Karti Tripathi
91a47fb42c
ci / rust (push) Successful in 2m26s
Governed compute for unified-memory AI hardware — the machines where CPU and GPU share one pool and there is no separate VRAM allocation to bounce off. Over-commit that pool and the box thrashes and wedges, SSH and ping included, before the OOM killer gets a turn. Compute does not run inference. It supervises the servers that do: - Admission control. A model starts only if committed + requested + margin fits the budget. The refusal is the feature. - A 1 Hz watchdog on MemAvailable that stops the newest model before thrash. - Scenes: named sets of models activated as one transactional unit, with pre-flight validation and rollback to the previously active Scene on failure. Scenes reference model ids, never weight paths or commands, so a Scene obtained from elsewhere cannot introduce code. - Process ownership bound to (boot_id, pid, start_time_ticks, pgid == pid), so a reused PID can never be group-killed. - A protocol-transparent TCP gateway, so clients keep one address while model runtimes move behind it. - An MCP server, so agents drive the node as tools rather than as a CLI. One binary, six direct dependencies, no async runtime outside the MCP surface. Published from the internal monorepo with a fresh history. The private development tree keeps its own history; nothing here carries it.
22 lines
686 B
YAML
22 lines
686 B
YAML
apiVersion: lumbridge/v1
|
|
kind: EvalSuite
|
|
metadata:
|
|
name: performance
|
|
version: 1
|
|
description: "Repeated streamed requests measuring TTFT, client-observed prefill, and decode throughput."
|
|
tags: [performance, latency, throughput]
|
|
defaults:
|
|
max_tokens: 256
|
|
temperature: 0.0
|
|
repeat: 3
|
|
system: "Answer directly in plain text."
|
|
cases:
|
|
- id: short-prefill
|
|
category: latency
|
|
prompt: "Explain why unified-memory admission control prevents system thrashing. Give a detailed answer."
|
|
max_tokens: 256
|
|
- id: structured-decode
|
|
category: throughput
|
|
prompt: "Write twenty numbered, one-sentence operational checks for an AI inference server."
|
|
max_tokens: 384
|