cache get_torque_params

[cl0cks4fe]

lots of io in get_torque_params (and it's is called twice in get_params for a lot of cars)

could also pass it through
re: #32536

I'll look for further ways to speed up

[github-actions]

Thanks for contributing to openpilot! In order for us to review your PR as quickly as possible, check the following:

• Convert your PR to a draft unless it's ready to review
• Read the contributing docs
• Before marking as "ready for review", ensure:
  • the goal is clearly stated in the description
  • all the tests are passing
  • the change is something we merge
  • include a route or your device' dongle ID if relevant

[sshane]

nice! speeds up the test with 8 procs from ~37 to ~30s for me! I wonder why the unit tests take 2x longer in CI though. I re-ran it

[sshane]

Can you see why the unit tests take much longer now?

[cl0cks4fe]

@sshane selfdrive unittests?
Looks like it's ~1m faster than this run.

this PR: 1491 passed, 364 skipped, 4010 subtests passed in 420.02s (0:07:00)
that one: 1491 passed, 364 skipped, 4015 subtests passed in 477.50s (0:07:57)

[cl0cks4fe]

@sshane
different runners, which explains why your PRs are also fast. should this change?

[deanlee]

#32586 is a further optimization.

@sshane: commaai/opendbc#942 improves the initialization performance of CanParser & CanPacker, which should boost test_car_interface's runtime speed.

update:
The test_car_interface retrieves the car interface from the cache rather than reinitializing it each time:
CarInterface, CarController, CarState = interfaces[car_name]

Therefore, commaai/opendbc#942 does not enhance the performance of test_car_interface.