Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

merge to newest code #2

Merged
merged 166 commits into from
Aug 3, 2021
Merged
Changes from 1 commit
Commits
Show all changes
166 commits
Select commit Hold shift + click to select a range
29e958d
[TIR][TVMScript] specialize (#8354)
Hzfengsy Jul 1, 2021
6f600f1
[Refactor] Remove dead code from depthwise_conv2d for Intel graphics …
echuraev Jul 2, 2021
ab01abc
[BugFix][Relay] Fix type relation for batch_matmul (#8376)
lygztq Jul 2, 2021
22204be
fix keras install (#8391)
Jul 2, 2021
7b898d0
Fix np.int and np.float usage in the tree. (#8389)
Jul 2, 2021
354d996
Add missing annotation for requires_gpu in test_topi_dense.py Require…
Jul 2, 2021
970aeff
Add "operator" style to Model Library Format (#8072)
areusch Jul 2, 2021
2e3d617
macOS is now supported (#8396)
hogepodge Jul 3, 2021
7e3f068
[microTVM] Add Nucleo stm32l4r5zi board to zephyr (#8386)
mehrdadh Jul 3, 2021
e19e979
[Torch] Remove unused conversion (#8397)
masahi Jul 3, 2021
e32d47e
[Arith] Inverse affine map (#8384)
vinx13 Jul 4, 2021
a00d211
Actually add Compute Library tests to the Jenkins File (#8394)
Jul 4, 2021
d17f753
Support aten::flip (#8398)
delldu Jul 4, 2021
d3fc562
[Relay][TOPI] Resize 1D (#8346)
Jul 5, 2021
6a3d950
[Docs] Fix for broken link in apps for wasm-standalone dir (#8045)
mikepapadim Jul 5, 2021
ec47129
add aten::masked_fill_ in pytorch frontend (#8403)
hsiehjackson Jul 5, 2021
6bcad2e
fix storage rewrite index remap (#8338)
wrongtest-intellif Jul 6, 2021
bbfc52c
Cleanup more uses of np.bool and np.int. (#8399)
Jul 6, 2021
bd5cd9f
[Fix] Update stale relay.Module API in docs/comments (#8411)
YuchenJin Jul 7, 2021
2628179
[ONNX] Wrap 'If' if it has multiple outputs (#8385)
karljang Jul 7, 2021
a4775c2
[DOCS] Add docs for Pass Instrument (#8220)
chiwwang Jul 7, 2021
8fb4cdf
Revert "Actually add Compute Library tests to the Jenkins File (#8394…
Jul 7, 2021
9c66587
Refactor the compile engine into a cleaner interface. (#7518)
csullivan Jul 8, 2021
0b39af7
[Relay] Fix index order in conv2d computation for Arm CPU. (#8361)
AnastasiaStulova Jul 8, 2021
e3e03df
[microTVM] Add fixture to zephyr test (#8393)
mehrdadh Jul 8, 2021
e7c5349
[Relay] Add support of conv2d with NHWC for Mali (#8422)
AnastasiaStulova Jul 9, 2021
ee65ab7
[PyLint] Minor updates to pass pylint locally. (#8424)
Lunderberg Jul 9, 2021
53cb8aa
[Frontend] Check LLVM enabled/installed (#8414)
kueitang Jul 9, 2021
4b67e9d
[Bug] Fix x86 dense schedule extern ops (#8420)
comaniac Jul 9, 2021
e934b7e
[Doc] Fix Relay pattern rewrite (#8425)
comaniac Jul 9, 2021
0fa4396
[CUDA] dense_tensorcore/batch_matmul_tensorcore support int8/int4 (#8…
wyc-ruiker Jul 9, 2021
683c5eb
[Arith] Simplify MatchFusePattern in InverseAffineMap (#8427)
vinx13 Jul 9, 2021
e51f5bb
[TOPI] Bugfix for topi.prod (#8416)
leeexyz Jul 9, 2021
f692fc7
Improve XGBTuner document (#8428)
juierror Jul 9, 2021
513fcf4
[TVMSCRIPT] TVMScript Parser support BufferSlice indices (#8408)
Beya2019 Jul 9, 2021
6141cac
Replace RuntimeError in _lookup_task with deferred error. (#8421)
mdw-octoml Jul 9, 2021
bdfbc86
fix flaky TF test (#8431)
Jul 9, 2021
59b204d
[microTVM][RVM] Fix clock skew on virtualbox (#8395)
mehrdadh Jul 10, 2021
972d7b5
[Relay] Add support of conv2d with NHWC for Bifrost (#8430)
AnastasiaStulova Jul 10, 2021
3a9a388
fix wrong log of tir pass VerifyMemory (#8445)
Jul 11, 2021
c3558a1
[Relay to onnx conversion fixes][Pool, Pad] (#8435)
schilkunda-amba Jul 12, 2021
3424005
[Relay to onnx conversion][New ops] (#8436)
schilkunda-amba Jul 12, 2021
1d7a9e9
[ROCM] Fix undefined symbols by adding library (#8446)
mvermeulen Jul 12, 2021
15bdf28
Fix address and port reported by android_rpc to tracker (#8405)
echuraev Jul 12, 2021
c81d533
[Bugfix] Fix broadcast type func with incomplete type (#8438)
comaniac Jul 12, 2021
cd5a20e
[COMMUNITY] @junrushao1994 -> PMC (#8450)
zhiics Jul 12, 2021
62adc77
[MyPy] Extend type checking and annotation for TIR (#8429)
mikepapadim Jul 13, 2021
d043cb9
[BugFix][TOPI] Fix the integer overflow problem of the scatter_nd op.…
zhuwenxi Jul 13, 2021
807373c
Add qnn batch_matmul operator (#8401)
elvin-n Jul 13, 2021
80f48c7
[microTVM] Fix Stack Size Issue for Zephyr AOT Demo on Physical Hardw…
mehrdadh Jul 13, 2021
957cc12
[Relay] Modify create_executor to pass params (#8418)
mikepapadim Jul 13, 2021
d67514b
[PROFILING] Use PAPI to collect hardware performance counters on CPU …
tkonolige Jul 13, 2021
136f218
[Relay][ONNX] Batch_matmul to dense optimization (#8440)
ymwangg Jul 13, 2021
f62917e
[TOPI] Add support for arbitrary dtypes to CSRMV and CSRMM (#8437)
tkonolige Jul 13, 2021
1a26733
[Refactor] Enforce attaching storage scope to PointerType (#8366)
masahi Jul 13, 2021
73b38e8
[Fix] Explicitly retain `__hash__` of `StringImm` (#8449)
UniverseFly Jul 13, 2021
5c1a1cf
[CUDA] Improve injective schedule to enable half2 (#8457)
comaniac Jul 14, 2021
c16d61b
[Fix] Remove unused variable in GraphExecutorCodegen (#8465)
YuchenJin Jul 14, 2021
a425d26
[Docs] Corrected typo in googletest build instructions. (#8459)
Lunderberg Jul 14, 2021
f15be8b
[RPC] Fix cpp_rpc connection to rpc_tracker (#8388)
echuraev Jul 14, 2021
e1b3ff4
[Relay][Frontend][ONNX] Add ConvInteger support. (#8456)
Jul 14, 2021
bbba5da
[COMMUNITY] comaniac added as new PMC member (#8470)
tqchen Jul 15, 2021
1a9bcc5
[UnitTests] Minor fixes to unit tests for cudnn/vulkan targets (#8462)
Lunderberg Jul 15, 2021
29f789f
[Codegen] Remove compile_enginer header (#8471)
YuchenJin Jul 15, 2021
11c5b6d
[Relay][Onnx][Frontend] Add RandomUniform converter and tests to onnx…
Jul 15, 2021
8a8c9b2
[AMP] Add default op attribute registration to __init__.py (#8460)
AndrewZhaoLuo Jul 15, 2021
bd88ee2
Fix auto-scheduling after 9c6658721 (#8478)
echuraev Jul 15, 2021
ce15ca6
[Relay][Frontend][ONNX] Allow importing models with malformed Loop no…
Jul 15, 2021
63c6df8
FoldScaleAxis became non-recursive (#8325)
d-smirnov Jul 15, 2021
ea8b69b
DeviceType enums match dlpack (#8407)
schell Jul 16, 2021
5bb01ef
fix typo (#8484)
Raghav-Chakravarthy Jul 16, 2021
c263f22
[TVMC][FIX] Compiler supports input with a slash (#8481)
leeexyz Jul 16, 2021
2b57907
fix minor misspelling (#8476)
mshr-h Jul 16, 2021
cba9cf3
[VM] Fix the shape function of conv nhwc (#8480)
jcf94 Jul 16, 2021
c8b9900
[BYOC] add multi functions support in partition pass (#8464)
Jul 16, 2021
7388715
Fix _get_yolo_detections (#8477)
apivovarov Jul 16, 2021
ff425c1
apps: microtvm: Disable `CONFIG_FPU ` for Zephyr runtime (#8055)
microbuilder Jul 16, 2021
44f1baf
[TVMSCRIPT] Support tir.abs node in tvm script (#8488)
Beya2019 Jul 17, 2021
c95d16e
[Frontend][Tensorflow2] Stridedslice and concat_v2 fix (#8483)
srinidhigoud Jul 17, 2021
5ecd6cd
[VM] Allow serialization of function attrs which are strings (#8485)
tkonolige Jul 17, 2021
7456cfc
[AutoTVM] Re-enable `ref_input` (#8113)
Tantalus13A98B5F Jul 17, 2021
0545962
[Bugfix] [tir] do not simplify 'Any() - Any()' to 0 (#8266)
hgt312 Jul 17, 2021
2c61afa
Switch from CompileEngine to TECompiler in Interpreter (#8486)
mbs-octoml Jul 17, 2021
d3db5d6
Fix dynamic batching when use_implicit_batch=False (#8461)
Jul 17, 2021
1534983
[ARITH] fix zero iter bug in arith (#8494)
spectrometerHBH Jul 18, 2021
f64ddff
Add missing shape functions for relay.nn operations (#8489)
rijulg Jul 19, 2021
4b2ccde
src/runtime/module.cc (#8496)
AndrewZhaoLuo Jul 19, 2021
6d88bdd
Update Docker CI (#8193)
Jul 19, 2021
2d1847c
[ETHOS-N] Re-enabled tests and updated module hashes (#8498)
leo-blonk Jul 19, 2021
2b74f97
Keep CODEOWNERS file up to date. (#8500)
tqchen Jul 20, 2021
d5818b4
[Frontend, pytorch] Vc/pytorch lstm (#8447)
vvchernov Jul 20, 2021
ca28dff
Rename runtime-config to executor-config and add documentation for Mo…
areusch Jul 20, 2021
1141709
Fix 8093, Enhance Buffer Index Simplify (#8204)
Johnson9009 Jul 20, 2021
1a1be09
[Refactor] Remove scope attribute from Buffer class (#8463)
masahi Jul 20, 2021
dbdfc44
Enable ONNX tests that needed onnxruntime 1.7.0 (#8502)
Jul 21, 2021
78142b6
Organize the CodeOwners file: (#8512)
tqchen Jul 21, 2021
eacc2cb
[TIR] Bugfix for zero number arguments tir functions. (#8515)
ZihengJiang Jul 21, 2021
e8c7f67
[Relay] Fix bug in test_op_level3 (#8508)
CaptainDuke Jul 21, 2021
2df0854
[TensorIR][M2a] Fuse, Split (#8467)
jinhongyii Jul 21, 2021
789ab1f
[Relay] Support resize in the ONNX conversion (#8455)
schilkunda-amba Jul 21, 2021
18491ea
update qemu install (#8518)
mehrdadh Jul 22, 2021
45497bd
[Topi][UnitTests] Parameterize conv2d and depthwise_conv2d tests (#8433)
Lunderberg Jul 22, 2021
e95f10f
[CUDA] Initial support for dynamic shared memory (#8466)
masahi Jul 22, 2021
59e96e0
[microTVM][Cortex-R5] Add zephyr cortex-r5 board to Zephyr (#8519)
mehrdadh Jul 22, 2021
07243a8
[TVMSCRIPT]Fix script printters StructuralEqual check failed (#8499)
Beya2019 Jul 22, 2021
bce0db5
[PROFILING] Add json output to profiling reports (#8503)
tkonolige Jul 24, 2021
e664ef0
[PRINTER] Fix the repeatitive cast in scripr printing (#8531)
tqchen Jul 24, 2021
8ab2074
[Frontend, Tensorflow2] Added support for TensorList ops (#8454)
Jul 25, 2021
18171e4
[CMake] Split out libinfo.cc into a separate target. (#8520)
Lunderberg Jul 25, 2021
3445532
[RUNTIME] Fix TypeKey2Index when for root Object (#8547)
tqchen Jul 25, 2021
9c63f4f
[TFLite] Mimic the TFLite's 2.4 reader's behaviour (#8538)
ekalda Jul 26, 2021
fc870f8
Remove unused variable in topi cpp test (#8549)
senychen Jul 26, 2021
ee207fd
[RPC] Add explicit type cast to print. (#8524)
Jul 26, 2021
a492db8
[Bugfix] Visit each input param of the function in ExprVisitor visit_…
JoeyChou-SiMa-ai Jul 27, 2021
3b7aed3
[FFI] Specifically check handle for recursion during shutdown (#8548)
tqchen Jul 27, 2021
f2354ee
Add a `--context-path` for build.sh, allowing to test Dockerfiles (#8…
leandron Jul 27, 2021
8954968
Fix #8510 (#8511)
Shpionus Jul 27, 2021
496f10b
[Frontend][TENSORFLOW] Add support for unpack with dim 0 after tensor…
Jul 28, 2021
5a96a61
[Bugfix] fix android rpc app undefined reference problem (#8530)
hope51607 Jul 28, 2021
8d9df2d
[TensorRT, BYOC] Handling a corner case in TRT RemoveDropout pass (#8…
rohanmukh Jul 28, 2021
a17ee9f
[BUGFIX] fix illegal memory access bug in reduce op schedule by const…
ZQPei Jul 28, 2021
2fa26b3
Re-enable Compute library tests. (#8573)
Jul 29, 2021
774c0e5
[Test] Fix AutoScheduler test to cover Conv2D Winograd (#8539)
sunjiweiswift Jul 29, 2021
fcbd2b6
fixenhance robustness of DefuseOps (#8564)
ganler Jul 29, 2021
6b2cbfe
[Contrib] Added default non-verbose to download_testdata(), pass to d…
Lunderberg Jul 29, 2021
92f9558
[Coreml] Fix Coreml Input Shape Handling (#8562)
AndrewZhaoLuo Jul 29, 2021
74715f4
[AutoScheduler] Fix task extraction with TE compiler (#8560)
comaniac Jul 29, 2021
b6b8269
add support for softmax and log_softmax with MIOpen (#8543)
altanh Jul 29, 2021
a993321
[TOPI][CUDA] minor change on assert statement in conv2d_NCHWc_int8.cu…
zotanika Jul 29, 2021
83ce7fe
[TOPI] Fix `nn.pool*d` issue with 'vectorize' function and add unit t…
MarioPeric-SiMa-ai Jul 29, 2021
00ad44e
[Bugfix] Fix #8536 Get Target When Heterogeneous Execution (#8537)
Johnson9009 Jul 29, 2021
df06c58
[Bugfux] wasm32-standalone app repaired (#8563)
ya0guang Jul 29, 2021
cb395ff
Disable pip cache when creating Docker images (#8575)
leandron Jul 29, 2021
850abb0
[TOPI] Add transpose_a/b & dynamic shape support for batch matmul (#8…
jcf94 Jul 29, 2021
e39204e
[Bugfix] Preserve IRModule type definition and imports in NameMangleE…
Jul 29, 2021
66b3cc9
[TIR] cast disparate floating point types for binary ops (#8517)
AndrewZhaoLuo Jul 29, 2021
bef7bf9
[Refactor] Remove AttrStmt with storage_scope key (#8516)
masahi Jul 29, 2021
88dd31b
[VM] Bug fix for numpy scalar input in vm (#8553)
jcf94 Jul 30, 2021
97f147c
[Torch] Reduce testing time of LSTM tests (#8583)
vvchernov Jul 30, 2021
8148028
[FIX][CI] hotfix check_grad perf regression (#8581)
altanh Jul 30, 2021
df96cba
[Vulkan] Prioritize discrete GPUs as device_id=0. (#8588)
Lunderberg Jul 30, 2021
22c7d61
speed up reference resize kernel (#8592)
Jul 30, 2021
ff71773
Delete pytest-results as part of CI workspace preparation. (#8594)
areusch Jul 30, 2021
b7df355
use sizevar when convert any to tir (#8555)
lygztq Jul 30, 2021
12e805f
Fix storage_access not visiting else branch (#8525)
masahi Jul 30, 2021
2b94a7e
[AOT][Stack Allocator] Fix Initial Memory Misalignment (#8487)
mehrdadh Jul 30, 2021
c8a892b
[Vulkan] Rewrote PointerValueTypeRewrite transform (#8528)
Lunderberg Jul 30, 2021
5012462
[TensorIR][M2a] Reduction Factoring (RFactor) (#8544)
MasterJH5574 Jul 31, 2021
2a8950b
[TensorIR] Support for match_buffer from subregion (#8585)
Hzfengsy Jul 31, 2021
7d8a774
[VTA] Recover rpc server support (#8604)
tqchen Jul 31, 2021
28de742
[Refactor] Unify the shared pass prefix between vm and graph (#8526)
YuchenJin Jul 31, 2021
4b67dac
[CUDA] Support multiple TIR-level dynamic shared memory allocations (…
masahi Jul 31, 2021
887324f
[TOPI][CUDA] Improve the performance of scatter_nd (#8479)
zhuwenxi Aug 1, 2021
9f29e2a
[BUILD] Add caching to CMake (#8373)
electriclilies Aug 1, 2021
7653972
[Meta Schedule][M3a] Instruction and Trace (#8615)
junrushao Aug 2, 2021
3113388
Add support for AOT in external code generation tests (#8591)
Mousius Aug 2, 2021
b3e832a
[CI] Fix global pip cache disable change (#8590)
leandron Aug 2, 2021
55916a2
[DOCS] Fix scipy docs inv (#8619)
tqchen Aug 2, 2021
47a5a3a
[runtime] Remove unused parameter. (#8580)
huajsj Aug 2, 2021
49d5879
Introduce --interface-api={c,packed} parameter (#8280)
Mousius Aug 2, 2021
f656a22
Docker env for Arm® Ethos™-U55 Port (#8514)
grant-arm Aug 2, 2021
9b97bad
[FIX,PROFILING] Add USE_PAPI configuration to config.cmake (#8567)
tkonolige Aug 2, 2021
3cb0e93
[Fix] Fix a typo in include/tvm/ir/function.h (#8617)
Aug 3, 2021
38fe522
[Relay][Quantization] Extend FakeQuantizationToInteger to more ops (#…
Aug 3, 2021
804ba27
Fix test_external_codegen, broken by #8591 (#8630)
areusch Aug 3, 2021
09e234d
[DOCS] TVM install addenda for M1 Macs (#8568)
AndrewZhaoLuo Aug 3, 2021
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
[Codegen] Remove compile_enginer header (apache#8471)
* [Codegen] remove compile_enginer header

* fix lint
YuchenJin authored Jul 15, 2021

Verified

This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
commit 29f789fcc07e5e51072faf88a26a3c41292809c7
35 changes: 17 additions & 18 deletions src/relay/backend/graph_executor_codegen.cc
Original file line number Diff line number Diff line change
@@ -36,7 +36,6 @@
#include <string>
#include <vector>

#include "compile_engine.h"
#include "te_compiler.h"
#include "utils.h"

@@ -184,7 +183,7 @@ class GraphOpNode : public GraphNode {
*/
class GraphExecutorCodegen : public backend::MemoizedExprTranslator<std::vector<GraphNodeRef>> {
public:
GraphExecutorCodegen(runtime::Module* mod, const TargetMap& targets) : mod_(mod) {
GraphExecutorCodegen(runtime::Module* mod, const tec::TargetMap& targets) : mod_(mod) {
targets_ = targets;
}

@@ -222,20 +221,20 @@ class GraphExecutorCodegen : public backend::MemoizedExprTranslator<std::vector<
device_context_map.insert({expr, dev});
}

auto lowered_module = tec::LowerTE(mod, targets_, device_context_map, memory_plan_, mod_name_,
[this](Function func) {
// We need to maintain the constant map for external
// functions so we pass this processing function which
// allows us to process each function as we lower it.
if (func->GetAttr<String>(attr::kCompiler).defined()) {
UpdateConstants(func, &params_);
}

// TODO(@areusch, @jroesch): We should refactor this to
// execute as a further pass, instead writing data to the
// lowering process directly.
UpdateFunctionMetadata(func, this->function_metadata_);
});
auto lowered_module = tec::LowerTE(
mod, targets_, device_context_map, memory_plan_, mod_name_, [this](Function func) {
// We need to maintain the constant map for external
// functions so we pass this processing function which
// allows us to process each function as we lower it.
if (func->GetAttr<String>(attr::kCompiler).defined()) {
UpdateConstants(func, &params_);
}

// TODO(@areusch, @jroesch): We should refactor this to
// execute as a further pass, instead writing data to the
// lowering process directly.
tec::UpdateFunctionMetadata(func, this->function_metadata_);
});

function_metadata_.Set(runtime::symbol::tvm_module_main, lowered_module.main_func_info);
auto main_module = lowered_module.main_module;
@@ -580,7 +579,7 @@ class GraphExecutorCodegen : public backend::MemoizedExprTranslator<std::vector<
/*! \brief variable map */
std::unordered_map<const Object*, std::vector<GraphNodeRef>> var_map_;
/*! \brief target device */
TargetMap targets_;
tec::TargetMap targets_;
/*!
* \brief parameters (i.e. ConstantNodes found in the graph).
* These are take as inputs to the GraphExecutor.
@@ -609,7 +608,7 @@ class GraphExecutorCodegenModule : public runtime::ModuleNode {
<< "runtime::Module mod and Map<int, Target> targets";
void* mod = args[0];
Map<Integer, tvm::Target> tmp = args[1];
TargetMap targets;
tec::TargetMap targets;
for (const auto& it : tmp) {
auto dev_type = it.first.as<tir::IntImmNode>();
ICHECK(dev_type);