igneum-network/igneum
1
0
Fork 0
Code Issues Pull requests Releases Activity

Prover floor: the 12 GB tier reads proves-alone (mine-and-prove not claimed, 9.9 GB over the 9.0 line), 16 GB mines and proves; aligned with the proving agent's table

Browse source 
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
This commit is contained in:
igneum-labs 2026-10-06 00:39:12 +00:00
parent d9599bce15
commit bf82dde2b1
2 changed files with 4 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
docs/analysis/prover-floor.md
View file

@ -177,7 +177,7 @@ threshold 16,851 MiB (20,516 stock); the server after Setup 6,535 MiB (v1 9,703,
| Card | Stock SP1 6.8.1 server | The v3 server (this branch), measured on the 5090's allocation | Profile |
|---|---|---|---|
| 8 GB | refused (the 24 GB panic) | not measured; the empty shard alone is 9.9 GB measured (7.9 GB the server's own), so no | mine only |
| 12 GB | refused | the v1 shard alone 10.3 GB measured (8.2 GB the server's own) 5.7 s; beside the miner 8.2 GB own + the miner's 1.7 GB, 24.4 s (sweep 4); the card's own display not measured | `SP1_GPU_ELEMENT_THRESHOLD=67108864`, mine and prove (1.4 to 1.9 GB spare on paper) |
| 12 GB | refused | the v1 shard alone 10.3 GB measured (8.2 GB the server's own) 5.7 s; beside the miner 8.2 GB own + the miner's 1.7 GB, 24.4 s (sweep 4); the card's own display not measured | `SP1_GPU_ELEMENT_THRESHOLD=67108864`, PROVES ALONE; mine-and-prove NOT claimed (8.2 + 1.7 = 9.9 GB before the display, over the 9.0 GB line the proving agent holds for the tier) |
| 16 GB | refused | the v1 shard alone 12.9 GB measured (10.95 own) at 2^27, 4.3 s; beside the miner 10.95 own + 1.7, 17.4 s (sweep 4) | 2^27, mine and prove (2.9 GB spare on paper) |
| 24 GB | the v1 shard alone (20.4 GB), never the prototype shard (28.3) | the v1 shard 12.9 GB at 2^27, the prototype shard 13.5 GB (16.8 s) | upstream's 24 GB threshold or 2^27 |
| 32 GB | everything (28.3 GB for the prototype shard beside the miner at 30.1) | 16.9 GB at upstream's threshold | unchanged |

5
docs/bench-log.md
View file

@ -1587,8 +1587,9 @@ against 8,202 alone), the miner costs 4.3x in time (24.4 s against 5.7 s a v1 sh
by 25x), and 2^25 buys no memory (12,066 MiB again at 43.5 s: the floor is now the Setup keys and the recursion
stage, 4.4 GB own after Setup plus about 3.8 GB at the recursion peak). So a 12 GB card (12,288 MiB) mining and
proving holds 8.2 GB of server plus 1.7 GB of miner plus its own display (0.5 to 1 GB, not measured): 10.4 to
10.9 GB, 1.4 to 1.9 GB spare at 2^26, and the "under 9.0 GB" row (the server's own) is met at 8.2; a 16 GB card
mines and proves at 2^27 (10.95 + 1.7 GB, 2.9 GB spare) at 17 s. All of it the 5090's allocation pattern until the
10.9 GB on paper, but 8.2 + 1.7 = 9.9 GB before the display is over the 9.0 GB line the proving agent holds for the
tier, so the 12 GB tier reads "proves alone" (10.3 GB, 5.7 s) and mine-and-prove is not claimed there; a 16 GB
card mines and proves at 2^27 (10.95 + 1.7 GB = 12.7 plus the display, under its 15.0 GB line) at 17.4 s. All of it the 5090's allocation pattern until the
on-order 12 GB card runs the same two points. Shipping it is the project's own signed build of SP1's prover at
every SP1 upgrade (the proving plan's packaging row before 0.3.12).