[Bug] bad performance on model Molmo-7B-D-0924

[zhulinJulia24]

Checklist

• 1. I have searched related issues but cannot get the expected help.
• 2. The bug has not been fixed in the latest version.
• 3. Please note that if the bug-related issue you submitted lacks corresponding environment info and a minimal reproducible demo, it will be challenging for us to reproduce and resolve the issue, reducing the likelihood of receiving feedback.

Describe the bug

bad performance on model Molmo-7B-D-0924

Reproduction

1. python3 benchmark/profile_throughput.py /nvme/qa_test_models/datasets/ShareGPT_V3_unfiltered_cleaned_split.json /nvme/qa_test_models/allenai/Molmo-7B-D-0924 --concurrency 256 --num-prompts 5000 --tp 1
   the result is
   =================================== Profile Throughput ===================================
   Benchmark duration 418.677
   Total requests 3000
   Successful requests 3000
   Concurrency 256
   Cancel rate 0
   Stream output true
   Skip tokenize false
   Skip detokenize false
   Total input tokens 680073
   Total generated tokens 623970
   Input throughput (tok/s) 1624.338
   Output throughput (tok/s) 1490.337
   Request throughput (req/s) 7.165

---

                                          mean       P50       P75       P95       P99

End-to-end Latency 34.174 35.039 36.022 38.719 40.092
Time to First Token (TTFT) 33.515 35.018 35.966 38.695 40.056
Time per Output Token (TPOT) 0.949 0.245 1.262 4.025 5.979
Inter-token Latency (ITL) 17.986 18.791 29.941 30.932 30.997
Tokens per Tick 200.633 129 313 642 798

Environment

sys.platform: linux
Python: 3.10.12 (main, Nov  6 2024, 20:22:13) [GCC 11.4.0]
CUDA available: True
MUSA available: False
numpy_random_seed: 2147483648
GPU 0,1,2,3,4,5,6,7: NVIDIA A100-SXM4-80GB
CUDA_HOME: /usr/local/cuda
NVCC: Cuda compilation tools, release 11.8, V11.8.89
GCC: x86_64-linux-gnu-gcc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0
PyTorch: 2.5.1+cu118
PyTorch compiling details: PyTorch built with:
  - GCC 9.3
  - C++ Version: 201703
  - Intel(R) oneAPI Math Kernel Library Version 2024.2-Product Build 20240605 for Intel(R) 64 architecture applications
  - Intel(R) MKL-DNN v3.5.3 (Git Hash 66f0cb9eb66affd2da3bf5f8d897376f04aae6af)
  - OpenMP 201511 (a.k.a. OpenMP 4.5)
  - LAPACK is enabled (usually provided by MKL)
  - NNPACK is enabled
  - CPU capability usage: AVX512
  - CUDA Runtime 11.8
  - NVCC architecture flags: -gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_80,code=sm_80;-gencode;arch=compute_86,code=sm_86;-gencode;arch=compute_37,code=sm_37;-gencode;arch=compute_90,code=sm_90
  - CuDNN 90.1
  - Magma 2.6.1
  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=11.8, CUDNN_VERSION=9.1.0, CXX_COMPILER=/opt/rh/devtoolset-9/root/usr/bin/c++, CXX_FLAGS= -D_GLIBCXX_USE_CXX11_ABI=0 -fabi-version=11 -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -DNDEBUG -DUSE_KINETO -DLIBKINETO_NOROCTRACER -DLIBKINETO_NOXPUPTI=ON -DUSE_FBGEMM -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wall -Wextra -Werror=return-type -Werror=non-virtual-dtor -Werror=bool-operation -Wnarrowing -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-unused-parameter -Wno-strict-overflow -Wno-strict-aliasing -Wno-stringop-overflow -Wsuggest-override -Wno-psabi -Wno-error=old-style-cast -Wno-missing-braces -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, TORCH_VERSION=2.5.1, USE_CUDA=ON, USE_CUDNN=ON, USE_CUSPARSELT=1, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_GLOO=ON, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=1, USE_NNPACK=ON, USE_OPENMP=ON, USE_ROCM=OFF, USE_ROCM_KERNEL_ASSERT=OFF, 

TorchVision: 0.20.1+cu118
LMDeploy: 0.7.0+df06ae3
transformers: 4.48.0
gradio: 5.12.0
fastapi: 0.115.6
pydantic: 2.10.5
triton: 3.1.0
NVIDIA Topology: 
        GPU0    GPU1    GPU2    GPU3    GPU4    GPU5    GPU6    GPU7    CPU Affinity    NUMA Affinity
GPU0     X      NV12    NV12    NV12    NV12    NV12    NV12    NV12    0-27,56-83      0
GPU1    NV12     X      NV12    NV12    NV12    NV12    NV12    NV12    0-27,56-83      0
GPU2    NV12    NV12     X      NV12    NV12    NV12    NV12    NV12    0-27,56-83      0
GPU3    NV12    NV12    NV12     X      NV12    NV12    NV12    NV12    0-27,56-83      0
GPU4    NV12    NV12    NV12    NV12     X      NV12    NV12    NV12    28-55,84-111    1
GPU5    NV12    NV12    NV12    NV12    NV12     X      NV12    NV12    28-55,84-111    1
GPU6    NV12    NV12    NV12    NV12    NV12    NV12     X      NV12    28-55,84-111    1
GPU7    NV12    NV12    NV12    NV12    NV12    NV12    NV12     X      28-55,84-111    1

Legend:

  X    = Self
  SYS  = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)
  NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node
  PHB  = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)
  PXB  = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)
  PIX  = Connection traversing at most a single PCIe bridge
  NV#  = Connection traversing a bonded set of # NVLinks

Error traceback