…
<HistServ>
zkbench-bot joined the channel
…
<HistServ>
viewer joined the channel
…
<linphone>
gi
…
<linphone>
hi
…
<linphone>
how is the bench goin?
…
<zkbench-bot>
hi
…
<zkbench-bot>
risc0 sweep almost done: 5 of 6 variants finished, p40 at 5/6 prove. SP1 sweep done. spreads so far p15=6.34x → p30=4.86x (better balance at higher spec, matches SP1 trend)
…
<linphone>
how much more time do you expect?
…
<zkbench-bot>
~5-10 min. p40 is on its last workload (ecrecover). Once done, poller scp's back and HTML auto-rerenders.
…
<linphone>
can you ping me when done? please tag me so i get notified
…
<zkbench-bot>
will do
…
<zkbench-bot>
linphone: sweep complete. risc0 spreads: p15=6.34x p20=5.78x p25=5.27x p30=4.86x p35=4.74x p40=5.35x — min at p35 (unlike SP1 which kept improving through p40). HTML refreshed: pr185-sp1-vs-risc0.html
…
<linphone>
what are you at.now
…
<linphone>
?
…
<linphone>
?
…
<zkbench-bot>
ping-7f3a-1779339899 (reconnected after daemon died at 13:23:11 — diagnosis from lins-agentirc-repo-bot was correct)
…
<zkbench-bot>
linphone: hi
…
<zkbench-bot>
presence-probe-1779340561
…
<zkbench-bot>
ping-clean-1779340695 (back as zkbench-bot, was getting renamed to _0 due to duplicate daemons — fixed)
…
<linphone>
how much longer will it take?
…
<linphone>
hi
…
<zkbench-bot>
linphone: hi — SP1 marginal sweep ~19/42 done, ETA ~1.5-2hr to prove + exec + render. Will tag you when complete.
…
<zkbench-bot>
linphone: rebuilt agent-irc from main; testing new auto-recovery. sweep at 25/42.
…
<zkbench-bot>
hi
…
<zkbench-bot>
linphone: SP1 marginal sweep done. HTML: pr185-sp1-marginal.html (10 opcodes, +20 spec). cycles/op & prove µs/op per opcode in summary table.
…
<zkbench-bot>
linphone: real +90 SP1 stress prove done. add=40s mulmod=32s keccak=34s blake2f=50s ecrecover=34s modexp=14s
…
<linphone>
how is the pace?
…
<zkbench-bot>
linphone: pace is good. SP1 p15-p40 + risc0 p15-p40 sweeps done. SP1 marginal sweep done (10 opcodes, +20 spec). real +90 SP1 just finished — spread 3.69× (vs 4.70× at +40). chart regenerated: pr185-sp1-vs-risc0.html. all systems green.
…
<zkbench-bot>
linphone: SP1 full marginal prove sweep done (alethia-reth p30, post-#185, +30 spec, sample-1). 441 prove JSONs, 0 crashes. Report: pass9-bench-tools/results/sp1-marg-p30-full/reports/zk-gas-report-unknown-proving-20260522-040814.html
…
<zkbench-bot>
linphone: risc0 sweep progress — done=2/429 crashed=0
…
<linphone>
how is it goin?
…
<linphone>
?
…
<zkbench-bot>
linphone: risc0 sweep progress — done=18/429 crashed=0 rate=248s/proof eta=28.3h