream/master' into d3d12 | 25 July 2018, 20:57:14 UTC |
7018190 | Steven Johnson | 25 July 2018, 20:50:05 UTC | Merge pull request #3150 from halide/hex_fix_predicated_store_test Only run predicated vector LUT test on X86 | 25 July 2018, 20:50:05 UTC |
6aa7ea8 | Steven Johnson | 25 July 2018, 17:30:34 UTC | Merge branch 'metaprogrammed_simplifier_rules' of https://github.com/halide/Halide into metaprogrammed_simplifier_rules | 25 July 2018, 17:30:34 UTC |
563b161 | Steven Johnson | 25 July 2018, 17:30:31 UTC | Merge branch 'master' into metaprogrammed_simplifier_rules | 25 July 2018, 17:30:31 UTC |
90d182b | Patricia Suriana | 25 July 2018, 15:50:25 UTC | Only run predicated vector LUT test on X86 | 25 July 2018, 15:50:25 UTC |
ef6eea5 | Pranav Bhandarkar | 24 July 2018, 23:43:32 UTC | Fix failing tests. The following tests were failing on HVX on master after interleave-bug was merged into master. correctness_argmax correctness_bool_compute_root_vectorize correctness_fuse correctness_in_place correctness_multi_way_select correctness_simd_op_check correctness_undef Fixed these tests now. The problem with all, except simd_op_check was that we weren't handling scalar loads and stores in HexagonAlignment.h We do, to a certain extent only though. As in we look for "too much" alignment but we do not crash any more like before. simd_op_check was failing all valign and vlalign issues and that was due to a bug in AlignLoads that could be traced to refactoring the alignment logic into HexagonAlignment.h | 24 July 2018, 23:43:32 UTC |
0c6c2ad | Pranav Bhandarkar | 24 July 2018, 23:42:20 UTC | Revert "Revert "Interleave bug"" This reverts commit 38e5ca1a4f2796a52fc888ab72a281f45018b560. | 24 July 2018, 23:42:20 UTC |
9d13fd5 | Patricia Suriana | 24 July 2018, 21:05:22 UTC | Merge pull request #2777 from halide/stencil_chain Add stencil chain apps | 24 July 2018, 21:05:22 UTC |
ea02650 | Marcos Slomp | 24 July 2018, 18:55:58 UTC | Merge branch 'master' into d3d12 | 24 July 2018, 18:55:58 UTC |
4635aba | Marcos Slomp | 24 July 2018, 18:38:23 UTC | complete skip of the d3d12compute code when building for 32bit targets (will produce an empty translation unit) | 24 July 2018, 18:38:23 UTC |
42a49f4 | Patricia Suriana | 24 July 2018, 17:42:57 UTC | Merge pull request #3147 from halide/auto-sched-string-bugs Fix bugs in auto-scheduler string generator | 24 July 2018, 17:42:57 UTC |
4fa2d39 | Patricia Suriana | 24 July 2018, 17:14:39 UTC | Merge pull request #3146 from halide/predicated_store_bugs Fix predicated store to use the new computed index | 24 July 2018, 17:14:39 UTC |
eef75f1 | Patricia Suriana | 23 July 2018, 23:27:26 UTC | Add get_pipeline call to auto-sched str | 23 July 2018, 23:27:26 UTC |
6ce4d9f | Patricia Suriana | 23 July 2018, 23:04:29 UTC | Fix bugs in auto-scheduler string version | 23 July 2018, 23:04:29 UTC |
8087921 | Patricia Suriana | 23 July 2018, 21:36:08 UTC | Fix predicated store to use the new computed index | 23 July 2018, 21:36:08 UTC |
4008657 | Steven Johnson | 23 July 2018, 20:34:14 UTC | Python: don't delete intermediates, don't rely on HALIDE_PATH | 23 July 2018, 20:34:14 UTC |
c2b7984 | Patricia Suriana | 23 July 2018, 20:10:57 UTC | Add boundary condition to stencil chain | 23 July 2018, 20:10:57 UTC |
ba88422 | Steven Johnson | 23 July 2018, 18:01:27 UTC | Use a fake static_assert, because C++11 not available for runtime | 23 July 2018, 18:01:27 UTC |
d1e55b7 | Steven Johnson | 23 July 2018, 17:32:55 UTC | Add static_assert | 23 July 2018, 17:32:55 UTC |
82723af | Steven Johnson | 23 July 2018, 17:18:43 UTC | Merge pull request #3143 from halide/revert-3030-interleave-bug Revert "Interleave bug" | 23 July 2018, 17:18:43 UTC |
bf9e1f4 | Marcos Slomp | 23 July 2018, 16:15:24 UTC | Merge branch 'master' into d3d12 | 23 July 2018, 16:15:24 UTC |
38e5ca1 | Steven Johnson | 23 July 2018, 15:45:53 UTC | Revert "Interleave bug" | 23 July 2018, 15:45:53 UTC |
d49c695 | Patricia Suriana | 23 July 2018, 15:29:07 UTC | Merge branch 'master' of github.com:halide/Halide into stencil_chain | 23 July 2018, 15:29:07 UTC |
34873ac | Steven Johnson | 21 July 2018, 01:33:48 UTC | Horrible workaround for CpuFeatures | 21 July 2018, 01:33:48 UTC |
c79a47d | Steven Johnson | 21 July 2018, 01:24:07 UTC | Merge pull request #3140 from halide/srj-const Buffer::translated() should be const | 21 July 2018, 01:24:07 UTC |
a7585c3 | Steven Johnson | 20 July 2018, 22:02:01 UTC | Work around Windows errors for cpu_features | 20 July 2018, 22:02:01 UTC |