feat(bench): benchmark MockChain transactions

[djolertrk]

It depends on the 0xMiden/miden-debug#127

[github-actions]

Miden examples benchmark

Candidate 840006ab0cfd compared with next 62c4318f44f2. Lower is better.

example	VM cycles (vs next)	MAST size (vs next)
auth-component-no-auth	7,459 (~0%)	6,300B (~0%)
auth-component-rpo-falcon512	80,179 (~0%)	12,654B (~0%)
basic-wallet	81,696 (~0%)	7,982B (~0%)
basic-wallet-tx-script	83,356 (~0%)	13,556B (~0%)
collatz	5,263 (~0%)	2,401B (~0%)
counter-contract	85,387 (~0%)	13,099B (~0%)
counter-note	85,387 (~0%)	3,197B (~0%)
fibonacci	869 (~0%)	3,219B (~0%)
is-prime	2,332,731 (~0%)	5,864B (~0%)
p2id-note	81,696 (~0%)	17,024B (~0%)
p2id-tx-script	85,331 (~0%)	11,455B (~0%)
p2ide-note	82,099 (~0%)	13,103B (~0%)
storage-example	82,688 (~0%)	15,249B (~0%)

SVG flamegraphs, replay snapshots, and compiled packages are attached to the workflow run.

[djolertrk]

hey @greenhat @bitwalker. how does this look to you? :)

[greenhat]

Nice!
Please make a task in the Makefile.toml to run the benchmark suitable for agent consumption (stdout only, no report file, no github comments, etc.).
As a side note, could you please skip posting github comment if there are no changes in size or cycles?

[greenhat]

I'd like to leave the storage example as is since it is intended to demonstrate the "escape hatch" of providing your own WIT file for power users.

[djolertrk]

As a side note, could you please skip posting github comment if there are no changes in size or cycles?

Yes, it does make sense

[djolertrk]

I added:

• cargo make bench-agent (@greenhat is it a good name?)
• CI now skips PR comments when no changes
• addressed the rest

[djolertrk]

The 0xMiden/miden-debug#127 is merged and the new debugger is released, but we need to wait for compiler to be ported to the new vm v0.30.

[greenhat]

I added:

• cargo make bench-agent (@greenhat is it a good name?)

Yep, it's fine.

• CI now skips PR comments when no changes
• addressed the rest

Great! Thank you!