Граф коммитов

447 Коммитов

Автор SHA1 Сообщение Дата
Yuwei HU 4c553d24f0 add 2D reduction into tutorials (#204) 2017-07-01 10:13:59 -04:00
Tianqi Chen 5a7a056c3c [LANG/BUFFER] Change buffer arguments to match DLPack order, add scope (#203) 2017-06-30 10:34:07 -04:00
ziheng 7b8218512a [CONTRIB/NNPACK] Add NNPack Fully Connected Functions (#199)
* Add NNPack Fully Connected Inference

* Add NNPack fully_connected_output

* Fix lint

* Fix
2017-06-29 21:32:34 -04:00
ziheng 29d5ffbbbe [INTRINSIC] Add sqrt (#202)
* [INTRINSIC] Add sqrt

* [INTRINSIC] Expose on cpp
2017-06-27 19:53:57 -04:00
ziheng b8e02348e6 Fix] Avoid Directly Pass Python Context Object (#201) 2017-06-26 21:46:00 -04:00
Tianqi Chen 3cad2478f1 [OP] Initial Stucture of Op Library (#198)
* [OP] Initial start of op library

* add gtest
2017-06-23 17:34:12 -07:00
Haichen Shen 69aefaa388 [LANG] Add all and any in the python API (#196)
* [LANG] Add all and any in the python API

* compatible with python3
2017-06-22 16:18:12 -07:00
ziheng 7cc92ace2c [LANG] Expose tvm.cast (#195)
* [LANG] Expose tvm.cast

* Update

* Add unittest
2017-06-22 12:47:41 -07:00
Tianqi Chen 7c6a71bae8 [CODEGEN] Make codegen registerable (#193)
* [CODEGEN] Make codegen registerable

* fix llvm disbaled
2017-06-20 23:23:27 -07:00
Tianqi Chen 1400edac33 [IR] Include PrefetchIR (#189) 2017-06-18 21:53:13 -07:00
Tianqi Chen eaf0fde333 [LLVM] More optimized option, allow emit assembly (#187) 2017-06-17 22:46:14 -07:00
Tianqi Chen 5445061469 [CODEGEN] More storage alignment info aware generation (#186)
* [CODEGEN] More storage alignment info aware generation

* fix

* fix

* fix warning
2017-06-17 19:00:47 -07:00
Tianqi Chen 3b8e70ae03 [RUNTIME] Move device_api to include (#185)
* [RUNTIME] Move device_api to include

* fix doxygen

* fix device api

* fx
2017-06-16 18:10:15 -07:00
ziheng fcfec961a5 [TUTORIAL] Cross Compilation and RPC (#184)
* [TUTORIAL] Add tutorial for RPC

* [TUTORIAL] Update tutorial

* [TUTORIAL] Update tutorial

* trigger update

* [TUTORIAL] Improve build
2017-06-16 14:15:00 -07:00
Tianqi Chen d0041efdbc [MODULE] support load back of .ll file into llvm module (#183) 2017-06-15 11:17:47 -07:00
Tianqi Chen 204ad63bd4 [NNVM] Example NNVM integration. (#182) 2017-06-15 09:45:18 -07:00
Bing Xu b1402b379f [Compile] Fix compile issue with LLVM 8.0 (#181) 2017-06-12 20:28:19 -07:00
ziheng 2ce5277f75 [TUTORIAL] Update tvm.make.Select to tvm.select (#177) 2017-06-11 22:34:39 -07:00
ziheng f467f66e32 Support for Tuple Inputs of Reducer and ComputeOp (#175)
* Support for batch ComputeOp

* Support for batch ComputeOp

* Fix CrossThreadReduction

* Fix lint

* Add UpdateArray, remove support for batch reduce

* Tuple input support for reduce

* rfactor works with multiple reducer; support multiple reducers with different types

* Small fix

* Small fix

* Change return type of rfactor to Array<Expr>

* Fix lint

* Improve

* Add tutorial

* Improve tutorial

* Improve tutorial
2017-06-10 23:28:52 -07:00
Tianqi Chen ef50162b95 [MODULE/RUNTIME] Remove Precompile, simplify module (#174) 2017-06-05 22:00:22 -07:00
ziheng 84aeaf4803 Change Schedule Array constructor to static make method (#170)
* Change Schedule Array constructor to static make method

* Add CreateSchedule

* Add doc

* Change CreateSchedule to create_schedule at cpp side
2017-06-04 17:34:14 -07:00
Tianqi Chen 3bf7246929 [BUILD/CODEGEN] Allow combine multiple functions in build stage. (#169)
* [BUILD/CODEGEN] Allow combine multiple functions in build stage.

* Enhance code module

* fix compile
2017-06-04 15:47:18 -07:00
Tianqi Chen 5912ed034e [PERF/TIMER] Add builtin timing logic (#168)
* [PERF/TIMER] Add buildin timing logic

* fix lint
2017-06-03 09:06:28 -07:00
Tianqi Chen 46b4a91404 [PASS] Refactor build config, allow implicit unroll pragma (#167) 2017-06-02 11:32:53 -07:00
Tianqi Chen 86e5682482 [ARITH] More aggressive CSE during canonical simplify (#166) 2017-06-01 15:32:50 -07:00
ziheng 67a314c14c Improve makefile (#165)
* Improve makefile

* Fix
2017-05-31 17:20:31 -07:00
Haichen Shen 9962fdca39 [CODEGEN] Change default max_auto_unroll from 256 to 32 (#164) 2017-05-31 15:44:58 -07:00
Haichen Shen b9eb17c7f0 [BUILD] Clean the HalideIR submodule during the make clean (#163) 2017-05-31 15:07:57 -07:00
Tianqi Chen 46c56e28d3 [EXAMPLE/PYTHON] Improve extension type, add operator lib (#162) 2017-05-30 22:03:52 -07:00
Tianqi Chen 560463e9dc [BUILD] Enable RTTI of most part of library, example extension pkg. (#161) 2017-05-30 21:17:52 -07:00
Tianqi Chen 1efc4ca005 [CODEGEN/RUNTIME] Cross Compile Test (#160) 2017-05-25 09:41:24 -07:00
Tianqi Chen 1d0d876b7c [TEST/CI] 32bit compatibility and CI. (#159) 2017-05-24 19:06:32 -07:00
Tianqi Chen 28edb46133 [CI] Force doc build pass to mark success (#158) 2017-05-24 12:16:37 -07:00
Tianqi Chen bc81982081 [DOCS] Jenkins deployment of docs, add FAQ (#157) 2017-05-23 23:26:22 -07:00
Tianqi Chen 80f36b340f Remove linux from travis (#156) 2017-05-22 19:47:12 -07:00
Tianqi Chen 5896761659 [TESTS] Jenkins test flow (#152) 2017-05-22 16:04:31 -07:00
Tianqi Chen 1ea72d6f29 [DLPACK] Upgrade to the latest version (#150) 2017-05-21 13:16:12 -07:00
Yizhi Liu 050bc91bbb add tvm.select (#148) 2017-05-20 20:37:55 -07:00
Tianqi Chen 134c6ba335 [RUNTIME] RPC runtime that support run testing on remote device. (#147)
* [RUNTIME] RPC runtime that support run testing on remote device.

* Fix ctypes in OSX.

* fix lint
2017-05-20 09:05:12 -07:00
Hu Shiwen b7fe61191c use auto source_group (#146) 2017-05-18 09:17:47 -07:00
Hu Shiwen b47a12489c fix TVMRetValue move constructor not clear old value (#144)
* fix TVMRetValue move constructor not clear old value lead to repeat delete

* fix
2017-05-17 22:21:50 -07:00
ziheng 971e76713f Fix build status (#145) 2017-05-17 10:10:11 -07:00
Yizhi Liu 23d72ee094 fix doc examples & easy install (#143) 2017-05-16 08:50:22 -07:00
Tianqi Chen 03b09f749f [PASS] Improve SSA conversion, add forbid list in loop-par (#142) 2017-05-15 17:32:13 -07:00
Yizhi Liu 867ad378d2 check attach_stage & group in schedule.copy() (#139) 2017-05-14 13:49:58 -07:00
ziheng 52ad69fcd2 [FIX] Add CombineInternal<Mod> & Fix LoopPartition (#138)
* Add CombineInternal<Mod> & Fix LoopPartition

* Add check for path
2017-05-12 15:11:36 -07:00
Tianqi Chen 979623e56e [Tutorial] External Tensor Op (#137) 2017-05-12 11:25:51 -07:00
Tianqi Chen 553657eb74 [PYTHON] Support DLTensor compatible API (#136)
* [PYTHON] Support DLTensor compatible API

* optimize for common path
2017-05-11 22:05:23 -07:00
Tianqi Chen e3695cad85 [BUGFIX/PASS] Fix Vectorize with If condition (#135) 2017-05-09 23:19:44 -07:00
ziheng e9debc9be4 [PASS] Use likely tag & enable LoopPartition by default (#132)
* [PASS] Use likely tag & enable LoopPartition by default

* [PASS] Support thread_axis partition

* Take IfThenElse branch method

* [PASS] Insert branch at the innermost thread scope

* [PASS] Select candidates before trying to partition & add test for select

* [PASS] Clean code

* Fix

* Remove print & assert vectorize happens
2017-05-09 22:44:30 -07:00