GPU ATTACHÉ CASE

S.T.A.R.S. // COMPUTE OBSERVATORY
FILE 09 / ASH
READ-ONLY · 3 SECOND FEED

TELEMETRY UPLINK

CONNECTINGWaiting for the first verified sample.
Captured UTC
Sample age
Unavailable
Sequence
Unavailable

Polling every 3 seconds while this page is visible.

CURRENT RUNNER

Runner
Unavailable
Model
Unavailable
Serving status
Unavailable
Input capabilities
Unavailable
Context window
Unavailable
Active gateway requests
Unavailable

Readiness is not active generation. Gateway requests are not a count of GPU jobs.

DEVICE CASES // ALLOCATION & UTILIZATION

Awaiting measured device telemetry. No allocation or utilization is assumed.

HOST RAM // SEPARATE BUDGET

Used / total
Unavailable
Available
Unavailable

Used is total minus available; available RAM includes reclaimable memory.

READ THE COUNTERS, NOT THE SILHOUETTES

VRAM allocation and GPU utilization are different measurements. Allocation includes all device allocations, not a measured per-model partition. Each case has 32 visual tiles, each representing 1/32 of that device's capacity.

Unallocated VRAM is not authorization to run another workload. This page exposes no inference or operational controls.