== 13fa57c4ee40 2026-09-29T03:12:32Z iab/src
03:13:39 INFO    TKV recent ring: w=128 slack=2 for 4 sequences (1024 rows); 16 attention layer(s) read it through 16 ring layer(s); 32.53 MiB in scratch.attn_codec, 8.132 MiB per sequence.
03:13:44 INFO    TKV recent ring: w=128 slack=2 for 4 sequences (1024 rows); 16 attention layer(s) read it through 16 ring layer(s); 32.53 MiB in scratch.attn_codec, 8.132 MiB per sequence.
03:18:42 INFO    TKV recent ring: w=128 slack=2 for 4 sequences (1024 rows); 16 attention layer(s) read it through 16 ring layer(s); 32.53 MiB in scratch.attn_codec, 8.132 MiB per sequence.
03:19:49 INFO    served context: 262144 tokens — the realized KV pool (1212 servable pages, prefix 7, 0-page margin) confirms the sizing bound.
03:19:51 INFO    engine ready, accepting requests on 0.0.0.0:8000
== ea41aeda1802 2026-09-29T03:20:33Z tree_off
03:22:50 INFO    served context: 262144 tokens — the realized KV pool (1219 servable pages, prefix 7, 0-page margin) confirms the sizing bound.
03:22:52 INFO    engine ready, accepting requests on 0.0.0.0:8000
== 89b90b2473e5 2026-09-29T03:25:23Z iab/src
03:25:37 INFO    TKV recent ring: w=128 slack=2 for 4 sequences (1024 rows); 16 attention layer(s) read it through 16 ring layer(s); 32.53 MiB in scratch.attn_codec, 8.132 MiB per sequence.
03:25:42 INFO    TKV recent ring: w=128 slack=2 for 4 sequences (1024 rows); 16 attention layer(s) read it through 16 ring layer(s); 32.53 MiB in scratch.attn_codec, 8.132 MiB per sequence.
03:26:02 INFO    TKV recent ring: w=128 slack=2 for 4 sequences (1024 rows); 16 attention layer(s) read it through 16 ring layer(s); 32.53 MiB in scratch.attn_codec, 8.132 MiB per sequence.
03:26:15 INFO    served context: 262144 tokens — the realized KV pool (1212 servable pages, prefix 7, 0-page margin) confirms the sizing bound.
03:26:17 INFO    engine ready, accepting requests on 0.0.0.0:8000
== bded2d175be9 2026-09-29T03:31:14Z tree_null
03:31:39 INFO    TKV recent ring: w=128 slack=2 for 4 sequences (1024 rows); 16 attention layer(s) read it through 16 ring layer(s); 32.53 MiB in scratch.attn_codec, 8.132 MiB per sequence.
03:31:45 INFO    TKV recent ring: w=128 slack=2 for 4 sequences (1024 rows); 16 attention layer(s) read it through 16 ring layer(s); 32.53 MiB in scratch.attn_codec, 8.132 MiB per sequence.
03:32:16 INFO    TKV recent ring: w=128 slack=2 for 4 sequences (1024 rows); 16 attention layer(s) read it through 16 ring layer(s); 32.53 MiB in scratch.attn_codec, 8.132 MiB per sequence.
03:32:31 INFO    served context: 262144 tokens — the realized KV pool (1212 servable pages, prefix 7, 0-page margin) confirms the sizing bound.
03:32:33 INFO    engine ready, accepting requests on 0.0.0.0:8000
