Download amd_mi250x/runtime_amendment_02.json from kobzaond/RLVRAMBench: direct link, hf CLI and curl.
- Browser
- Download file 5.03 kB
-
https://huggingface.co/datasets/kobzaond/RLVRAMBench/resolve/main/amd_mi250x/runtime_amendment_02.json
- Command line
-
hf download hf://datasets/kobzaond/RLVRAMBench/amd_mi250x/runtime_amendment_02.json
-
curl -L -o runtime_amendment_02.json https://huggingface.co/datasets/kobzaond/RLVRAMBench/resolve/main/amd_mi250x/runtime_amendment_02.json
5.03 kB
| { | |
| "amendment": "runtime-amendment-02", | |
| "recorded": "2026-10-21", | |
| "defects": { | |
| "shared_tmp": "As amendment 01, for jobs that were already running when it was recorded.", | |
| "shm_psm_missing": "vLLM start-up failed because a psm_* POSIX shared-memory segment in the node-wide /dev/shm disappeared (FileNotFoundError). The cause was not established; Singularity cannot bind a private directory onto /dev, so no fix is applied.", | |
| "port_race": "vLLM's TCPStore failed to bind an ephemeral port taken by a co-located job. run_grpo_mi250x.slurm now sets a job-specific VLLM_PORT base." | |
| }, | |
| "not_rerun": "Job 23032954 ('Memory usage increased after sleeping', a vLLM sleep-mode self-check) shows none of these defects and keeps its recorded non-memory outcome. Diagnosed memory failures are never rerun.", | |
| "settings_changed": false, | |
| "attempts_per_slot_limit": "one rerun under this amendment", | |
| "evidence": { | |
| "hardware_transfer-amd-9680c83e3c59777b-grpo-lora-s113": { | |
| "job_id": "23032950", | |
| "defect": "vLLM psm_* shared-memory segment missing at start-up" | |
| }, | |
| "hardware_transfer-amd-624145edbc91703d-grpo-lora-s83": { | |
| "job_id": "23030167", | |
| "defect": "shared /tmp collision partner: timed out before step 1 under the pre-fix harness" | |
| }, | |
| "hardware_transfer-amd-b26b51d5745e7038-grpo-lora-s94": { | |
| "job_id": "23030168", | |
| "defect": "shared /tmp collision partner: timed out before step 1 under the pre-fix harness" | |
| }, | |
| "hardware_transfer-amd-b26b51d5745e7038-grpo-lora-s95": { | |
| "job_id": "23030169", | |
| "defect": "shared /tmp ZMQ socket collision (invalid device pointer)" | |
| }, | |
| "hardware_transfer-amd-b26b51d5745e7038-grpo-lora-s96": { | |
| "job_id": "23030170", | |
| "defect": "shared /tmp collision partner: timed out before step 1 under the pre-fix harness" | |
| }, | |
| "hardware_transfer-amd-d89e11c1db30d91c-grpo-lora-s112": { | |
| "job_id": "23030190", | |
| "defect": "shared /tmp ZMQ socket collision (invalid device pointer)" | |
| }, | |
| "hardware_transfer-amd-d89e11c1db30d91c-grpo-lora-s113": { | |
| "job_id": "23030191", | |
| "defect": "shared /tmp collision partner: timed out before step 1 under the pre-fix harness" | |
| }, | |
| "hardware_transfer-amd-9e2e15f277fffe36-grpo-lora-s111": { | |
| "job_id": "23030211", | |
| "defect": "shared /tmp ZMQ socket collision (invalid device pointer)" | |
| }, | |
| "hardware_transfer-amd-9e2e15f277fffe36-grpo-lora-s112": { | |
| "job_id": "23030212", | |
| "defect": "shared /tmp collision partner: timed out before step 1 under the pre-fix harness" | |
| }, | |
| "hardware_transfer-amd-996a3920c206da4c-grpo-lora-s113": { | |
| "job_id": "23030253", | |
| "defect": "shared /tmp collision partner: timed out before step 1 under the pre-fix harness" | |
| }, | |
| "hardware_transfer-amd-bc00117c458b5d5d-grpo-lora-s91": { | |
| "job_id": "23030266", | |
| "defect": "shared /tmp ZMQ socket collision (invalid device pointer)" | |
| }, | |
| "hardware_transfer-amd-75ad17324769e2b3-grpo-lora-s94": { | |
| "job_id": "23030275", | |
| "defect": "vLLM psm_* shared-memory segment missing at start-up" | |
| }, | |
| "hardware_transfer-amd-75ad17324769e2b3-grpo-lora-s95": { | |
| "job_id": "23030276", | |
| "defect": "shared /tmp collision partner: timed out before step 1 under the pre-fix harness" | |
| }, | |
| "hardware_transfer-amd-75ad17324769e2b3-grpo-lora-s96": { | |
| "job_id": "23030277", | |
| "defect": "shared /tmp ZMQ socket collision (invalid device pointer)" | |
| }, | |
| "hardware_transfer-amd-4b92cc3197922be7-grpo-lora-s113": { | |
| "job_id": "23030284", | |
| "defect": "vLLM psm_* shared-memory segment missing at start-up" | |
| }, | |
| "hardware_transfer-amd-211b84c6cb8b5347-grpo-lora-s42": { | |
| "job_id": "23032991", | |
| "defect": "TCP port race with a co-located job" | |
| }, | |
| "reservation_sweep-amd-00fc9f14ba4ae94a-grpo-lora-s301": { | |
| "job_id": "23034023", | |
| "defect": "TCP port race with a co-located job" | |
| } | |
| }, | |
| "rerun_run_ids": [ | |
| "hardware_transfer-amd-9680c83e3c59777b-grpo-lora-s113", | |
| "hardware_transfer-amd-624145edbc91703d-grpo-lora-s83", | |
| "hardware_transfer-amd-b26b51d5745e7038-grpo-lora-s94", | |
| "hardware_transfer-amd-b26b51d5745e7038-grpo-lora-s95", | |
| "hardware_transfer-amd-b26b51d5745e7038-grpo-lora-s96", | |
| "hardware_transfer-amd-d89e11c1db30d91c-grpo-lora-s112", | |
| "hardware_transfer-amd-d89e11c1db30d91c-grpo-lora-s113", | |
| "hardware_transfer-amd-9e2e15f277fffe36-grpo-lora-s111", | |
| "hardware_transfer-amd-9e2e15f277fffe36-grpo-lora-s112", | |
| "hardware_transfer-amd-996a3920c206da4c-grpo-lora-s113", | |
| "hardware_transfer-amd-bc00117c458b5d5d-grpo-lora-s91", | |
| "hardware_transfer-amd-75ad17324769e2b3-grpo-lora-s94", | |
| "hardware_transfer-amd-75ad17324769e2b3-grpo-lora-s95", | |
| "hardware_transfer-amd-75ad17324769e2b3-grpo-lora-s96", | |
| "hardware_transfer-amd-4b92cc3197922be7-grpo-lora-s113", | |
| "hardware_transfer-amd-211b84c6cb8b5347-grpo-lora-s42", | |
| "reservation_sweep-amd-00fc9f14ba4ae94a-grpo-lora-s301" | |
| ] | |
| } | |