使用 Cudagraphs 進行 Torch 匯出

此互動式腳本旨在概述在 ir=”dynamo” 路徑中使用 Torch-TensorRT Cudagraphs 整合的流程。該功能在 torch.compile 路徑中的運作方式類似。

導入和模型定義

import torch
import torch_tensorrt
import torchvision.models as models

使用預設設定透過 torch_tensorrt.compile 進行編譯

# We begin by defining and initializing a model
model = models.resnet18(pretrained=True).eval().to("cuda")

# Define sample inputs
inputs = torch.randn((16, 3, 224, 224)).cuda()

# Next, we compile the model using torch_tensorrt.compile
# We use the `ir="dynamo"` flag here, and `ir="torch_compile"` should
# work with cudagraphs as well.
opt = torch_tensorrt.compile(
    model,
    ir="dynamo",
    inputs=torch_tensorrt.Input(
        min_shape=(1, 3, 224, 224),
        opt_shape=(8, 3, 224, 224),
        max_shape=(16, 3, 224, 224),
        dtype=torch.float,
        name="x",
    ),
)

使用 Cudagraphs 整合進行推論

# We can enable the cudagraphs API with a context manager
with torch_tensorrt.runtime.enable_cudagraphs(opt) as cudagraphs_module:
    out_trt = cudagraphs_module(inputs)

# Alternatively, we can set the cudagraphs mode for the session
torch_tensorrt.runtime.set_cudagraphs_mode(True)
out_trt = opt(inputs)

# We can also turn off cudagraphs mode and perform inference as normal
torch_tensorrt.runtime.set_cudagraphs_mode(False)
out_trt = opt(inputs)

# If we provide new input shapes, cudagraphs will re-record the graph
inputs_2 = torch.randn((8, 3, 224, 224)).cuda()
inputs_3 = torch.randn((4, 3, 224, 224)).cuda()

with torch_tensorrt.runtime.enable_cudagraphs(opt) as cudagraphs_module:
    out_trt_2 = cudagraphs_module(inputs_2)
    out_trt_3 = cudagraphs_module(inputs_3)

具有包含圖形中斷的模組的 Cuda 圖形

當 CUDA 圖形應用於包含圖形中斷的 TensorRT 模型時，每個中斷都會引入額外的 overhead。這是因為圖形中斷阻止整個模型作為單個、連續的最佳化單元執行。因此，CUDA 圖形通常提供的一些效能優勢，例如減少核心啟動 overhead 並提高執行效率，可能會降低。將包裝的運行時模組與 CUDA 圖形一起使用，可讓您將操作序列封裝到可以有效執行的圖形中，即使存在圖形中斷也是如此。如果 TensorRT 模組有圖形中斷，CUDA Graph context manager 會返回一個 wrapped_module。這個模組捕獲整個執行圖，通過減少核心啟動 overhead 並提高效能，從而在後續的推論期間實現有效的重播。請注意，使用包裝器模組進行初始化涉及一個熱身階段，該模組將執行多次。此熱身可確保記憶體分配和初始化不會記錄在 CUDA 圖形中，這有助於維持一致的執行路徑並優化效能。

class SampleModel(torch.nn.Module):
    def forward(self, x):
        return torch.relu((x + 2) * 0.5)


model = SampleModel().eval().cuda()
input = torch.randn((1, 3, 224, 224)).to("cuda")

# The 'torch_executed_ops' compiler option is used in this example to intentionally introduce graph breaks within the module.
# Note: The Dynamo backend is required for the CUDA Graph context manager to handle modules in an Ahead-Of-Time (AOT) manner.
opt_with_graph_break = torch_tensorrt.compile(
    model,
    ir="dynamo",
    inputs=[input],
    min_block_size=1,
    pass_through_build_failures=True,
    torch_executed_ops={"torch.ops.aten.mul.Tensor"},
)

如果模組有圖形中斷，則整個子模組將被 cuda 圖形記錄和重播

with torch_tensorrt.runtime.enable_cudagraphs(
    opt_with_graph_break
) as cudagraphs_module:
    cudagraphs_module(input)