Replies: 3 comments
|
hi @vvb333007 These are exactly the right questions to ask at, so let's say what I know cold vs. what I still need to put a. First, a clarification on the execution model, because it reframes a couple of these. There's one interpreter, single-threaded. Two modes:
as for the other points:
Furthermore, both the P4 and S3 support multi-core processing; I am not currently utilizing this—not because I haven't considered it, but because it is a somewhat delicate task. I am already implementing this for the event-driven mode, which naturally uses all cores to handle multiple tasks in parallel, so both of the aforementioned modes will support it in the future. If you have any doubts, questions or difficulties in managing, please write, I will be happy to answer you, furthermore if you have any suggestions they are absolutely welcome ❤️ |
|
hi @vvb333007 I put together a benchmark to test the aspects that you asked You can find the benchmark here: examples/benchmark After thinking about it for a long time and doing a couple of tests I ask that this is a correct and reliable way to observe this data Since I saw that you work with esp32, if you have any suggestions/corrections/advice, let me know. I'll bring you the report below:
Let me know if you think there are other things that should be monitored |
|
Thank you for elaborate testing and report! |
Uh oh!
There was an error while loading. Please reload this page.
Great job!
Did you measure PSRAM usage? E.g. how much takes one medium-size script, compiled?
How many compiled scripts can be executed simultaneously?
Did you measure memory usage during script execution?
Did you measure the tight loops speed? I.e. setting GPIO in a loop - what maximum gpio switch frequency can be achieved?
All reactions