[Build] Propagate build option for CUDA minimal to TRT #20695
Conversation
|
Can you rebase it to main? |
gedoensmax
force-pushed
the
trt10_cuda_minial
branch
from
6566d23
to
c6d7bb0
Compare
|
@chilo-ms Sure, done. |
|
/azp run Linux CPU CI Pipeline, Linux CPU Minimal Build E2E CI Pipeline, Linux GPU CI Pipeline, Linux GPU TensorRT CI Pipeline, Linux OpenVINO CI Pipeline, MacOS CI Pipeline, ONNX Runtime Web CI Pipeline, onnxruntime-binary-size-checks-ci-pipeline, Linux QNN CI Pipeline |
|
/azp run Windows CPU CI Pipeline, Windows GPU CI Pipeline, Windows GPU TensorRT CI Pipeline, Windows ARM64 QNN CI Pipeline, orttraining-linux-ci-pipeline, orttraining-linux-gpu-ci-pipeline, orttraining-ortmodule-distributed, ONNX Runtime React Native CI Pipeline, Windows x64 QNN CI Pipeline |
|
/azp run Linux MIGraphX CI Pipeline, orttraining-amd-gpu-ci-pipeline |
|
Azure Pipelines successfully started running 2 pipeline(s). |
|
Azure Pipelines successfully started running 9 pipeline(s). |
1 similar comment
|
Azure Pipelines successfully started running 9 pipeline(s). |
| file(GLOB_RECURSE onnxruntime_providers_cuda_cu_srcs CONFIGURE_DEPENDS | ||
| "${ONNXRUNTIME_ROOT}/core/providers/cuda/*.cu" | ||
| "${ONNXRUNTIME_ROOT}/core/providers/cuda/*.cuh" | ||
| ) | ||
| else() | ||
| set(onnxruntime_providers_cuda_cu_srcs | ||
| "${ONNXRUNTIME_ROOT}/core/providers/cuda/math/unary_elementwise_ops_impl.cu" |
There was a problem hiding this comment.
It seems we need to include unary_elementwise_impl.cuh as well?
The definition of UnaryElementWiseImpl() is in that file and used by cuda::Impl_Cast<SrcT, DstT> which TRT EP will call to cast DOUBLE <- -> FLOAT or INT64 <-->INT32.
There was a problem hiding this comment.
The include should be added as the inlclude directory is set correctly. If you want I can add it to sources, but usually this is not the case in CMake I would say.
There was a problem hiding this comment.
yeah true, then no need to add it to sources
Description
Extend cuda minimal option to TRT provider, as with TRT 10 no linking to cuDNN is required anymore
.
Besides that with the new engine dump feature it is also possible to embed an engine in to an ONNX and not ship a builder lib.
In addition to that this has roughly the same deserialization time/session setup time that using TRT standalone has.
Motivation and Context