bench · bez · Tangled

burrito.space / bez

This repository has no description

bez bench

6 folders · 56 files

at track/scripted-e2e

dietrich ayala · Models given at submit time: JOBS_MODEL and JOBS_BATCH_MODELS · 3w ago

Models given at submit time: JOBS_MODEL and JOBS_BATCH_MODELS

Post once for every landed feature job, map move or not

Model variety for generation runs: qwen3.8-27b-fast-2 and deepseek-v4.1-flash

Model variety for generation runs: qwen3.8-27b-fast-2 and deepseek-v4.1-flash

Add rig-capacity benchmark scripts and crate

Read NEV_* variables with a GENERA_* fallback in the bench scripts

Build lane: distributed compilation and a shared cache across the EPYC nodes

absolute-mixed-batch.sh

Run model batches only on a lane; probe candidates in one wide build

Pass the landed job's result to the announcement after every landing

Announce landings that move a web feature forward, with a per-track ledger

Measure rig capacity: CPU-side benchmarks, GPU-tier formulas, capacity model

Add rig-capacity benchmark scripts and crate

Add rig-capacity benchmark scripts and crate

compile-batch.sh

Measure rig capacity: CPU-side benchmarks, GPU-tier formulas, capacity model

Run model batches only on a lane; probe candidates in one wide build

Run model batches only on a lane; probe candidates in one wide build

job-run --inventory: check the lane's snapshot before the first stage, exit 3 on a refusal

dashboard-update-test.sh

Landing refreshes the coverage dashboard

dashboard-update.sh

Landing refreshes the coverage dashboard

expected-features-check.py

Say after a landing when a job declared features that did not move

feature-announce-test.sh

Post once for every landed feature job, map move or not

feature-announce.py

Post once for every landed feature job, map move or not

folknet_snapshot.py

Merge branch 'main' into track/dagu-trial

folknet-probe-install.sh

Merge branch 'main' into track/dagu-trial

folknet-probe.sh

Merge branch 'main' into track/dagu-trial

folknet-status.sh

Merge branch 'main' into track/dagu-trial

Kimi as a declared provider: alias provider in folknet.json, pi compat from configuration

fragment-floor.sh

Add bench/land.sh so a tick can land approved spec tracks

gen_candidates.py

Measure rig capacity: CPU-side benchmarks, GPU-tier formulas, capacity model

generated-floors-update.sh

Gate: floors are files next to each generated function and rise by themselves

generated-scores-test.sh

Tests: a zero raise for a function with no placeholder is refused

generated-scores.sh

Gate: floors are files next to each generated function and rise by themselves

job-events-lane-test.sh

bench/job-events-lane.sh: run the job event store on a lane, with offline scenario tests

job-events-lane.sh

bench/job-events-lane.sh: run the job event store on a lane, with offline scenario tests

Pass the landed job's result to the announcement after every landing

Pass the landed job's result to the announcement after every landing

lane-apply-test.sh

Merge branch 'main' into track/dagu-trial

Merge branch 'main' into track/dagu-trial

lane-create-lxc.sh

Build lane: distributed compilation and a shared cache across the EPYC nodes

Build lane: distributed compilation and a shared cache across the EPYC nodes

Dagu job: make the share's ledger directory on first use

lane-provision.sh

jobs: untested drafts of the pi scripts, submit and land, and job-check result

lane-sweep-test.sh

Build lanes remove copies of tracks that no longer exist

lane-sweep: a single-lane sweep does not claim a place in the list; handoff current

lane-writer-fetch.sh

Roster: lane-writer launch and fetch scripts, config gains a lanes[] array

lane-writer-launch-test.sh

Lane writers start through the harness-neutral writer command

lane-writer-launch.sh

Merge branch 'main' into track/dagu-trial

lane.example.json

Roster: lane-writer launch and fetch scripts, config gains a lanes[] array

Announce landings that move a web feature forward, with a per-track ledger

on-lane-test.sh

on-lane.sh and remote-lane.sh run in place when NEV_ON_LANE is set

on-lane.sh and remote-lane.sh run in place when NEV_ON_LANE is set

oracle-install.sh

Dagu job steps use the NAS npm cache; oracle install prefers it offline

oracle-scaling.sh

Measure rig capacity: CPU-side benchmarks, GPU-tier formulas, capacity model

overflow-mixed-run.sh

Run model batches only on a lane; probe candidates in one wide build

probe-replay-measure.sh

Probe builds without a compiler wrapper by default

on-lane.sh and remote-lane.sh run in place when NEV_ON_LANE is set

rig_throughput.py

Read NEV_* variables with a GENERA_* fallback in the bench scripts

rig-inventory.example.json

Announce landings that move a web feature forward, with a per-track ledger

rig-snapshot.sh

Model variety for generation runs: qwen3.8-27b-fast-2 and deepseek-v4.1-flash

run-inner-loop.sh

Add rig-capacity benchmark scripts and crate

bench/verify.sh: run the job event store tests in the track gate

web-features-update.sh

Add the web-features map generator, diff tool and landing step