Filter public library records by library, version, chip and outcome. Measured records describe recorded attempts. Reported records describe outside sources. Each result guides a fresh Verify on your configuration.

100 library records shown (up to 100 matches)

Reset library filters

Reported record

llvm-gpu-backends

Chip unspecified

github issue

reported fix: AMDGPU MIR serialization missing handling of MinNumAGPRs

Source evidence →

Reported fix Reported by an outside source. Fresh verification required. 2025-10-07T00:00:02Z

Reported record

openvino

Chip unspecified

github issue

closed without a linked fix: [Performance]: 265k NPU much slower than CPU at LLM inference

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-07T04:38:33Z

Reported record

torchtitan

Chip unspecified

github issue

reported fix: ReordererSequenceParallel (ETP=1) doesn't work with torch.compile + AC

Source evidence →

Reported fix Reported by an outside source. Fresh verification required. 2025-10-07T05:18:27Z

Reported record

llama-cpp 1.3.275

Chip unspecified

github issue

closed without a linked fix: Misc. bug: Slower performance on newer models Apriel-1.5-15b, granite-4.0-h-tiny

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-07T05:34:36Z

Reported record

syzkaller

Chip unspecified

github issue

open issue: syzbot, syz-cluster: avoid duplication of base syzkaller configs

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-07T12:30:43Z

Reported record

liger-kernel

Chip unspecified

github issue

open issue: Support for Qwen3-VL models

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-07T13:26:40Z

Reported record

iree

Chip unspecified / gfx950

github issue

open issue: [Codegen] Buffer read across devices instead of copied

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-07T14:36:50Z

Reported record

vllm

Chip unspecified

github issue

closed without a linked fix: [Performance]: Use int over list[int] as output_tokens to reduce GC overhead

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-07T17:23:35Z

Reported record

flash-attention

Chip unspecified / sm_100

github issue

open issue: [CuteDSL] ICE for flash combine kernel on blackwell

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-07T18:37:28Z

Reported record

bpftrace

Chip unspecified

github issue

closed without a linked fix: Fields of anonymous structures in arrays can't be accessed (BTF)

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-07T20:08:14Z

Reported record

tritonbench

H100

github issue

closed without a linked fix: [AMD] Validate perf numbers on AMD MI350

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-07T20:31:25Z

Reported record

nvidia-open-gpu-kernel-modules

Chip unspecified

github issue

open issue: NVRM: memmgrRestorePowerMgmtState_KERNEL: !!!!! Calling Resume on an active GPU or the previous Suspend call might have failed !!!!!!

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-07T20:54:33Z

Reported record

iree

Chip unspecified

github issue

open issue: [Codegen][GPU] Refactor and consolidate padding passes and configurations

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-07T21:26:59Z

Reported record

xla 580.95.05

B200

github issue

open issue: CUDNN jax.nn.dot_product_attention fails on CUDA 13

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:04:07Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_logging is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:26:52Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_metrics is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:26:55Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_report_handler is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:26:59Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_result is flaky

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:01Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_serialization is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:04Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_storage is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:07Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_sync_actor is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:09Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_thread_runner is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:11Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_torch_trainer is flaky

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:13Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_util is flaky

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:16Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_validation_manager is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:23Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_worker is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:26Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_worker_group_poll_status is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:28Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/tune:test_actor_reuse is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:28Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:lightning_exp_tracking_mlflow is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:35Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:lightning_exp_tracking_tensorboard is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:37Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:mlflow_simple_example is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:39Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_api_migrations is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:41Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_checkpoint is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:43Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_checkpoint_manager is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:46Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_e2e_wandb_integration is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:48Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_lightgbm_predictor is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:51Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_predictor is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:53Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_session is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:55Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_storage is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:57Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_tensorflow_checkpoint is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:27:59Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_torch_checkpoint is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:28:02Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_torch_utils is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:28:04Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_train_head is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:28:07Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_utils is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:28:09Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_xgboost_predictor is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:28:11Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_state_export is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:29:57Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_telemetry is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:29:59Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:tensorflow_regression_example is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:30:16Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_base_trainer is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:31:19Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/tune:test_actor_caching is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:31:32Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://doc:doctest[train] is consistently_failing

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T00:31:41Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train:test_torch_predictor is flaky

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T05:45:39Z

Reported record

openblas

Chip unspecified

github issue

closed without a linked fix: Thread Usage Limitation in OpenBLAS on Windows on ARM64 Setup

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T05:46:25Z

Reported record

ray

Chip unspecified

github issue

closed without a linked fix: CI test linux://python/ray/train/v2:test_torch_gpu is flaky

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T05:52:58Z

Reported record

triton-inference-server

Chip unspecified

github issue

closed without a linked fix: Evaluating Whisper Large V3 Concurrency on TensorRT Deployment

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T06:00:53Z

Reported record

tvm 0.21.0

Chip unspecified

github issue

closed without a linked fix: [bug][relax.frontend.torch] FFI segfault in tvm::relax::Tuple::Tuple when importing torch.export graph with 4D advanced-indexing write (aten.index_put_) and tuple outputs

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T09:28:36Z

Reported record

litert 2.0.2

Chip unspecified

github issue

closed without a linked fix: Support 32-bit ARM (armeabi-v7a) ABI for LiteRT Next on legacy Android devices

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T11:25:28Z

Reported record

flashmla

Chip unspecified

github issue

closed without a linked fix: [QST] Why sparse flashmla prefill only support batch_size=1？

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T11:44:58Z

Reported record

dra-driver-nvidia-gpu

Chip unspecified

github issue

closed without a linked fix: CD daemon teardown may leave behind stale IMEX daemon (`Address already in use`)

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T12:15:21Z

Reported record

verl

Chip unspecified

github issue

closed without a linked fix: Multi-turn tool use RL training with SGLang has incorrect response trajectory

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T13:59:20Z

Reported record

sglang 0.5.1

Chip unspecified

github issue

closed without a linked fix: [Bug] Importing `sgl_kernel` now hard-requires CUDA toolkit (regression vs `v0.5.1`)

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T14:08:49Z

Reported record

llvm-gpu-backends

Chip unspecified

github issue

reported fix: [AMDGPU] `sort` in `AMDGPURewriteAGPRCopyMFMAImpl::eliminateSpillsOfReassignedVGPRs()` needs to have strict weak ordering

Source evidence →

Reported fix Reported by an outside source. Fresh verification required. 2025-10-08T14:27:22Z

Reported record

dra-driver-nvidia-gpu

Chip unspecified

github issue

open issue: System Mediated MPS-sharing of GPUs

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T14:48:03Z

Reported record

dra-driver-nvidia-gpu

Chip unspecified

github issue

closed without a linked fix: Publish DeviceTaints to report / track GPU health

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T14:53:00Z

Reported record

megatron-lm

Chip unspecified

github issue

closed without a linked fix: Automatically select training micro batch size based on GPU memory

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T18:49:40Z

Reported record

nccl-tests

H100

github issue

closed without a linked fix: Multi-node NCCL performance drop due to cross numa issues

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-08T22:08:48Z

Reported record

triton-inference-server

Chip unspecified

github issue

closed without a linked fix: Python backend stub is becoming zombie process

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T00:42:16Z

Reported record

gptqmodel

Chip unspecified

github issue

open issue: How to quantize both vision encoder and llm together?

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T02:24:04Z

Reported record

sglang 0.5.3rc1

H20

github issue

closed without a linked fix: [Bug] Server hang during Kimi-K2 serving ( KV-Cache flushed)

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T02:43:29Z

Reported record

verl 535.129.03

A100

github issue

open issue: NCCL Broadcast Timeout in Megatron-LM Training of Qwen3-235B with Multi-Turn Tool Calling (max_turns=5)

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T03:39:22Z

Reported record

flashinfer

Chip unspecified

github issue

closed without a linked fix: [Bug] flashinfer v0.4.0 jit failure

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T04:16:30Z

Reported record

deepep

B200

github issue

closed without a linked fix: Trying to understand performance

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T04:20:56Z

Reported record

litert

Chip unspecified

github issue

closed without a linked fix: [LiteRT Play Services] inference on NPU

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T06:28:36Z

Reported record

tensorrt 10.13.3

Chip unspecified

github issue

open issue: Could not find any implementation for node {ForeignNode[/model/Unsqueeze + /model/Unsqueeze_1.../Cast]} failure of TensorRT 10.13.3 when running trtexec on GPU 3090

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T06:36:04Z

Reported record

llama-cpp 12.2.0

Chip unspecified

github issue

closed without a linked fix: bug: segfault on q4_0 repacking on windows avx2 cpu only

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T06:48:02Z

Reported record

deepgemm

Chip unspecified

github issue

closed without a linked fix: Query on performance-modeling of MoE Grouped GEMM

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T07:26:01Z

Reported record

tensorrt-llm 1.2.0rc1

B200

github issue

closed without a linked fix: [Bug]: EXAONE 4.0 FP8 trtllm-serve failure

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T08:24:33Z

Reported record

intel-xpu-triton 1.6.35096

Chip unspecified

github issue

closed without a linked fix: [pytorch upstream] softmax bwd on BMG is slow

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T08:39:01Z

Reported record

torchtitan

Chip unspecified

github issue

closed without a linked fix: More robust distributed loss

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T09:23:48Z

Reported record

dcgm

Chip unspecified

github issue

open issue: Can DCGM independently view the health information of the CPU？

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T09:36:55Z

Reported record

syzkaller

Chip unspecified

github issue

reported fix: dashboard: more graceful error handling for user commands

Source evidence →

Reported fix Reported by an outside source. Fresh verification required. 2025-10-09T09:57:35Z

Reported record

flashmla

Chip unspecified

github issue

open issue: Missing sparse backward kernel.

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T09:59:27Z

Reported record

llm-compressor

Chip unspecified

github issue

closed without a linked fix: Any example for static fp8 quantization?

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T11:30:53Z

Reported record

esmi-ib-library 24.04.3

Chip unspecified

github issue

open issue: Trouble loading HSMP module into Kernel

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T11:47:45Z

Reported record

llm-compressor

Chip unspecified

github issue

closed without a linked fix: Can't using llmcompressor for awq quantization and using autoawq completely replace each other?

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T12:21:34Z

Reported record

open-mpi

Chip unspecified

github issue

open issue: segfault due to thread-unsafe getenv() call

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T13:09:07Z

Reported record

tensorrt-llm

Chip unspecified

github issue

closed without a linked fix: [AutoDeploy]: Improve MoE performance for TP>1

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T15:15:46Z

Reported record

tensorrt-llm

Chip unspecified

github issue

open issue: [Feature]: FP8 for DS-R1

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T15:36:52Z

Reported record

nemo-speech

Chip unspecified

github issue

closed without a linked fix: Accuracy issue with MALSD decoding with LM fusion

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T16:26:58Z

Reported record

megatron-lm

Chip unspecified

github issue

open issue: Empty pipeline stages deadlock with custom pipeline-parallel-layout

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T17:23:12Z

Reported record

vllm

B200

github issue

closed without a linked fix: [Performance]: Deepseek3.2 Compile Time Regression on B200 with Pytorch 2.9

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T17:46:55Z

Reported record

iree

Chip unspecified / gfx950

github issue

open issue: [gfx950][mxfp4] Inefficient codegen for undersized tiles

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T18:26:03Z

Reported record

flashinfer

Chip unspecified

github issue

closed without a linked fix: JIT install in non-editable mode doesn't work?

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T20:11:48Z

Reported record

verl

Chip unspecified

github issue

closed without a linked fix: vllm sleep mode patch for AMD GPU no longer works

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T21:13:28Z

Reported record

mlx

github issue

reported fix: [BUG] CUDA conv fails to find execution plan

Source evidence →

Reported fix Reported by an outside source. Fresh verification required. 2025-10-09T22:39:11Z

Reported record

bpftrace 0.21.0

Chip unspecified

github issue

closed without a linked fix: regression attaching tracepoint:syscalls:sys_enter_landlock_add_rule in v0.24 (field has incomplete type)

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-09T22:58:22Z

Reported record

nemo-speech

GB200

github issue

closed without a linked fix: Seeing a performance drop when running Nemotron 340B workload using container: nvcr.io/nvidia/nemo:25.09 compared to when using container nvcr.io/nvidia/nemo:25.07.01 on GB200 and B200 systems.

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-10T01:45:58Z

Reported record

nccl-tests

B200

github issue

open issue: can ncu profiling peer memory traffic

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-10T02:20:43Z

Reported record

verl 3.11.9

H20

github issue

closed without a linked fix: vllm hang when use lora with qwen2.5-72b

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-10T02:30:37Z

Reported record

open-mpi

Chip unspecified

github issue

open issue: with clang 21.1.6 as wrapper compiler for cuda-aware openmpi-5.0.9 and compile options -fopenmp and -fopenmp-targets=nvptx64-nvidia-cuda one gets the message "The call to cuMemHostRegister(0x7fb38c000000, 134217728, 0) failed." for every code with OpenMPI [Patch is proposed as temporary solution in the last comment]

Source evidence →

Held Reported by an outside source. Fresh verification required. 2025-10-10T03:01:38Z

CHOOSE A TARGET

Build toward the hardware you need.

Explore 46 maker entries across cloud, data centre and edge. Register a specific chip for qualification; availability and support are confirmed during scoping.