sigcue
← Reported problems
Reported problemNot a verified signal

Multi-model loads evict each other on AMD Strix Halo (gfx1151): scheduler caps available VRAM at host RAM free instead of the VRAM carveout

Multi-model loads evict each other on AMD Strix Halo (gfx1151): scheduler caps available VRAM at host RAM free instead of the VRAM carveout

SourceGitHub
Communityollama/ollama
DateJun 14, 2026
Independent reports1

Important limitations

  • This may be one public report, not widespread demand.
  • Current independent report count: 1.
  • This is currently a single independent report. It has not passed clustering or human review as a verified signal.
  • No market size, willingness to pay, competition, or trend claims are inferred here.

What was reported

Who is affected
Application developer using the reported software
Job to be done
Deploy a service reliably
Workflow
0.30+ (Container Deployment)" and #16529 (open, 2026-06-04) "AMD Strix Halo
Desired outcome
Both models stay resident: ~24 GiB + ~24 GiB ≈ 48 GiB easily fits in the 96 GiB carveout.
Severity
Not enough evidence yet.
Urgency
Not enough evidence yet.

Evidence indicators

Evidence records1
Independent reports1
Reproduction notesPresent
Metrics mentionedPresent
Logs mentionedPresent
Corroborationuncorroborated

What would upgrade this

  • Additional independent reports from distinct communities or authors.
  • Human clustering into a review signal, then explicit publication approval.
  • Counter-evidence review and durable original-source verification.

Original source

Open the original public report