trtllm-bench throughput, Error: No such option: --ep version 0.18.1

System Info

v0.18.1 docker image build on GB200

Who can help?

No response

Information

• The official example scripts
• My own modified scripts

Tasks

• An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
• My own task or dataset (give details below)

Reproduction

Hello, I am following this guide

https://github.com/NVIDIA/TensorRT-LLM/blob/main/docs/source/blogs/Best_perf_practice_on_DeepSeek-R1_in_TensorRT-LLM.md

run this following cmd with v0.18.1, got the following error. please advise, which version has the --ep option.

2025-04-15 16:03:38,167 - INFO - flashinfer.jit: Prebuilt kernels not found, using JIT backend
[TensorRT-LLM] TensorRT-LLM version: 0.18.1
Usage: trtllm-bench throughput [OPTIONS]
Try 'trtllm-bench throughput --help' for help.

Error: No such option: --ep (Possible options: --help, --pp, --tp)

trtllm-bench -m nvidia/DeepSeek-R1-FP4
throughput
--tp 8
--ep 8
--warmup 0
--dataset ./dataset.txt
--backend pytorch
--max_batch_size 384
--max_num_tokens 1536
--num_requests 49152
--concurrency 3072
--kv_cache_free_gpu_mem_fraction 0.85
--extra_llm_api_options ./extra-llm-api-config.yml

Expected behavior

benchmark results

actual behavior

error, no --ep option

additional notes

for future readme, suggested to add working trtllm version

Start with the linked Best_perf_practice_on_DeepSeek-R1_in_TensorRT-LLM.md guide and run trtllm-bench throughput --help against version 0.18.1. Compare the documented command with the available options, determine the supported version or required documentation correction, and verify that the benchmark command reaches its expected results.