Issue Station
Sign In
Tools hereToolbox

vllm - 💡(How to fix) Fix [Bug]: OOM during FlashInfer JIT compile of gdn_prefill_sm90 on H100 due to many concurrent cicc processes [2 comments, 1 participants]

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…
GitHub stats
vllm-project/vllm#37279•Fetched 2026-04-08 00:48:19
View on GitHub
Comments
2
Participants
1
Timeline
4
Reactions
0
Author
stephenmcconnachiestephenmcconnachie
Participants
stephenmcconnachiestephenmcconnachie
Timeline (top)
commented ×2closed ×1labeled ×1

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): 64 On-line CPU(s) list: 0-63 Vendor ID: GenuineIntel Model name: Intel(R) Xeon(R) Gold 6426Y CPU family: 6 Model: 143 Thread(s) per core: 2 Core(s) per socket: 16 Socket(s): 2 Stepping: 8 BogoMIPS: 5000.00 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 tsc_known_freq 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 cat_l2 cdp_l3 intel_ppin cdp_l2 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 user_shstk avx_vnni avx512_bf16 wbnoinvd dtherm ida arat pln pts hfi vnmi avx512vbmi umip pku ospke waitpkg avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq la57 rdpid bus_lock_detect cldemote movdiri movdir64b enqcmd fsrm md_clear serialize tsxldtrk pconfig arch_lbr ibt amx_bf16 avx512_fp16 amx_tile amx_int8 flush_l1d arch_capabilities Virtualization: VT-x L1d cache: 1.5 MiB (32 instances) L1i cache: 1 MiB (32 instances) L2 cache: 64 MiB (32 instances) L3 cache: 75 MiB (2 instances) NUMA node(s): 2 NUMA node0 CPU(s): 0-15,32-47 NUMA node1 CPU(s): 16-31,48-63 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: 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; PBRSB-eIBRS SW sequence; BHI BHI_DIS_S Vulnerability Srbds: Not affected Vulnerability Tsa: Not affected Vulnerability Tsx async abort: Not affected Vulnerability Vmscape: Mitigation; IBPB before exit to userspace

Questions

  • Is this a known issue for gdn_prefill_sm90 on H100?
  • Is there a supported way to limit FlashInfer JIT parallelism?
  • Is there a way to disable/avoid this path as a workaround?

Code Example

==============================
        System Info
==============================
OS                           : Ubuntu 24.04.4 LTS (x86_64)
GCC version                  : (Ubuntu 13.3.0-6ubuntu2~24.04.1) 13.3.0
Clang version                : Could not collect
CMake version                : version 4.2.1
Libc version                 : glibc-2.39

==============================
       PyTorch Info
==============================
PyTorch version              : 2.10.0+cu128
Is debug build               : False
CUDA used to build PyTorch   : 12.8
ROCM used to build PyTorch   : N/A

==============================
      Python Environment
==============================
Python version               : 3.11.11 (main, Jan 14 2025, 22:49:08) [Clang 19.1.6 ] (64-bit runtime)
Python platform              : Linux-6.17.0-19-generic-x86_64-with-glibc2.39

==============================
       CUDA / GPU Info
==============================
Is CUDA available            : True
CUDA runtime version         : 12.8.93
CUDA_MODULE_LOADING set to   :
GPU models and configuration : GPU 0: NVIDIA H100 PCIe
Nvidia driver version        : 570.211.01
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):                                  64
On-line CPU(s) list:                     0-63
Vendor ID:                               GenuineIntel
Model name:                              Intel(R) Xeon(R) Gold 6426Y
CPU family:                              6
Model:                                   143
Thread(s) per core:                      2
Core(s) per socket:                      16
Socket(s):                               2
Stepping:                                8
BogoMIPS:                                5000.00
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 tsc_known_freq 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 cat_l2 cdp_l3 intel_ppin cdp_l2 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 user_shstk avx_vnni avx512_bf16 wbnoinvd dtherm ida arat pln pts hfi vnmi avx512vbmi umip pku ospke waitpkg avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq la57 rdpid bus_lock_detect cldemote movdiri movdir64b enqcmd fsrm md_clear serialize tsxldtrk pconfig arch_lbr ibt amx_bf16 avx512_fp16 amx_tile amx_int8 flush_l1d arch_capabilities
Virtualization:                          VT-x
L1d cache:                               1.5 MiB (32 instances)
L1i cache:                               1 MiB (32 instances)
L2 cache:                                64 MiB (32 instances)
L3 cache:                                75 MiB (2 instances)
NUMA node(s):                            2
NUMA node0 CPU(s):                       0-15,32-47
NUMA node1 CPU(s):                       16-31,48-63
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:             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; PBRSB-eIBRS SW sequence; BHI BHI_DIS_S
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Mitigation; IBPB before exit to userspace

==============================
Versions of relevant libraries
==============================
[pip3] flashinfer-python==0.6.6
[pip3] numpy==2.4.3
[pip3] nvidia-cublas-cu12==12.8.4.1
[pip3] nvidia-cuda-cupti-cu12==12.8.90
[pip3] nvidia-cuda-nvrtc-cu12==12.8.93
[pip3] nvidia-cuda-runtime-cu12==12.8.90
[pip3] nvidia-cudnn-cu12==9.10.2.21
[pip3] nvidia-cudnn-frontend==1.18.0
[pip3] nvidia-cufft-cu12==11.3.3.83
[pip3] nvidia-cufile-cu12==1.13.1.3
[pip3] nvidia-curand-cu12==10.3.9.90
[pip3] nvidia-cusolver-cu12==11.7.3.90
[pip3] nvidia-cusparse-cu12==12.5.8.93
[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==13.590.48
[pip3] nvidia-nccl-cu12==2.27.5
[pip3] nvidia-nvjitlink-cu12==12.8.93
[pip3] nvidia-nvshmem-cu12==3.4.5
[pip3] nvidia-nvtx-cu12==12.8.90
[pip3] pyzmq==27.1.0
[pip3] torch==2.10.0
[pip3] torch_c_dlpack_ext==0.1.5
[pip3] torchaudio==2.10.0
[pip3] torchvision==0.25.0
[pip3] transformers==5.3.0
[pip3] triton==3.6.0
[conda] Could not collect

==============================
         vLLM Info
==============================
ROCM Version                 : Could not collect
vLLM Version                 : 0.17.1
vLLM Build Flags:
  CUDA Archs: Not Set; ROCm: Disabled
GPU Topology:
        GPU0    CPU Affinity    NUMA Affinity   GPU NUMA ID
GPU0     X      0-15,32-47      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
==============================
LD_LIBRARY_PATH=/usr/local/cuda/lib64:
PYTORCH_NVML_BASED_CUDA_CHECK=1
TORCHINDUCTOR_COMPILE_THREADS=1
`TORCHINDUCTOR_CACHE_DIR=/tmp/torchinductor_datadigipres`

---

CUDA_VISIBLE_DEVICES=0 \
vllm serve Qwen/Qwen3.5-2B   \
--attention-backend FLASH_ATTN   \
--enforce-eager   \
--max-model-len 4096   \
--max-num-seqs 2   \
--swap-space 0   \
--gpu-memory-utilization 0.50

---

curl http://localhost:8000/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
    "model": "Qwen/Qwen3.5-2B",
    "messages": [{"role": "user", "content": "Describe the benefits of vLLM over Llama.cpp"}],
    "max_tokens": 100
  }'

---

~/.cache/flashinfer/0.6.6/90a/cached_ops/gdn_prefill_sm90/
/usr/local/cuda/bin/nvcc ...
/usr/local/cuda/bin/../nvvm/bin/cicc ...

---

Mem:   125Gi total, 117Gi used, 2.9Gi free, 8.1Gi available
Swap:  8.0Gi total, ~623Mi used

---

5885556 KiB
5774388 KiB
5437976 KiB
5399000 KiB
5372580 KiB

---

/usr/local/cuda -> /usr/local/cuda-12.8
nvcc 12.8.93

---

pkill -f 'nvcc|cicc|vllm'
rm -rf ~/.cache/flashinfer
mkdir -p ~/.cache/flashinfer
RAW_BUFFERClick to expand / collapse

Your current environment

<details> <summary>The output of <code>python collect_env.py</code></summary> 
==============================
        System Info
==============================
OS                           : Ubuntu 24.04.4 LTS (x86_64)
GCC version                  : (Ubuntu 13.3.0-6ubuntu2~24.04.1) 13.3.0
Clang version                : Could not collect
CMake version                : version 4.2.1
Libc version                 : glibc-2.39

==============================
       PyTorch Info
==============================
PyTorch version              : 2.10.0+cu128
Is debug build               : False
CUDA used to build PyTorch   : 12.8
ROCM used to build PyTorch   : N/A

==============================
      Python Environment
==============================
Python version               : 3.11.11 (main, Jan 14 2025, 22:49:08) [Clang 19.1.6 ] (64-bit runtime)
Python platform              : Linux-6.17.0-19-generic-x86_64-with-glibc2.39

==============================
       CUDA / GPU Info
==============================
Is CUDA available            : True
CUDA runtime version         : 12.8.93
CUDA_MODULE_LOADING set to   :
GPU models and configuration : GPU 0: NVIDIA H100 PCIe
Nvidia driver version        : 570.211.01
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):                                  64
On-line CPU(s) list:                     0-63
Vendor ID:                               GenuineIntel
Model name:                              Intel(R) Xeon(R) Gold 6426Y
CPU family:                              6
Model:                                   143
Thread(s) per core:                      2
Core(s) per socket:                      16
Socket(s):                               2
Stepping:                                8
BogoMIPS:                                5000.00
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 tsc_known_freq 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 cat_l2 cdp_l3 intel_ppin cdp_l2 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 user_shstk avx_vnni avx512_bf16 wbnoinvd dtherm ida arat pln pts hfi vnmi avx512vbmi umip pku ospke waitpkg avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq la57 rdpid bus_lock_detect cldemote movdiri movdir64b enqcmd fsrm md_clear serialize tsxldtrk pconfig arch_lbr ibt amx_bf16 avx512_fp16 amx_tile amx_int8 flush_l1d arch_capabilities
Virtualization:                          VT-x
L1d cache:                               1.5 MiB (32 instances)
L1i cache:                               1 MiB (32 instances)
L2 cache:                                64 MiB (32 instances)
L3 cache:                                75 MiB (2 instances)
NUMA node(s):                            2
NUMA node0 CPU(s):                       0-15,32-47
NUMA node1 CPU(s):                       16-31,48-63
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:             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; PBRSB-eIBRS SW sequence; BHI BHI_DIS_S
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Mitigation; IBPB before exit to userspace

==============================
Versions of relevant libraries
==============================
[pip3] flashinfer-python==0.6.6
[pip3] numpy==2.4.3
[pip3] nvidia-cublas-cu12==12.8.4.1
[pip3] nvidia-cuda-cupti-cu12==12.8.90
[pip3] nvidia-cuda-nvrtc-cu12==12.8.93
[pip3] nvidia-cuda-runtime-cu12==12.8.90
[pip3] nvidia-cudnn-cu12==9.10.2.21
[pip3] nvidia-cudnn-frontend==1.18.0
[pip3] nvidia-cufft-cu12==11.3.3.83
[pip3] nvidia-cufile-cu12==1.13.1.3
[pip3] nvidia-curand-cu12==10.3.9.90
[pip3] nvidia-cusolver-cu12==11.7.3.90
[pip3] nvidia-cusparse-cu12==12.5.8.93
[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==13.590.48
[pip3] nvidia-nccl-cu12==2.27.5
[pip3] nvidia-nvjitlink-cu12==12.8.93
[pip3] nvidia-nvshmem-cu12==3.4.5
[pip3] nvidia-nvtx-cu12==12.8.90
[pip3] pyzmq==27.1.0
[pip3] torch==2.10.0
[pip3] torch_c_dlpack_ext==0.1.5
[pip3] torchaudio==2.10.0
[pip3] torchvision==0.25.0
[pip3] transformers==5.3.0
[pip3] triton==3.6.0
[conda] Could not collect

==============================
         vLLM Info
==============================
ROCM Version                 : Could not collect
vLLM Version                 : 0.17.1
vLLM Build Flags:
  CUDA Archs: Not Set; ROCm: Disabled
GPU Topology:
        GPU0    CPU Affinity    NUMA Affinity   GPU NUMA ID
GPU0     X      0-15,32-47      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
==============================
LD_LIBRARY_PATH=/usr/local/cuda/lib64:
PYTORCH_NVML_BASED_CUDA_CHECK=1
TORCHINDUCTOR_COMPILE_THREADS=1
`TORCHINDUCTOR_CACHE_DIR=/tmp/torchinductor_datadigipres`
</details>

🐛 Describe the bug

On an H100 (sm90a), using vLLM 0.17.1, running inference against API after vllm serve triggers FlashInfer JIT compilation of:

~/.cache/flashinfer/0.6.6/90a/cached_ops/gdn_prefill_sm90/

During this step, many parallel nvcc jobs are launched, which fan out into many concurrent cicc processes. Each cicc process uses several GiB of host RAM, so total memory usage quickly grows until the host hits OOM.

On a machine with 125 GiB RAM, memory reached about 117 GiB used during this compile burst. The OOM killer then started killing unrelated processes before inference could start.

This looks like excessive JIT compile parallelism / memory usage in the FlashInfer gdn_prefill_sm90 path, rather than a generic CUDA initialization issue.

Environment

  • vllm==0.17.1
  • flashinfer-python==0.6.6
  • CUDA 12.8 (/usr/local/cuda -> /usr/local/cuda-12.8, nvcc 12.8.93)
  • GCC/G++ 13.3.0
  • Kernel 6.17.0-19-generic
  • GPU: H100

Exact command used for vLLM serve

CUDA_VISIBLE_DEVICES=0 \
vllm serve Qwen/Qwen3.5-2B   \
--attention-backend FLASH_ATTN   \
--enforce-eager   \
--max-model-len 4096   \
--max-num-seqs 2   \
--swap-space 0   \
--gpu-memory-utilization 0.50

Exact command used for vLLM inference

curl http://localhost:8000/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
    "model": "Qwen/Qwen3.5-2B",
    "messages": [{"role": "user", "content": "Describe the benefits of vLLM over Llama.cpp"}],
    "max_tokens": 100
  }'

Observed behavior

  • vllm serve / first inference triggers JIT compile of gdn_prefill_sm90
  • many nvcc processes spawn
  • these fan out into many cicc subprocesses
  • each cicc uses about 4–6 GiB RSS
  • host memory spikes until OOM
  • unrelated processes get killed

Example paths/processes seen:

~/.cache/flashinfer/0.6.6/90a/cached_ops/gdn_prefill_sm90/
/usr/local/cuda/bin/nvcc ...
/usr/local/cuda/bin/../nvvm/bin/cicc ...

Example memory snapshot during the spike:

Mem:   125Gi total, 117Gi used, 2.9Gi free, 8.1Gi available
Swap:  8.0Gi total, ~623Mi used

Representative cicc RSS values:

5885556 KiB
5774388 KiB
5437976 KiB
5399000 KiB
5372580 KiB

Expected behavior JIT compilation should not spawn enough parallel compiler processes to exhaust host RAM.

Notes

  • CUDA/toolchain appears aligned:
/usr/local/cuda -> /usr/local/cuda-12.8
nvcc 12.8.93
  • I cleared the FlashInfer cache and reproduced the same issue again:
pkill -f 'nvcc|cicc|vllm'
rm -rf ~/.cache/flashinfer
mkdir -p ~/.cache/flashinfer
  • The problem appears specific to the FlashInfer gdn_prefill_sm90 JIT path.

Questions

  • Is this a known issue for gdn_prefill_sm90 on H100?
  • Is there a supported way to limit FlashInfer JIT parallelism?
  • Is there a way to disable/avoid this path as a workaround?

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

Fix Plan

To address the issue of excessive JIT compile parallelism and memory usage in the FlashInfer gdn_prefill_sm90 path, consider the following steps:

  • Limit FlashInfer JIT parallelism: Set the TORCHINDUCTOR_COMPILE_THREADS environment variable to a lower value, e.g., TORCHINDUCTOR_COMPILE_THREADS=4, to reduce the number of concurrent compiler processes.
  • Increase swap space: Increase the swap space to provide more memory for the system to use during the JIT compilation process. This can be done by adding a swap file or increasing the size of the existing swap partition.
  • Disable gdn_prefill_sm90 path: As a workaround, try disabling the gdn_prefill_sm90 path by setting the FLASHINFER_DISABLE_GDN_PREFILL environment variable to 1. However, this may impact performance.

Example code to set environment variables:

import os

# Set TORCHINDUCTOR_COMPILE_THREADS to 4
os.environ['TORCHINDUCTOR_COMPILE_THREADS'] = '4'

# Set FLASHINFER_DISABLE_GDN_PREFILL to 1 (if available)
os.environ['FLASHINFER_DISABLE_GDN_PREFILL'] = '1'

Command to increase swap space:

# Add a

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…
#api#ssr#installation#tensor shape#retriever error#indexing error#inference speed#environment variable

Still need to ship something?

×6

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

Back to top recommendations

TRENDING

vllm - 💡(How to fix) Fix [Bug]: OOM during FlashInfer JIT compile of gdn_prefill_sm90 on H100 due to many concurrent cicc processes [2 comments, 1 participants]