vllm - 💡(How to fix) Fix [Bug]: Inconsistent KV Cache reporting and system hang on long context requests (Gemma-4 26B AWQ Int4) [1 participants]

vllm2026-04-16 10:45:32

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.

GitHub issue URL

Helpful · Quick feedback

GitHub stats

vllm-project/vllm#40002•Fetched 2026-04-17 08:27:45

View on GitHub

Comments

Participants

Timeline

Reactions

Author

GitEventhandler

Participants

GitEventhandler

Timeline (top)

cross-referenced ×1labeled ×1

Root Cause

1. Inconsistent KV Cache Reporting

In vllm/v1/core/kv_cache_utils.py, there is a discrepancy between the reported GPU KV cache size and the Maximum concurrency calculation. (not sure whether the issue is caused by swa)

Fix Action

Fix / Workaround

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

Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Address sizes: 46 bits physical, 57 bits virtual Byte Order: Little Endian CPU(s): 48 On-line CPU(s) list: 0-47 Vendor ID: GenuineIntel Model name: Intel(R) Xeon(R) Silver 4310 CPU @ 2.10GHz CPU family: 6 Model: 106 Thread(s) per core: 2 Core(s) per socket: 12 Socket(s): 2 Stepping: 6 CPU(s) scaling MHz: 33% CPU max MHz: 3300.0000 CPU min MHz: 800.0000 BogoMIPS: 4200.00 Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 ssbd mba ibrs ibpb stibp ibrs_enhanced tpr_shadow flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb intel_pt avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local split_lock_detect wbnoinvd dtherm ida arat pln pts vnmi avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg tme avx512_vpopcntdq la57 rdpid fsrm md_clear pconfig flush_l1d arch_capabilities Virtualization: VT-x L1d cache: 1.1 MiB (24 instances) L1i cache: 768 KiB (24 instances) L2 cache: 30 MiB (24 instances) L3 cache: 36 MiB (2 instances) NUMA node(s): 2 NUMA node0 CPU(s): 0-11,24-35 NUMA node1 CPU(s): 12-23,36-47 Vulnerability Gather data sampling: Mitigation; Microcode Vulnerability Itlb multihit: Not affected Vulnerability L1tf: Not affected Vulnerability Mds: Not affected Vulnerability Meltdown: Not affected Vulnerability Mmio stale data: Mitigation; Clear CPU buffers; SMT vulnerable 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; usercopy/swapgs barriers and __user pointer sanitization Vulnerability Spectre v2: Mitigation; Enhanced / Automatic IBRS; IBPB conditional; RSB filling; PBRSB-eIBRS SW sequence; BHI SW loop, KVM SW loop Vulnerability Srbds: Not affected Vulnerability Tsx async abort: Not affected Vulnerability Vmscape: Not affected

Code Example

Collecting environment information...
==============================
        System Info
==============================
OS                           : Ubuntu 22.04.5 LTS (x86_64)
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              : 2.10.0+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-6.8.0-107-generic-x86_64-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 GeForce RTX 4060 Ti
GPU 1: NVIDIA GeForce RTX 4060 Ti

Nvidia driver version        : 595.45.04
cuDNN version                : Could not collect
HIP runtime version          : N/A
MIOpen runtime version       : N/A
Is XNNPACK available         : True

==============================
          CPU Info
==============================
Architecture:                         x86_64
CPU op-mode(s):                       32-bit, 64-bit
Address sizes:                        46 bits physical, 57 bits virtual
Byte Order:                           Little Endian
CPU(s):                               48
On-line CPU(s) list:                  0-47
Vendor ID:                            GenuineIntel
Model name:                           Intel(R) Xeon(R) Silver 4310 CPU @ 2.10GHz
CPU family:                           6
Model:                                106
Thread(s) per core:                   2
Core(s) per socket:                   12
Socket(s):                            2
Stepping:                             6
CPU(s) scaling MHz:                   33%
CPU max MHz:                          3300.0000
CPU min MHz:                          800.0000
BogoMIPS:                             4200.00
Flags:                                fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 ssbd mba ibrs ibpb stibp ibrs_enhanced tpr_shadow flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb intel_pt avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local split_lock_detect wbnoinvd dtherm ida arat pln pts vnmi avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg tme avx512_vpopcntdq la57 rdpid fsrm md_clear pconfig flush_l1d arch_capabilities
Virtualization:                       VT-x
L1d cache:                            1.1 MiB (24 instances)
L1i cache:                            768 KiB (24 instances)
L2 cache:                             30 MiB (24 instances)
L3 cache:                             36 MiB (2 instances)
NUMA node(s):                         2
NUMA node0 CPU(s):                    0-11,24-35
NUMA node1 CPU(s):                    12-23,36-47
Vulnerability Gather data sampling:   Mitigation; Microcode
Vulnerability Itlb multihit:          Not affected
Vulnerability L1tf:                   Not affected
Vulnerability Mds:                    Not affected
Vulnerability Meltdown:               Not affected
Vulnerability Mmio stale data:        Mitigation; Clear CPU buffers; SMT vulnerable
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; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:             Mitigation; Enhanced / Automatic IBRS; IBPB conditional; RSB filling; PBRSB-eIBRS SW sequence; BHI SW loop, KVM SW loop
Vulnerability Srbds:                  Not affected
Vulnerability Tsx async abort:        Not affected
Vulnerability Vmscape:                Not affected

==============================
Versions of relevant libraries
==============================
[pip3] flashinfer-python==0.6.7
[pip3] numpy==2.2.6
[pip3] nvidia-cublas-cu12==12.9.1.4
[pip3] nvidia-cuda-cupti-cu12==12.9.79
[pip3] nvidia-cuda-nvrtc-cu12==12.9.86
[pip3] nvidia-cuda-runtime-cu12==12.9.79
[pip3] nvidia-cudnn-cu12==9.10.2.21
[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==4.4.2
[pip3] nvidia-cutlass-dsl-libs-base==4.4.2
[pip3] nvidia-ml-py==13.595.45
[pip3] nvidia-nccl-cu12==2.27.5
[pip3] nvidia-nvjitlink-cu12==12.9.86
[pip3] nvidia-nvshmem-cu12==3.4.5
[pip3] nvidia-nvtx-cu12==12.9.79
[pip3] pyzmq==27.1.0
[pip3] torch==2.10.0+cu129
[pip3] torch_c_dlpack_ext==0.1.5
[pip3] torchaudio==2.10.0+cu129
[pip3] torchvision==0.25.0+cu129
[pip3] transformers==5.5.0
[pip3] triton==3.6.0
[conda] Could not collect

==============================
         vLLM Info
==============================
ROCM Version                 : Could not collect
vLLM Version                 : 0.18.2rc1.dev73+gdb7a17ecc (git sha: db7a17ecc)
vLLM Build Flags:
  CUDA Archs: 7.0 7.5 8.0 8.9 9.0 10.0 12.0; ROCm: Disabled; XPU: Disabled
GPU Topology:
        GPU0    GPU1    CPU Affinity    NUMA Affinity   GPU NUMA ID
GPU0     X      NODE    0-11,24-35      0               N/A
GPU1    NODE     X      0-11,24-35      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_VISIBLE_DEVICES=void
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
TORCH_CUDA_ARCH_LIST=7.0 7.5 8.0 8.9 9.0 10.0 12.0
NVIDIA_DRIVER_CAPABILITIES=compute,utility
VLLM_USAGE_SOURCE=production-docker-image
CUDA_VERSION=12.9.1
VLLM_ENABLE_CUDA_COMPATIBILITY=0
LD_LIBRARY_PATH=/usr/local/nvidia/lib64:/usr/local/cuda/lib64:/usr/local/cuda/lib64
NVIDIA_CTK_LIBCUDA_DIR=/usr/lib/x86_64-linux-gnu
PYTORCH_NVML_BASED_CUDA_CHECK=1
TORCHINDUCTOR_COMPILE_THREADS=1
TORCHINDUCTOR_CACHE_DIR=/tmp/torchinductor_root

---

vllm  | (EngineCore pid=171) INFO 04-15 09:11:57 [kv_cache_utils.py:1319] GPU KV cache size: 79,002 tokens
vllm  | (EngineCore pid=171) INFO 04-15 09:11:57 [kv_cache_utils.py:1324] Maximum concurrency for 262,144 tokens per request: 3.24x

---

vllm serve \
    --model ${PATH_TO_MODELS}/gemma-4-26B-A4B-it-AWQ-4bit \
    --max-model-len 262144 \
    --max-num-seqs 4 \
    --tensor-parallel-size 2 \
    --enable-expert-parallel \
    --tokenizer-mode auto \
    --gpu-memory-utilization 0.90 \
    --dtype auto \
    --runner generate \
    --kv-cache-dtype fp8 \
    --structured-outputs-config '{"backend":"xgrammar"}' \
    --enable-prefix-caching \
    --enable-chunked-prefill \
    --reasoning-parser gemma4 \
    --enable-auto-tool-choice \
    --tool-call-parser gemma4

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 (x86_64)
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              : 2.10.0+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-6.8.0-107-generic-x86_64-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 GeForce RTX 4060 Ti
GPU 1: NVIDIA GeForce RTX 4060 Ti

Nvidia driver version        : 595.45.04
cuDNN version                : Could not collect
HIP runtime version          : N/A
MIOpen runtime version       : N/A
Is XNNPACK available         : True

==============================
          CPU Info
==============================
Architecture:                         x86_64
CPU op-mode(s):                       32-bit, 64-bit
Address sizes:                        46 bits physical, 57 bits virtual
Byte Order:                           Little Endian
CPU(s):                               48
On-line CPU(s) list:                  0-47
Vendor ID:                            GenuineIntel
Model name:                           Intel(R) Xeon(R) Silver 4310 CPU @ 2.10GHz
CPU family:                           6
Model:                                106
Thread(s) per core:                   2
Core(s) per socket:                   12
Socket(s):                            2
Stepping:                             6
CPU(s) scaling MHz:                   33%
CPU max MHz:                          3300.0000
CPU min MHz:                          800.0000
BogoMIPS:                             4200.00
Flags:                                fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 ssbd mba ibrs ibpb stibp ibrs_enhanced tpr_shadow flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb intel_pt avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local split_lock_detect wbnoinvd dtherm ida arat pln pts vnmi avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg tme avx512_vpopcntdq la57 rdpid fsrm md_clear pconfig flush_l1d arch_capabilities
Virtualization:                       VT-x
L1d cache:                            1.1 MiB (24 instances)
L1i cache:                            768 KiB (24 instances)
L2 cache:                             30 MiB (24 instances)
L3 cache:                             36 MiB (2 instances)
NUMA node(s):                         2
NUMA node0 CPU(s):                    0-11,24-35
NUMA node1 CPU(s):                    12-23,36-47
Vulnerability Gather data sampling:   Mitigation; Microcode
Vulnerability Itlb multihit:          Not affected
Vulnerability L1tf:                   Not affected
Vulnerability Mds:                    Not affected
Vulnerability Meltdown:               Not affected
Vulnerability Mmio stale data:        Mitigation; Clear CPU buffers; SMT vulnerable
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; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:             Mitigation; Enhanced / Automatic IBRS; IBPB conditional; RSB filling; PBRSB-eIBRS SW sequence; BHI SW loop, KVM SW loop
Vulnerability Srbds:                  Not affected
Vulnerability Tsx async abort:        Not affected
Vulnerability Vmscape:                Not affected

==============================
Versions of relevant libraries
==============================
[pip3] flashinfer-python==0.6.7
[pip3] numpy==2.2.6
[pip3] nvidia-cublas-cu12==12.9.1.4
[pip3] nvidia-cuda-cupti-cu12==12.9.79
[pip3] nvidia-cuda-nvrtc-cu12==12.9.86
[pip3] nvidia-cuda-runtime-cu12==12.9.79
[pip3] nvidia-cudnn-cu12==9.10.2.21
[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==4.4.2
[pip3] nvidia-cutlass-dsl-libs-base==4.4.2
[pip3] nvidia-ml-py==13.595.45
[pip3] nvidia-nccl-cu12==2.27.5
[pip3] nvidia-nvjitlink-cu12==12.9.86
[pip3] nvidia-nvshmem-cu12==3.4.5
[pip3] nvidia-nvtx-cu12==12.9.79
[pip3] pyzmq==27.1.0
[pip3] torch==2.10.0+cu129
[pip3] torch_c_dlpack_ext==0.1.5
[pip3] torchaudio==2.10.0+cu129
[pip3] torchvision==0.25.0+cu129
[pip3] transformers==5.5.0
[pip3] triton==3.6.0
[conda] Could not collect

==============================
         vLLM Info
==============================
ROCM Version                 : Could not collect
vLLM Version                 : 0.18.2rc1.dev73+gdb7a17ecc (git sha: db7a17ecc)
vLLM Build Flags:
  CUDA Archs: 7.0 7.5 8.0 8.9 9.0 10.0 12.0; ROCm: Disabled; XPU: Disabled
GPU Topology:
        GPU0    GPU1    CPU Affinity    NUMA Affinity   GPU NUMA ID
GPU0     X      NODE    0-11,24-35      0               N/A
GPU1    NODE     X      0-11,24-35      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_VISIBLE_DEVICES=void
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
TORCH_CUDA_ARCH_LIST=7.0 7.5 8.0 8.9 9.0 10.0 12.0
NVIDIA_DRIVER_CAPABILITIES=compute,utility
VLLM_USAGE_SOURCE=production-docker-image
CUDA_VERSION=12.9.1
VLLM_ENABLE_CUDA_COMPATIBILITY=0
LD_LIBRARY_PATH=/usr/local/nvidia/lib64:/usr/local/cuda/lib64:/usr/local/cuda/lib64
NVIDIA_CTK_LIBCUDA_DIR=/usr/lib/x86_64-linux-gnu
PYTORCH_NVML_BASED_CUDA_CHECK=1
TORCHINDUCTOR_COMPILE_THREADS=1
TORCHINDUCTOR_CACHE_DIR=/tmp/torchinductor_root

</details>

🐛 Describe the bug

Hi vLLM team! I encountered two interrelated issues when using the gemma-4-26B-A4B-it-AWQ-4bit model with the official Docker image.

1. Inconsistent KV Cache Reporting

In vllm/v1/core/kv_cache_utils.py, there is a discrepancy between the reported GPU KV cache size and the Maximum concurrency calculation. (not sure whether the issue is caused by swa)

Logs:

vllm  | (EngineCore pid=171) INFO 04-15 09:11:57 [kv_cache_utils.py:1319] GPU KV cache size: 79,002 tokens
vllm  | (EngineCore pid=171) INFO 04-15 09:11:57 [kv_cache_utils.py:1324] Maximum concurrency for 262,144 tokens per request: 3.24x

2. System Hang on Long Context Requests

When submitting a request with a sequence length of approximately 170k tokens (which is within the max-model-len of 262,144 but exceeds the actual GPU KV cache size of 79,002), the engine hangs: CPU usage spikes to 100% (pinned to a single core) while GPU utilization remains at 0%. Increasing --renderer-num-workers 12 does not resolve the issue; the workload remains stuck on a single core.

Reproduction (with docker image `vllm/vllm-openai:gemma4`)

vllm serve \
    --model ${PATH_TO_MODELS}/gemma-4-26B-A4B-it-AWQ-4bit \
    --max-model-len 262144 \
    --max-num-seqs 4 \
    --tensor-parallel-size 2 \
    --enable-expert-parallel \
    --tokenizer-mode auto \
    --gpu-memory-utilization 0.90 \
    --dtype auto \
    --runner generate \
    --kv-cache-dtype fp8 \
    --structured-outputs-config '{"backend":"xgrammar"}' \
    --enable-prefix-caching \
    --enable-chunked-prefill \
    --reasoning-parser gemma4 \
    --enable-auto-tool-choice \
    --tool-call-parser gemma4

Environment Variables:
- OMP_NUM_THREADS=12
- VLLM_MEMORY_PROFILER_ESTIMATE_CUDAGRAPHS=1

Expected behavior

The tokenization process completes within a short period of time, with prefill performed by the GPU.

I think this BUG should be the same type as the one here (#39914 ). When the context approaches or exceeds the length indicated by [kv_cache_utils.py:1319] GPU KV cache size: $N$ tokens, the system hangs.

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.

extent analysis

TL;DR

The issue is likely caused by the discrepancy between the reported GPU KV cache size and the maximum concurrency calculation, leading to system hangs when submitting long context requests.

Guidance

Verify KV cache size calculation: Review the kv_cache_utils.py file to ensure the GPU KV cache size calculation is accurate and consistent with the maximum concurrency calculation.
Check system configuration: Confirm that the system configuration, including environment variables and Docker image settings, is consistent with the expected behavior.
Test with adjusted settings: Attempt to reproduce the issue with adjusted settings, such as reducing the --max-model-len or increasing the --gpu-memory-utilization threshold.
Monitor system resources: Use tools like top or nvidia-smi to monitor CPU and GPU utilization during the tokenization process to identify potential bottlenecks.

Example

No code snippet is provided as the issue is related to system configuration and environment variables.

Notes

The issue may be related to a known bug (#39914), and the provided reproduction steps and environment variables should be reviewed carefully to ensure consistency with the expected behavior.

Recommendation

Apply a workaround by adjusting the --max-model-len or --gpu-memory-utilization settings to avoid exceeding the GPU KV cache size, until a permanent fix is available.

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?

FAQ

Expected behavior

The tokenization process completes within a short period of time, with prefill performed by the GPU.

#api #model loading #dependency error #configuration error #environment variable

Still need to ship something?

×6

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

Data

Security

Network

Code

UI/UX

Text

System

Multimedia

Protocol

API

Engineering

vllm - 💡(How to fix) Fix [Bug]: Inconsistent KV Cache reporting and system hang on long context requests (Gemma-4 26B AWQ Int4) [1 participants]

Recommended Tools

GitHub issue graph ai analysis

Root Cause

1. Inconsistent KV Cache Reporting

Fix Action

Fix / Workaround

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

Code Example

Your current environment

🐛 Describe the bug

1. Inconsistent KV Cache Reporting

2. System Hang on Long Context Requests

Reproduction (with docker image `vllm/vllm-openai:gemma4`)

Expected behavior

Before submitting a new issue...

extent analysis

TL;DR

Guidance

Example

Notes

Recommendation

FAQ

Expected behavior

Still need to ship something?

TRENDING

vllm - 💡(How to fix) Fix [Bug]: Inconsistent KV Cache reporting and system hang on long context requests (Gemma-4 26B AWQ Int4) [1 participants]

Recommended Tools

GitHub issue graph ai analysis

Root Cause

1. Inconsistent KV Cache Reporting

Fix Action

Fix / Workaround

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

Code Example

Your current environment

🐛 Describe the bug

1. Inconsistent KV Cache Reporting

2. System Hang on Long Context Requests

Reproduction (with docker image vllm/vllm-openai:gemma4)

Expected behavior

Before submitting a new issue...

extent analysis

TL;DR

Guidance

Example

Notes

Recommendation

FAQ

Expected behavior

Still need to ship something?

RELATED_DISCOVERY

TRENDING

Reproduction (with docker image `vllm/vllm-openai:gemma4`)