Run Barretenberg benchmarks on a dedicated remote EC2 machine.
npx skills add https://github.com/HabibTorjmen/Blockchain --skill remote-bench
Please help me install this Agent Skill. Skill: remote-bench Source: https://github.com/HabibTorjmen/Blockchain/tree/main/aztec-packages/barretenberg/.claude/skills/remote-bench Command: npx skills add https://github.com/HabibTorjmen/Blockchain --skill remote-bench
Producing reliable benchmark numbers is hard when your execution machine is shared or noisy, causing run-to-run variance that makes comparisons untrustworthy.
Ask the AI to run the remote benchmark for target bb with the command you intend to measure, after confirming BB_SSH_KEY, BB_SSH_INSTANCE, and BB_SSH_CPP_PATH are configured for the remote machine.
High-intent search queries and answers about installing and using this skill.
To run stable Barretenberg benchmarks without noise, execute the targets on a dedicated remote EC2 instance via SSH. This isolates execution to produce reliable, single-run performance measurements for accurate comparisons.
Remote benchmark execution prevents concurrent run corruption by validating SSH environment variables and enforcing a remote lock. This lock mechanism blocks simultaneous benchmark processes from interfering with each other on the shared remote machine.
Yes, remote SSH benchmark execution requires validated environment variables. You must configure BB_SSH_KEY, BB_SSH_INSTANCE, and BB_SSH_CPP_PATH to enable secure binary transfer, remote execution, and result collection on the EC2 machine.
Yes, remote execution supports both native and WASM benchmark targets. The workload-agnostic transport securely transfers the required binary and inputs to the remote machine, retrieves outputs, and applies to remote chonk workloads.
The best way to measure performance comparison for Barretenberg targets is pairing remote execution with other benchmark command skills. This combination defines the actual BB command and inputs while the remote instance ensures noise-free measurements.
Local benchmark runs produce unreliable performance comparison numbers due to shared or noisy execution environments causing run-to-run variance. Moving execution to a dedicated remote EC2 machine isolates the workload and yields trustworthy results.
HabibTorjmenGenerate Perfetto-compatible traces for Chonk prover profiling.
lucas-flatwhiteAutomates web performance audits to detect regressions via page timing metrics.
HabibTorjmenBenchmark Chonk client-IOV proving with pinned IVC inputs for native and WASM runs.
korchasaDiagnose flowai benchmark failures by comparing run artifacts against a failure taxonomy.
