eaea99c558
[TOPI] Example for convolution in GPU ( #212 )
...
* [TOPI] Example for convolution
* update conv ex
* fix submodule HalideIR
* update conv impl
* python3
* minor fix
* fix pylint error
* Add test code
* x
* fix
* fix
* move python helper function into topi.testing
* fix pylint
2017-07-18 19:38:40 -07:00
01cbc61a89
[API] Prefetch schedule supported ( #258 )
...
* prefetch interface added
* prefetch python comments modified. prefetch info data structure maintained.
* start injecting prefetches. first step (domain touch) implemented.
* domain touch tested.
* Prefetch ir_mutator and ir_visitor dispatch registered.
* modify domain touched from passing a func_ref to passing a tensor
* modify domain touched from passing a func_ref to passing a tensor
* modify Tensor copy to Tensor ref
* temp commit for rebase
* debug info removed, typo fixed, ready to rebase
* prefetch flatten test add!
* roll back builtin functions to side effect functions
* lint error fixed!
* add cache line size to storage flatten argument
* forgot modifications add
* change code style to dmlc-like; get rid of can_prove, use manually compute instead
* python lint error fixed
* modify instrinsic name to pass tests
* [TEST] get rid of str(), replace them by accessing attributes
* change map to list comprehension
* redundant numpy import removed
2017-07-18 09:40:54 -07:00
7b6427e3a8
[RPC] Allow RPCServer to run without decorator ( #257 )
2017-07-17 13:40:42 -07:00
f12e781c92
fix build in windows ( #256 )
2017-07-17 10:24:51 -07:00
8e51af2f94
[PASS] CombineContextCall ( #255 )
2017-07-16 22:48:31 -07:00
36f20b5472
[PASS] Add storage alignment info to heap allocated data ( #254 )
2017-07-16 21:23:33 -07:00
204e9cb429
[EXECUTOR] Fix bug and improve ( #252 )
...
* [EXECUTOR] Fix bug and improve
* [EXECUTOR] Enhance test case
2017-07-16 14:54:08 -07:00
f433373de4
[CODEGEN] Generate main compute function separately with alias info ( #253 )
2017-07-16 14:16:09 -07:00
8d241b9d86
[RPC] Allow back pressure from writer ( #250 )
...
* [RPC] Allow backpressure from writer
* fix
* fix
2017-07-15 18:23:20 -07:00
c6d4f5af20
[EXECUTOR] Enable load executor remotely ( #245 )
...
* [EXECUTOR] Enable load executor remotely
* [EXECUTOR] Pipeline
* Pass bytearray directly
* Enable load dynamic library in rpc_server.py
* Fix
* lint
* Return Module from remote side directly
* Remove unused header file
* Fix
* fix
2017-07-15 13:03:30 -07:00
3f3bf29d5a
[DOC] Make range related function consistent ( #249 )
2017-07-15 10:43:20 -07:00
2ab0bfb589
[RENAME] nvcc_compiler->nvcc, cc_compiler->cc, metal_compiler->xcode ( #248 )
2017-07-15 09:41:44 -07:00
d3efd7fc35
[WIP][Frontend] Scala/Java package ( #176 )
...
* JVM package skeleton
* [JVM] link libtvm.so and list function names
* [JVM] Function & NDArray skeleton
* [JVM] TVMFuncCall in JNI
* [JVM] handle string arg in TVMFuncCall
* [JVM] get module function
* [JVM] entry function for Module
* [JVM] construct Module from function return value
* [JVM] TVMContext, TVMArray attributes
* [JVM] NDArray from / to java array
* [JVM] load so and compute on cpu
* [JVM] move PackedFunc to individual modules
* [JVM] assembly package & native library loader
* [JVM] unit test & codestyle check settings
* [JVM] NDArray from & to different dtypes
* [JVM] NDArray from native double array. Add linux-cpu profile.
* [JVM] modify Makefile
* [JVM] add linux-x86_64-gpu profile
* [tvm4j] delay load libtvm_runtime.so
* [tvm4j] refactor to pure java
* [tvm4j] remove scalastyle-config.xml
* [tvm4j] remove link HalideIR, remove Shape, remove scala binary versions
* [tvm4j] only allow convert from/to same type array
* [tvm4j] make NDArray api more readable
* [tvm4j] refactor for c api
* [tvm4j] add Jenkins tests
* [tvm4j] fix duplicate Dockerfile cmd
* [tvm4j] fix ut script filename
* [tvm4j] add module load tests
* [tvm4j] add javadoc, remove types package
* [tvm4j] fix test script
* [tvm4j] remove ut temp dir
* [tvm4j] fix missing package types
* [tvm4j] java code style check
* [tvm4j] fix java lint
* [tvm4j] downgrade checkstyle plugin for JDK7
* [tvm4j] add stylecheck in jenkins tests
* [tvm4j] specify source file encoding
* [tvm4j] lazy init function; add Function.call() api; allow manully release Module,NDArray,Function
* [tvm4j] fix ModFree
* [tvm4j] cache Function in API
2017-07-14 23:04:39 -07:00
86ff24abb5
[DOC] Fix doxygen comments ( #247 )
2017-07-14 17:33:40 -07:00
f33147c229
Update index.rst
2017-07-14 10:15:56 -07:00
8ca7576943
[INTRIN] prefetch support ( #246 )
...
* [INTRIN] prefetch support
* lint
* add buildin
2017-07-14 09:53:00 -07:00
7bcb3f538b
[REFACTOR] collections->container, RPC returns func, time_evaluator r… ( #244 )
...
* [REFACTOR] collections->container, RPC returns func, time_evaluator returns struct
* fix executor
2017-07-13 22:30:22 -07:00
c324494f1a
[RUNTIME][RPC] Change RPCServer to Event Driven Code ( #243 )
...
* [RUNTIME][RPC] Change RPCServer to Event Driven Code
* fix
2017-07-13 18:44:33 -07:00
f4d1dddbcd
[EXECUTOR] Save/Load Params ( #242 )
...
* [EXECUTOR] Save/Load Params
* [EXECUTOR] Improve Save/Load, fix Makefile
* [EXECUTOR] Make save independent with executor
2017-07-12 16:25:36 -07:00
9c36d9f044
[RUNTIME] Fix Metal runtime compile ( #241 )
2017-07-11 16:32:07 -07:00
9c954adab3
[DOC] Fix c++ doc build ( #240 )
2017-07-10 19:33:04 -07:00
0a07411d54
[JS][WEB][BACKEND] Javascript(webassembly) backend. ( #239 )
2017-07-10 18:23:47 -07:00
619e529a1c
Enable identity layout ( #238 )
2017-07-10 13:23:08 -07:00
eb3fc6c630
[CI] Add JVM Env ( #237 )
...
* [CI] Add JVM Env
* add update
2017-07-10 09:18:06 -07:00
321218650b
[PASS] Layout transform pass ( #233 )
...
* [PASS] Layout transform pass
* Fix according to comment
* Fix
2017-07-09 19:20:18 -07:00
adc06e6fbb
[RUNTIME][ABI] Remove TVMValue as argument, use address ( #236 )
2017-07-09 18:02:56 -07:00
cd29c18c34
[RUNTIME][ABI] Flat structure arguments ( #232 )
2017-07-08 16:26:36 -07:00
2dec0510c5
[TEST][TOPI] of depthwise_conv2d ( #230 )
...
* test of depthwise_conv2d
* fix nose test error
* python3 fix
2017-07-08 16:18:45 -07:00
55ba9cb830
[TEST] Add scipy to test dep ( #231 )
2017-07-08 11:14:28 -07:00
9d84cb0777
[RUNTIME] Add workspace pool ( #229 )
...
* [RUNTIME] Add workspace pool
* fix doc
* fix the free list
* avoid zero size
2017-07-07 23:19:57 -07:00
5cdc860492
[TAG] Fix signature of decorated function ( #228 )
...
* [TAG] Fix signature of decorated function
* Add dep
2017-07-07 16:02:35 -07:00
72fcd4e6ee
[RUNTIME] Add System Lib ( #227 )
...
* [RUNTIME] Add System Lib
* lint
* lint
* fix compile
2017-07-07 12:16:37 -07:00
b759d0f3d4
Small fix of the Depthwise Convolution example in python3 ( #224 )
...
* fix for python3
fix for python3
* Update depthwise_conv2d_map_test.py
remove sys.append
2017-07-07 10:21:57 -07:00
bfe6d95eb1
[DOC/TOPI] Add API doc for topi ( #226 )
...
* [DOC/TOPI] Add API doc for topi
* fix lint
2017-07-07 09:44:43 -07:00
5fec5c5abe
[TOPI/TEST] Add Testcase folder for TOPI ( #225 )
2017-07-07 08:50:26 -07:00
4580e690b8
[TOPI] Example for depthwise convolution ( #197 )
...
* first commit
* move to topi/recipe
* refactor, almost rewrite
* 2-D sum reduction; implement SAME pad; improve schedule
* add util.py; separate test script
* conv + bn + relu fusion
* auto fusion
* separate declare and schedule; using op tag
* divide large image into blocks
* move to topi; improve blocking schedule
* restructure
* add doc
* using time_evaluator
2017-07-07 08:06:56 -07:00
825566ccff
[SCHEDULE] tensorize ( #223 )
2017-07-06 13:56:39 -07:00
28120f554c
[C API] Make DSL API registerable, add copy from/to raw bytes ( #222 )
...
* [C API] Make DSL API registerable, add copy from/to raw bytes
* fix cython
2017-07-06 11:34:08 -07:00
0a19b16ac0
[CODEGEN/PASS] add restricted, alignment option ( #221 )
...
* [CODEGEN/PASS] add restricted, alignment option
* fix lint
* Fix the alloca
2017-07-05 23:21:42 -07:00
00506a62e7
[IR] Add body to AssertStmt ( #220 )
...
* [IR] Add body to AssertStmt
* fix lint
2017-07-05 17:55:34 -07:00
c9da7254ba
[TAG] Add tvm.tag module for tagging operator ( #217 )
...
* [TAG] Add op_tag module for tagging operator
* Fix accroading to comments
* Add example
* Add into doc
* Add --fix-missing for docker
2017-07-04 22:27:48 -07:00
8a66ac230f
[PASS/OP/REFACTOR] IRDeepCompare, isolate computeop part, allow fuzzy bind ( #218 )
2017-07-04 21:53:15 -07:00
8e2ea2c483
[APP] Improve GraphExecutor ( #216 )
...
* Remove 'final' in GraphExecutor for extension
* Dynamic num of inputs/outputs for tvm_op
2017-07-04 14:09:06 -07:00
4bb3c35a05
[REFACTOR/PASS] Formalize argument bind and match util ( #214 )
...
* [REFACTOR/PASS] Formalize argument bind and match util
* grammar
2017-07-04 10:30:24 -07:00
3c19159587
[DOCS] fix doc builder ( #213 )
...
* [DOCS] fix doc builder
* fix
* fix
* fix doc builder
2017-07-03 21:01:42 -04:00
623748a698
Switch off global barrier detection by default ( #211 )
2017-07-03 16:31:50 -04:00
d0cbf94f0b
[REFACTOR] examples->apps ( #210 )
2017-07-03 14:33:55 -04:00
34d2aae351
[BUFFER/REFACTOR] Buffer byte_offset-> elem_offset, add buffer_bind_scope ( #209 )
2017-07-03 14:13:52 -04:00
b0e41b9ae5
[CODEGEN] Concise typecast for threadIdx ( #208 )
2017-07-03 01:55:32 -04:00
bf97724b7a
[FIX] Fix allocate size^2 in graph_executor ( #207 )
2017-07-03 01:45:19 -04:00
Haichen Shen
Jian Weng
Tianqi Chen
Xingjian Shi
ziheng
Yizhi Liu
Yuwei HU