mirror of
https://github.com/saymrwulf/onnxruntime.git
synced 2026-07-05 04:17:53 +00:00
7c17e33c07
### Description
CUDA inference speed heavily relies on Tensor Cores. To have tensor
cores achieve the optimal throughput they require the data layout to be
NHWC rather than NCHW.
### Motivation and Context
Especially for convolutional networks this is very important. I will
illustrate this using a very simple network:
```
import torch
import torch.nn as nn
class Net1(nn.Module):
def __init__(self):
super(Net1, self).__init__()
# 1 input image channel, 6 output channels, 5x5 square convolution
# kernel
self.m = nn.ModuleList([
nn.Conv2d(in_channels=8, out_channels=32, kernel_size=5, stride=1),
nn.Conv2d(in_channels=32, out_channels=64, kernel_size=3, stride=1),
nn.Conv2d(in_channels=64, out_channels=128, kernel_size=3, stride=1),
nn.Conv2d(in_channels=128, out_channels=128, kernel_size=3, stride=1, bias=False),
nn.Conv2d(in_channels=128, out_channels=128, kernel_size=3, stride=1, bias=False),
])
def forward(self, x):
for module in self.m:
x = module(x)
return x
if __name__ == "__main__":
dtype = torch.half
device = "cuda"
dummy_input = torch.randn(8, 8, 512, 512, dtype=dtype, device=device)
model = Net1().to(dtype=dtype, device=device)
input_names = ["input1"]
output_names = ["output1"]
torch.onnx.export(model, dummy_input, "test.onnx",
input_names=input_names, output_names=output_names)
```
I profiled the launch of `./build/RelWithDebInfo/onnxruntime_perf_test
-e cuda -I -q -t 5 test.onnx` using sys and nvtx ranges.
Current master launches below kernels:
If I add the introduced `-l` flag we see below kernels:
Notice the missing NCHW<>NHWC kernels per operation. The layout
optimizer introduced a transpose op as first and last op of the whole
network. The `op_generic_tensor_kernel` shows the bias used which should
also be optimized out next.
Measured across some very basic models:
| CUDA EP | **NCHW** [ms] | **NHWC** [ms] | Speedup |
|:------------------------|--------------------------------------:|-----------------------------------------:|------------------:|
| | -e cuda -t 5 -q | -e cuda -t 5 -q -l | |
| resnet101-v2-7_bs8_fp16 | 18.33 | 13.07 | 1.4 |
| resnet101-v2-7_bs8 | 21.8 | 12.06 | 1.81 |
| test | 102.07 | 73.62 | 1.39 |
Average speedup: 1.53
## Outlook
Next the mission will be to first write a templated unit test to check
for correctness of NHWC vs NCHW ops. After that we have to transition
more ops to measure perf improvements on a broader range of models.
Currently this is not easily possible as we can do not support all ops
in the NHWC domain.
---------
Co-authored-by: Tianlei Wu <tlwu@microsoft.com>
|
||
|---|---|---|
| .. | ||
| external | ||
| patches | ||
| tensorboard | ||
| adjust_global_compile_flags.cmake | ||
| CMakeLists.txt | ||
| CMakeSettings.json | ||
| codeconv.runsettings | ||
| deps.txt | ||
| deps_update_and_upload.py | ||
| EnableVisualStudioCodeAnalysis.props | ||
| gdk_toolchain.cmake | ||
| Info.plist.in | ||
| libonnxruntime.pc.cmake.in | ||
| nuget_helpers.cmake | ||
| onnxruntime.cmake | ||
| onnxruntime_codegen_tvm.cmake | ||
| onnxruntime_common.cmake | ||
| onnxruntime_compile_triton_kernel.cmake | ||
| onnxruntime_config.h.in | ||
| onnxruntime_csharp.cmake | ||
| onnxruntime_flatbuffers.cmake | ||
| onnxruntime_framework.cmake | ||
| onnxruntime_framework.natvis | ||
| onnxruntime_fuzz_test.cmake | ||
| onnxruntime_graph.cmake | ||
| onnxruntime_ios.toolchain.cmake | ||
| onnxruntime_java.cmake | ||
| onnxruntime_java_unittests.cmake | ||
| onnxruntime_kernel_explorer.cmake | ||
| onnxruntime_language_interop_ops.cmake | ||
| onnxruntime_mlas.cmake | ||
| onnxruntime_nodejs.cmake | ||
| onnxruntime_objectivec.cmake | ||
| onnxruntime_opschema_lib.cmake | ||
| onnxruntime_optimizer.cmake | ||
| onnxruntime_providers.cmake | ||
| onnxruntime_providers_acl.cmake | ||
| onnxruntime_providers_armnn.cmake | ||
| onnxruntime_providers_azure.cmake | ||
| onnxruntime_providers_cann.cmake | ||
| onnxruntime_providers_coreml.cmake | ||
| onnxruntime_providers_cpu.cmake | ||
| onnxruntime_providers_cuda.cmake | ||
| onnxruntime_providers_dml.cmake | ||
| onnxruntime_providers_dnnl.cmake | ||
| onnxruntime_providers_js.cmake | ||
| onnxruntime_providers_migraphx.cmake | ||
| onnxruntime_providers_nnapi.cmake | ||
| onnxruntime_providers_openvino.cmake | ||
| onnxruntime_providers_qnn.cmake | ||
| onnxruntime_providers_rknpu.cmake | ||
| onnxruntime_providers_rocm.cmake | ||
| onnxruntime_providers_tensorrt.cmake | ||
| onnxruntime_providers_tvm.cmake | ||
| onnxruntime_providers_vitisai.cmake | ||
| onnxruntime_providers_webnn.cmake | ||
| onnxruntime_providers_xnnpack.cmake | ||
| onnxruntime_pyop.cmake | ||
| onnxruntime_python.cmake | ||
| onnxruntime_rocm_hipify.cmake | ||
| onnxruntime_session.cmake | ||
| onnxruntime_snpe_provider.cmake | ||
| onnxruntime_training.cmake | ||
| onnxruntime_unittests.cmake | ||
| onnxruntime_util.cmake | ||
| onnxruntime_webassembly.cmake | ||
| precompiled_header.cmake | ||
| Sdl.ruleset | ||
| set_winapi_family_desktop.h | ||
| target_delayload.cmake | ||
| uwp_stubs.h | ||
| wcos_rules_override.cmake | ||
| winml.cmake | ||
| winml_cppwinrt.cmake | ||
| winml_sdk_helpers.cmake | ||
| winml_unittests.cmake | ||