Speed sweep

ling-3-0-flash-q4-0-rocmfp4-strix-lean-radeon-ai-pro-r9700-32gb-llama-cpp-tp3-sweep

ling-3-0-flash-q4-0-rocmfp4-strix-lean-radeon-ai-pro-r9700-32gb-llama-cpp-tp3-sweep

Record

Recipe
ling-3-0-flash-q4-0-rocmfp4-strix-lean-radeon-ai-pro-r9700-32gb-llama-cpp-tp3
Measured
2026-08-05T20:51:18.428Z
Points
1

Measured speed

ConcurrencyContextPrefillDecodeTTFT msStatus
14,096542.639.3observed

Remaining fields

Identity, launch, related records, and measured speed are shown above. This is the rest of the normalized record.

id
ling-3-0-flash-q4-0-rocmfp4-strix-lean-radeon-ai-pro-r9700-32gb-llama-cpp-tp3-sweep
measured at
2026-08-05T20:51:18.428Z
recipe id
ling-3-0-flash-q4-0-rocmfp4-strix-lean-radeon-ai-pro-r9700-32gb-llama-cpp-tp3
schema version
local-ai-registry/v1

metrics

concurrency
1
inference engine version
ROCmFPX db6844d + BailingMoe3 port
latest point at
2026-08-05T20:51:18.428Z
max context tokens
4,096
peak generation tps
39.31
peak prompt tps
542.58
point count
1

rows

concurrencycontext tokensdecode tok sdecode tok s per streamoutput tokenspeak vram gbprefill tok ssamples
14,09639.3139.310Unknown542.581
Provenance & metadata (1)