Issue Station
Sign In
Tools hereToolbox

vllm - 💡(How to fix) Fix [Bug]: SM12.1 / GB10 still fails in CutlassFp8BlockScaledMMKernel after #41215

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…

Error Message

RuntimeError: cutlass_gemm_caller, /workspace/csrc/libtorch_stable/quantization/w8a8/cutlass/c3x/cutlass_gemm_caller.cuh:61, Error Internal

Root Cause

RuntimeError: Engine core initialization failed. See root cause above.

Fix Action

Fix / Workaround

============================== CPU Info

Architecture: aarch64 CPU op-mode(s): 64-bit Byte Order: Little Endian CPU(s): 20 On-line CPU(s) list: 0-19 Vendor ID: ARM Model: 1 Thread(s) per core: 1 Core(s) per socket: 5 Socket(s): 1 Stepping: r0p1 Frequency boost: disabled CPU max MHz: 2808.0000 CPU min MHz: 338.0000 BogoMIPS: 2000.00 Flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh bti ecv afp wfxt Model: 1 Thread(s) per core: 1 Core(s) per socket: 5 Socket(s): 1 Stepping: r0p1 CPU max MHz: 3900.0000 CPU min MHz: 1378.0000 BogoMIPS: 2000.00 Flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh bti ecv afp wfxt Model: 1 Thread(s) per core: 1 Core(s) per socket: 5 Socket(s): 1 Stepping: r0p1 CPU max MHz: 2808.0000 CPU min MHz: 338.0000 BogoMIPS: 2000.00 Flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh bti ecv afp wfxt Model: 1 Thread(s) per core: 1 Core(s) per socket: 5 Socket(s): 1 Stepping: r0p1 CPU max MHz: 3900.0000 CPU min MHz: 1378.0000 BogoMIPS: 2000.00 Flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh bti ecv afp wfxt L1d cache: 1.3 MiB (20 instances) L1i cache: 1.3 MiB (20 instances) L2 cache: 25 MiB (20 instances) L3 cache: 24 MiB (2 instances) NUMA node(s): 1 NUMA node0 CPU(s): 0-19 Vulnerability Gather data sampling: Not affected Vulnerability Ghostwrite: Not affected Vulnerability Indirect target selection: Not affected Vulnerability Itlb multihit: Not affected Vulnerability L1tf: Not affected Vulnerability Mds: Not affected Vulnerability Meltdown: Not affected Vulnerability Mmio stale data: Not affected Vulnerability Old microcode: Not affected Vulnerability Reg file data sampling: Not affected Vulnerability Retbleed: Not affected Vulnerability Spec rstack overflow: Not affected Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl Vulnerability Spectre v1: Mitigation; __user pointer sanitization Vulnerability Spectre v2: Mitigation; CSV2, BHB Vulnerability Srbds: Not affected Vulnerability Tsa: Not affected Vulnerability Tsx async abort: Not affected Vulnerability Vmscape: Not affected

With that workaround, vLLM selects: Selected TritonFp8BlockScaledMMKernel for Fp8LinearMethod

The workaround is diagnostic only. We do not want GB10 / SM12.1 permanently routed away from the optimized CUTLASS FP8 block-scaled path.

Code Example

Collecting environment information...
==============================
        System Info
==============================
OS                           : Ubuntu 22.04.5 LTS (aarch64)
GCC version                  : (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0
Clang version                : Could not collect
CMake version                : Could not collect
Libc version                 : glibc-2.35

==============================
       PyTorch Info
==============================
PyTorch version              : 2.11.0+cu129
Is debug build               : False
CUDA used to build PyTorch   : 12.9
ROCM used to build PyTorch   : N/A
XPU used to build PyTorch    : N/A

==============================
      Python Environment
==============================
Python version               : 3.12.13 (main, Mar  4 2026, 09:23:07) [GCC 11.4.0] (64-bit runtime)
Python platform              : Linux-6.17.0-1018-nvidia-aarch64-with-glibc2.35
    
==============================
       CUDA / GPU Info
==============================
Is CUDA available            : True
CUDA runtime version         : 12.9.86
CUDA_MODULE_LOADING set to   : 
GPU models and configuration : GPU 0: NVIDIA GB10
Nvidia driver version        : 580.159.03
cuDNN version                : Could not collect
HIP runtime version          : N/A
MIOpen runtime version       : N/A
Is XNNPACK available         : True

==============================
          CPU Info
==============================
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  20
On-line CPU(s) list:                     0-19
Vendor ID:                               ARM
Model:                                   1
Thread(s) per core:                      1
Core(s) per socket:                      5
Socket(s):                               1
Stepping:                                r0p1
Frequency boost:                         disabled
CPU max MHz:                             2808.0000
CPU min MHz:                             338.0000
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh bti ecv afp wfxt
Model:                                   1
Thread(s) per core:                      1
Core(s) per socket:                      5
Socket(s):                               1
Stepping:                                r0p1
CPU max MHz:                             3900.0000
CPU min MHz:                             1378.0000
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh bti ecv afp wfxt
Model:                                   1
Thread(s) per core:                      1
Core(s) per socket:                      5
Socket(s):                               1
Stepping:                                r0p1
CPU max MHz:                             2808.0000
CPU min MHz:                             338.0000
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh bti ecv afp wfxt
Model:                                   1
Thread(s) per core:                      1
Core(s) per socket:                      5
Socket(s):                               1
Stepping:                                r0p1
CPU max MHz:                             3900.0000
CPU min MHz:                             1378.0000
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh bti ecv afp wfxt
L1d cache:                               1.3 MiB (20 instances)
L1i cache:                               1.3 MiB (20 instances)
L2 cache:                                25 MiB (20 instances)
L3 cache:                                24 MiB (2 instances)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-19
Vulnerability Gather data sampling:      Not affected
Vulnerability Ghostwrite:                Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Old microcode:             Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

==============================
Versions of relevant libraries
==============================
[pip3] flashinfer-python==0.6.11.post2
[pip3] numpy==2.2.6
[pip3] nvidia-cublas-cu12==12.9.1.4
[pip3] nvidia-cuda-cccl==13.2.75
[pip3] nvidia-cuda-cccl-cu12==12.9.27
[pip3] nvidia-cuda-crt==13.2.78
[pip3] nvidia-cuda-cupti-cu12==12.9.79
[pip3] nvidia-cuda-nvcc==13.2.78
[pip3] nvidia-cuda-nvcc-cu12==12.9.86
[pip3] nvidia-cuda-nvrtc==13.2.78
[pip3] nvidia-cuda-nvrtc-cu12==12.9.86
[pip3] nvidia-cuda-runtime==13.2.75
[pip3] nvidia-cuda-runtime-cu12==12.9.79
[pip3] nvidia-cudnn-cu12==9.17.1.4
[pip3] nvidia-cudnn-frontend==1.18.0
[pip3] nvidia-cufft-cu12==11.4.1.4
[pip3] nvidia-cufile-cu12==1.14.1.1
[pip3] nvidia-curand-cu12==10.3.10.19
[pip3] nvidia-cusolver-cu12==11.7.5.82
[pip3] nvidia-cusparse-cu12==12.5.10.65
[pip3] nvidia-cusparselt-cu12==0.7.1
[pip3] nvidia-cutlass-dsl==4.5.0
[pip3] nvidia-cutlass-dsl-libs-base==4.5.0
[pip3] nvidia-cutlass-dsl-libs-cu13==4.5.0
[pip3] nvidia-ml-py==13.595.45
[pip3] nvidia-nccl-cu12==2.28.9
[pip3] nvidia-nvjitlink-cu12==12.9.86
[pip3] nvidia-nvshmem-cu12==3.4.5
[pip3] nvidia-nvtx-cu12==12.9.79
[pip3] nvidia-nvvm==13.2.78
[pip3] pyzmq==27.1.0
[pip3] tokenspeed-triton==3.7.10.post20260505
[pip3] torch==2.11.0+cu129
[pip3] torch_c_dlpack_ext==0.1.5
[pip3] torchaudio==2.11.0+cu129
[pip3] torchvision==0.26.0+cu129
[pip3] transformers==5.9.0
[pip3] triton==3.6.0
[conda] Could not collect

==============================
         vLLM Info
==============================
ROCM Version                 : Could not collect
vLLM Version                 : 0.21.1rc1.dev169+ga6682d1d2 (git sha: a6682d1d2)
vLLM Build Flags:
  CUDA Archs: 8.0 8.7 8.9 9.0 10.0 12.0; ROCm: Disabled; XPU: Disabled
GPU Topology:
  	 [4mGPU0	CPU Affinity	NUMA Affinity	GPU NUMA ID [0m
GPU0	 X 	0-19	0		N/A

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

==============================
     Environment Variables
==============================
NVIDIA_REQUIRE_CUDA=cuda>=12.9 brand=unknown,driver>=535,driver<536 brand=grid,driver>=535,driver<536 brand=tesla,driver>=535,driver<536 brand=nvidia,driver>=535,driver<536 brand=quadro,driver>=535,driver<536 brand=quadrortx,driver>=535,driver<536 brand=nvidiartx,driver>=535,driver<536 brand=vapps,driver>=535,driver<536 brand=vpc,driver>=535,driver<536 brand=vcs,driver>=535,driver<536 brand=vws,driver>=535,driver<536 brand=cloudgaming,driver>=535,driver<536 brand=unknown,driver>=550,driver<551 brand=grid,driver>=550,driver<551 brand=tesla,driver>=550,driver<551 brand=nvidia,driver>=550,driver<551 brand=quadro,driver>=550,driver<551 brand=quadrortx,driver>=550,driver<551 brand=nvidiartx,driver>=550,driver<551 brand=vapps,driver>=550,driver<551 brand=vpc,driver>=550,driver<551 brand=vcs,driver>=550,driver<551 brand=vws,driver>=550,driver<551 brand=cloudgaming,driver>=550,driver<551 brand=unknown,driver>=560,driver<561 brand=grid,driver>=560,driver<561 brand=tesla,driver>=560,driver<561 brand=nvidia,driver>=560,driver<561 brand=quadro,driver>=560,driver<561 brand=quadrortx,driver>=560,driver<561 brand=nvidiartx,driver>=560,driver<561 brand=vapps,driver>=560,driver<561 brand=vpc,driver>=560,driver<561 brand=vcs,driver>=560,driver<561 brand=vws,driver>=560,driver<561 brand=cloudgaming,driver>=560,driver<561 brand=unknown,driver>=565,driver<566 brand=grid,driver>=565,driver<566 brand=tesla,driver>=565,driver<566 brand=nvidia,driver>=565,driver<566 brand=quadro,driver>=565,driver<566 brand=quadrortx,driver>=565,driver<566 brand=nvidiartx,driver>=565,driver<566 brand=vapps,driver>=565,driver<566 brand=vpc,driver>=565,driver<566 brand=vcs,driver>=565,driver<566 brand=vws,driver>=565,driver<566 brand=cloudgaming,driver>=565,driver<566 brand=unknown,driver>=570,driver<571 brand=grid,driver>=570,driver<571 brand=tesla,driver>=570,driver<571 brand=nvidia,driver>=570,driver<571 brand=quadro,driver>=570,driver<571 brand=quadrortx,driver>=570,driver<571 brand=nvidiartx,driver>=570,driver<571 brand=vapps,driver>=570,driver<571 brand=vpc,driver>=570,driver<571 brand=vcs,driver>=570,driver<571 brand=vws,driver>=570,driver<571 brand=cloudgaming,driver>=570,driver<571
CUDA_VERSION=12.9.1
LD_LIBRARY_PATH=/usr/local/nvidia/lib64:/usr/local/cuda/lib64:/usr/local/cuda/lib64
NVIDIA_DRIVER_CAPABILITIES=compute,utility
VLLM_ENABLE_CUDA_COMPATIBILITY=0
TORCH_CUDA_ARCH_LIST=8.0 8.7 8.9 9.0 10.0 12.0
VLLM_USAGE_SOURCE=production-docker-image
NVIDIA_VISIBLE_DEVICES=all
PYTORCH_NVML_BASED_CUDA_CHECK=1
TORCHINDUCTOR_COMPILE_THREADS=1
TORCHINDUCTOR_CACHE_DIR=/tmp/torchinductor_root

---

import torch
from vllm import _custom_ops as ops

m, n, k = 1, 16384, 5120

a = torch.randn((m, k), device="cuda").to(torch.float8_e4m3fn)

# Match vLLM's B.T path from the failing Qwen3.6 compiled graph:
# logical B shape is (k, n) with transpose-like stride.
b = torch.randn((n, k), device="cuda").to(torch.float8_e4m3fn).t()

scale_a = torch.ones((m, k // 128), device="cuda", dtype=torch.float32)
scale_b = torch.ones((k // 128, n // 128), device="cuda", dtype=torch.float32)

out = ops.cutlass_scaled_mm(a, b, scale_a, scale_b, torch.bfloat16, None)
torch.cuda.synchronize()
print(out.shape)

---

RuntimeError: cutlass_gemm_caller,
/workspace/csrc/libtorch_stable/quantization/w8a8/cutlass/c3x/cutlass_gemm_caller.cuh:61,
Error Internal

---

File "/root/.cache/vllm/torch_compile_cache/.../inductor_cache/...py", line 348, in call
  torch.ops._C.cutlass_scaled_mm.default(
      buf7,
      buf2,
      reinterpret_tensor(arg4_1, (5120, 16384), (1, 5120), 0),
      reinterpret_tensor(buf3, (s18, 40), (1, s18), 0),
      reinterpret_tensor(arg5_1, (40, 128), (1, 40), 0),
      None)

RuntimeError: cutlass_gemm_caller,
/workspace/csrc/libtorch_stable/quantization/w8a8/cutlass/c3x/cutlass_gemm_caller.cuh:61,
Error Internal

---

vllm/v1/worker/gpu_worker.py:392 determine_available_memory
vllm/v1/worker/gpu_model_runner.py:6133 profile_run
vllm/v1/worker/gpu_model_runner.py:5793 _dummy_run
vllm/model_executor/models/qwen3_5.py:658 forward
torch.ops._C.cutlass_scaled_mm.default(...)
RuntimeError: cutlass_gemm_caller ... Error Internal
RAW_BUFFERClick to expand / collapse

Your current environment

<details> <summary>The output of <code>python collect_env.py</code></summary> 
Collecting environment information...
==============================
        System Info
==============================
OS                           : Ubuntu 22.04.5 LTS (aarch64)
GCC version                  : (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0
Clang version                : Could not collect
CMake version                : Could not collect
Libc version                 : glibc-2.35

==============================
       PyTorch Info
==============================
PyTorch version              : 2.11.0+cu129
Is debug build               : False
CUDA used to build PyTorch   : 12.9
ROCM used to build PyTorch   : N/A
XPU used to build PyTorch    : N/A

==============================
      Python Environment
==============================
Python version               : 3.12.13 (main, Mar  4 2026, 09:23:07) [GCC 11.4.0] (64-bit runtime)
Python platform              : Linux-6.17.0-1018-nvidia-aarch64-with-glibc2.35
    
==============================
       CUDA / GPU Info
==============================
Is CUDA available            : True
CUDA runtime version         : 12.9.86
CUDA_MODULE_LOADING set to   : 
GPU models and configuration : GPU 0: NVIDIA GB10
Nvidia driver version        : 580.159.03
cuDNN version                : Could not collect
HIP runtime version          : N/A
MIOpen runtime version       : N/A
Is XNNPACK available         : True

==============================
          CPU Info
==============================
Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  20
On-line CPU(s) list:                     0-19
Vendor ID:                               ARM
Model:                                   1
Thread(s) per core:                      1
Core(s) per socket:                      5
Socket(s):                               1
Stepping:                                r0p1
Frequency boost:                         disabled
CPU max MHz:                             2808.0000
CPU min MHz:                             338.0000
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh bti ecv afp wfxt
Model:                                   1
Thread(s) per core:                      1
Core(s) per socket:                      5
Socket(s):                               1
Stepping:                                r0p1
CPU max MHz:                             3900.0000
CPU min MHz:                             1378.0000
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh bti ecv afp wfxt
Model:                                   1
Thread(s) per core:                      1
Core(s) per socket:                      5
Socket(s):                               1
Stepping:                                r0p1
CPU max MHz:                             2808.0000
CPU min MHz:                             338.0000
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh bti ecv afp wfxt
Model:                                   1
Thread(s) per core:                      1
Core(s) per socket:                      5
Socket(s):                               1
Stepping:                                r0p1
CPU max MHz:                             3900.0000
CPU min MHz:                             1378.0000
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 sm3 sm4 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 svesm4 flagm2 frint svei8mm svebf16 i8mm bf16 dgh bti ecv afp wfxt
L1d cache:                               1.3 MiB (20 instances)
L1i cache:                               1.3 MiB (20 instances)
L2 cache:                                25 MiB (20 instances)
L3 cache:                                24 MiB (2 instances)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0-19
Vulnerability Gather data sampling:      Not affected
Vulnerability Ghostwrite:                Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Old microcode:             Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Not affected
Vulnerability Spec rstack overflow:      Not affected
Vulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:                Mitigation; __user pointer sanitization
Vulnerability Spectre v2:                Mitigation; CSV2, BHB
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Not affected

==============================
Versions of relevant libraries
==============================
[pip3] flashinfer-python==0.6.11.post2
[pip3] numpy==2.2.6
[pip3] nvidia-cublas-cu12==12.9.1.4
[pip3] nvidia-cuda-cccl==13.2.75
[pip3] nvidia-cuda-cccl-cu12==12.9.27
[pip3] nvidia-cuda-crt==13.2.78
[pip3] nvidia-cuda-cupti-cu12==12.9.79
[pip3] nvidia-cuda-nvcc==13.2.78
[pip3] nvidia-cuda-nvcc-cu12==12.9.86
[pip3] nvidia-cuda-nvrtc==13.2.78
[pip3] nvidia-cuda-nvrtc-cu12==12.9.86
[pip3] nvidia-cuda-runtime==13.2.75
[pip3] nvidia-cuda-runtime-cu12==12.9.79
[pip3] nvidia-cudnn-cu12==9.17.1.4
[pip3] nvidia-cudnn-frontend==1.18.0
[pip3] nvidia-cufft-cu12==11.4.1.4
[pip3] nvidia-cufile-cu12==1.14.1.1
[pip3] nvidia-curand-cu12==10.3.10.19
[pip3] nvidia-cusolver-cu12==11.7.5.82
[pip3] nvidia-cusparse-cu12==12.5.10.65
[pip3] nvidia-cusparselt-cu12==0.7.1
[pip3] nvidia-cutlass-dsl==4.5.0
[pip3] nvidia-cutlass-dsl-libs-base==4.5.0
[pip3] nvidia-cutlass-dsl-libs-cu13==4.5.0
[pip3] nvidia-ml-py==13.595.45
[pip3] nvidia-nccl-cu12==2.28.9
[pip3] nvidia-nvjitlink-cu12==12.9.86
[pip3] nvidia-nvshmem-cu12==3.4.5
[pip3] nvidia-nvtx-cu12==12.9.79
[pip3] nvidia-nvvm==13.2.78
[pip3] pyzmq==27.1.0
[pip3] tokenspeed-triton==3.7.10.post20260505
[pip3] torch==2.11.0+cu129
[pip3] torch_c_dlpack_ext==0.1.5
[pip3] torchaudio==2.11.0+cu129
[pip3] torchvision==0.26.0+cu129
[pip3] transformers==5.9.0
[pip3] triton==3.6.0
[conda] Could not collect

==============================
         vLLM Info
==============================
ROCM Version                 : Could not collect
vLLM Version                 : 0.21.1rc1.dev169+ga6682d1d2 (git sha: a6682d1d2)
vLLM Build Flags:
  CUDA Archs: 8.0 8.7 8.9 9.0 10.0 12.0; ROCm: Disabled; XPU: Disabled
GPU Topology:
  	 [4mGPU0	CPU Affinity	NUMA Affinity	GPU NUMA ID [0m
GPU0	 X 	0-19	0		N/A

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

==============================
     Environment Variables
==============================
NVIDIA_REQUIRE_CUDA=cuda>=12.9 brand=unknown,driver>=535,driver<536 brand=grid,driver>=535,driver<536 brand=tesla,driver>=535,driver<536 brand=nvidia,driver>=535,driver<536 brand=quadro,driver>=535,driver<536 brand=quadrortx,driver>=535,driver<536 brand=nvidiartx,driver>=535,driver<536 brand=vapps,driver>=535,driver<536 brand=vpc,driver>=535,driver<536 brand=vcs,driver>=535,driver<536 brand=vws,driver>=535,driver<536 brand=cloudgaming,driver>=535,driver<536 brand=unknown,driver>=550,driver<551 brand=grid,driver>=550,driver<551 brand=tesla,driver>=550,driver<551 brand=nvidia,driver>=550,driver<551 brand=quadro,driver>=550,driver<551 brand=quadrortx,driver>=550,driver<551 brand=nvidiartx,driver>=550,driver<551 brand=vapps,driver>=550,driver<551 brand=vpc,driver>=550,driver<551 brand=vcs,driver>=550,driver<551 brand=vws,driver>=550,driver<551 brand=cloudgaming,driver>=550,driver<551 brand=unknown,driver>=560,driver<561 brand=grid,driver>=560,driver<561 brand=tesla,driver>=560,driver<561 brand=nvidia,driver>=560,driver<561 brand=quadro,driver>=560,driver<561 brand=quadrortx,driver>=560,driver<561 brand=nvidiartx,driver>=560,driver<561 brand=vapps,driver>=560,driver<561 brand=vpc,driver>=560,driver<561 brand=vcs,driver>=560,driver<561 brand=vws,driver>=560,driver<561 brand=cloudgaming,driver>=560,driver<561 brand=unknown,driver>=565,driver<566 brand=grid,driver>=565,driver<566 brand=tesla,driver>=565,driver<566 brand=nvidia,driver>=565,driver<566 brand=quadro,driver>=565,driver<566 brand=quadrortx,driver>=565,driver<566 brand=nvidiartx,driver>=565,driver<566 brand=vapps,driver>=565,driver<566 brand=vpc,driver>=565,driver<566 brand=vcs,driver>=565,driver<566 brand=vws,driver>=565,driver<566 brand=cloudgaming,driver>=565,driver<566 brand=unknown,driver>=570,driver<571 brand=grid,driver>=570,driver<571 brand=tesla,driver>=570,driver<571 brand=nvidia,driver>=570,driver<571 brand=quadro,driver>=570,driver<571 brand=quadrortx,driver>=570,driver<571 brand=nvidiartx,driver>=570,driver<571 brand=vapps,driver>=570,driver<571 brand=vpc,driver>=570,driver<571 brand=vcs,driver>=570,driver<571 brand=vws,driver>=570,driver<571 brand=cloudgaming,driver>=570,driver<571
CUDA_VERSION=12.9.1
LD_LIBRARY_PATH=/usr/local/nvidia/lib64:/usr/local/cuda/lib64:/usr/local/cuda/lib64
NVIDIA_DRIVER_CAPABILITIES=compute,utility
VLLM_ENABLE_CUDA_COMPATIBILITY=0
TORCH_CUDA_ARCH_LIST=8.0 8.7 8.9 9.0 10.0 12.0
VLLM_USAGE_SOURCE=production-docker-image
NVIDIA_VISIBLE_DEVICES=all
PYTORCH_NVML_BASED_CUDA_CHECK=1
TORCHINDUCTOR_COMPILE_THREADS=1
TORCHINDUCTOR_CACHE_DIR=/tmp/torchinductor_root
</details>

🐛 Describe the bug

CutlassFp8BlockScaledMMKernel still fails on GB10 / SM12.1 in the refreshed cu129-nightly-aarch64 image that includes #41215. The server fails during startup/profile when vLLM selects the CUTLASS FP8 block-scaled kernel: Selected CutlassFp8BlockScaledMMKernel for Fp8LinearMethod

Disabling only that kernel makes the same model/config start successfully through TritonFp8BlockScaledMMKernel, so the failure appears isolated to the SM120 CUTLASS blockwise FP8 path rather than the model, CUDA visibility, long context, or server configuration.

export VLLM_DISABLED_KERNELS=CutlassFp8BlockScaledMMKernel

With that workaround, vLLM selects: Selected TritonFp8BlockScaledMMKernel for Fp8LinearMethod

The workaround is diagnostic only. We do not want GB10 / SM12.1 permanently routed away from the optimized CUTLASS FP8 block-scaled path.

Minimal reproducer

This reproduces the same cutlass_gemm_caller.cuh:61 Error Internal outside the OpenAI server path, using only the vLLM custom op and synthetic tensors. It does not require downloading the full model. Run inside vllm/vllm-openai:cu129-nightly-aarch64 on GB10 / SM12.1:

import torch
from vllm import _custom_ops as ops

m, n, k = 1, 16384, 5120

a = torch.randn((m, k), device="cuda").to(torch.float8_e4m3fn)

# Match vLLM's B.T path from the failing Qwen3.6 compiled graph:
# logical B shape is (k, n) with transpose-like stride.
b = torch.randn((n, k), device="cuda").to(torch.float8_e4m3fn).t()

scale_a = torch.ones((m, k // 128), device="cuda", dtype=torch.float32)
scale_b = torch.ones((k // 128, n // 128), device="cuda", dtype=torch.float32)

out = ops.cutlass_scaled_mm(a, b, scale_a, scale_b, torch.bfloat16, None)
torch.cuda.synchronize()
print(out.shape)

Expected result: the CUTLASS scaled matmul completes and prints:

torch.Size([1, 16384])

Observed result:

RuntimeError: cutlass_gemm_caller,
/workspace/csrc/libtorch_stable/quantization/w8a8/cutlass/c3x/cutlass_gemm_caller.cuh:61,
Error Internal

I also tested the same (N=16384, K=5120) shape with these M values: M = 1, 2, 4, 8, 15, 16, 17, 32, 33, 64, 65, 100, 128, 256, 257, 512 All failed with the same cutlass_gemm_caller.cuh:61 Error Internal, so this does not appear limited to only the small-M swap path. Full server failure

The full server path fails during determine_available_memory() / profile_run() / _dummy_run(). The failing call in the compiled graph is:

File "/root/.cache/vllm/torch_compile_cache/.../inductor_cache/...py", line 348, in call
  torch.ops._C.cutlass_scaled_mm.default(
      buf7,
      buf2,
      reinterpret_tensor(arg4_1, (5120, 16384), (1, 5120), 0),
      reinterpret_tensor(buf3, (s18, 40), (1, s18), 0),
      reinterpret_tensor(arg5_1, (40, 128), (1, 40), 0),
      None)

RuntimeError: cutlass_gemm_caller,
/workspace/csrc/libtorch_stable/quantization/w8a8/cutlass/c3x/cutlass_gemm_caller.cuh:61,
Error Internal

Relevant stack location:

vllm/v1/worker/gpu_worker.py:392 determine_available_memory
vllm/v1/worker/gpu_model_runner.py:6133 profile_run
vllm/v1/worker/gpu_model_runner.py:5793 _dummy_run
vllm/model_executor/models/qwen3_5.py:658 forward
torch.ops._C.cutlass_scaled_mm.default(...)
RuntimeError: cutlass_gemm_caller ... Error Internal

Then the API server exits with:

RuntimeError: Engine core initialization failed. See root cause above.

The minimal direct ops.cutlass_scaled_mm reproducer above isolates the failure to the CUTLASS custom op without the full server stack.

Before submitting a new issue...

  • Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING