Cache in spike prime

[techcatgato]

Does spike prime have any cache that might be able to be utilized by my PyBricks programs?

[laurensvalk]

You can make arrays or matrices to keep data in memory.

If you need a small amount of persistent data, you can use the hub.system.storage method.

[dlech]

When I measured it, I got: my_func took 0.1782 ms, so it seems already quite fast. We could probably knock off a few more microseconds by avoiding the exponents, but I don't think that would make much of difference in your program if you are only calling this 10 or 100 times per second.

[techcatgato]

When I measured it, I got: my_func took 0.1782 ms, so it seems already quite fast. We could probably knock off a few more microseconds by avoiding the exponents, but I don't think that would make much of difference in your program if you are only calling this 10 or 100 times per second.

When I send the output to a motor Is it still going to be fast? Dont have a motor on me right now so I cant check

[techcatgato]

When I measured it, I got: my_func took 0.1782 ms, so it seems already quite fast. We could probably knock off a few more microseconds by avoiding the exponents, but I don't think that would make much of difference in your program if you are only calling this 10 or 100 times per second.

Did you run yours on your PC or on the SPIKE PRIME?

[dlech]

On the SPIKE Prime.

[laurensvalk]

When I measured it, I got: my_func took 0.1782 ms, so it seems already quite fast. We could probably knock off a few more microseconds by avoiding the exponents, but I don't think that would make much of difference in your program if you are only calling this 10 or 100 times per second.

When I send the output to a motor Is it still going to be fast? Dont have a motor on me right now so I cant check

Sending it to the motor is almost instant. Especially if you control acceleration yourself using the dc() method.

But it is only going to respond as fast as the mechanical components in the motor can move. Usually with this type of system, any computation under 10 ms isn't going to slow things down all that much.

For reference, a self balancing robot needs to respond pretty quickly to stay balanced but it still works great with a 20ms or even a 50 ms loop time.

More formally speaking, a LEGO motor (or any similarly sized electro mechanical system) acts as a low pass filter. Even if you throw very fast changing data at it, it only cares about the average value over time.