How much memory may a HPPL program use on a G2?

[komame]

(09-12-2023 05:10 PM)chromos Wrote:  Since the speed of different algorithms is compared here, another forum thread might be useful if you are not familiar with it.

Yes, I'm familiar with it - I read it a long time ago. Unfortunately, those results are worthless today. These tests were conducted in 2014, so it was firmware 6975 or older. Since then, there have been 15 firmware versions for the HP Prime, and believe me, if you were to conduct these tests on each of them, the results would be different for each. Some would favor lists, while others would favor text strings, and that has changed over time. All those measurements, ceased to be relevant on the day the next firmware version was released, and that was 8 years ago... After that, the results kept changing, and unfortunately, when averaged out, with each new firmware release, they got worse and worse. The current firmware 14730 is the slowest in history, as I mentioned in this thread (a very long text, but in the 1st and 6th post, you can see charts for very similar measurements I performed in the context of different firmware releases):
https://www.hpmuseum.org/forum/thread-20403.html

Perhaps this 'variability' issue only applies to the PPL environment, as Python seems to be more resilient to various firmware versions when it comes to performance changes. However, some differences can still be observed (only four firmware versions with Python have been released so far, but that has allowed me to measure certain things).
I really hope this negative trend will change, and someone will pay attention to performance in the next firmware release.

Best regards,
Piotr Kowalewski