Fix flaky banking stage tests

[jstarry]

Problem

Some banking stage tests try to process some transactions within the first block of a cluster with slow poh. If for some reason the transaction isn't included before poh ticks to the 2nd to last tick in the first block (takes about ~3.2s (8 * 400ms)), the test will never end because the poh loop stops recording transaction mixins.

Summary of Changes

• Make slow poh even slower by increase ticks per slot by a factor of 1024 (blocks will be over 6min long)
• Exit low power poh loop as soon as exit is set

Fixes #

[jstarry]

Here's the previous fix for what seems to be a similar issue: solana-labs#6251

[apfitzge]

lgtm - aside from the ticks function, it was not really clear to me how the tick hack worked.
Longer slots should help if these tests were flaky due to poh reaching near tick-height.

CI passing, so it seems changing the poh service thread's functionality doesn't break anything!

[apfitzge]

@jstarry https://discord.com/channels/428295358100013066/560503042458517505/1346943500884906014 let's add a 2.2 label

[mergify]

Backports to the beta branch are to be avoided unless absolutely necessary for fixing bugs, security issues, and perf regressions. Changes intended for backport should be structured such that a minimum effective diff can be committed separately from any refactoring, plumbing, cleanup, etc that are not strictly necessary to achieve the goal. Any of the latter should go only into master and ride the normal stabilization schedule. Exceptions include CI/metrics changes, CLI improvements and documentation updates on a case by case basis.