-
Story
-
Resolution: Unresolved
-
Undefined
-
None
-
13
-
False
-
-
False
-
-
Investigate AI workloads for power usage results, aspirationally allowing comparisons between GPU vs CPU runtime environments (time to solution vs power consumption).
If a candidate workload is identified, incorporate a GPU related workload into Power Usage automation.
CPU-only inference looks to be a reasonable initial target.
- is cloned by
-
RHELPERF-138 Integrate Power Usage, PCPrecord_systemd into LLM test harness
-
- In Progress
-
-
RHELPERF-127 Investigate vLLM podman build issues
-
- Closed
-
- mentioned in
-
Page Loading...