Upload time #20
Comments
|
Have you tried the RC on test.pypi.org? |
|
thanks, I don't know Python, so if I run pip install pymcuprog, does that pick the latest version or do I have to download it? this is what I get if I run pip install pymcuprog from a terminal in platformio: |
|
Ah, sorry. |
|
thanks for your help, if i use that version, the upload fails: |
|
oh - how does it fail? |
|
just get this: |
|
after programming the chip again using pyupdi, pymcuprog from the command line seems to work again - 30 seconds to upload the 17k.. |
|
Could you add in a -v info or -v debug to add logging verbosity and see if there are more breadcrumbs as to why this is failing? |
|
ok, here is the output with -v info: |
|
Well, that one looks like it succeeds - I was thinking about the failure case... |
|
I'll keep the -v in the command line and if I get a failure again, I'll post. With this test version of pymcuprog, the upload time for the 17kB takes about 33 seconds. |
|
OK, do that. |
|
Just to ask if there's any news on improving the upload speed? I am using the circuit as per diagram for upload on a 4809 chip. |
|
Hi @paulskirk53 - the other issue (#10) you posted about upload time was resolved as fixed in the 3.13 release in May. The main culprit there was the ACK turnaround time - so ACK is now disabled. I was not sure if this was a duplicate report, and left it open as it mentions a crash which appears to be non-reproducable. Maybe we need to re-sync:
|
|
Thanks, I just did an install and now find that on 11.5k of code, pymcuprog takes 22 seconds and pyupdi takes 32 seconds. So there is an improvement. I am not an electronics or programming professional and don't claim to understand much of the detail of this stuff. Perhaps UPDI is inherently slower than methods such as AVRISPs which I've used on other AVRs which are very fast. Thanks for providing pymcuprog, I appreciate it. |
|
I think there is still a fair amount of overhead with the lengthy turnarounds via python - if speed becomes a concern, its always possible to pick up any one of the Curiosity Nano kits, saw off the target-end, and use it as a programmer. From pymcuprog you just remove the -t and -u switches and it should just work... |
|
thanks very much for the info, I'll see how it goes and I might try a sawn off nano :) |
|
Not sure if this helps, but are you using a USB serial adapter? I always wondered how flashing a device is so slow for the tiny amount of data and only recently found a note that said that FTDI adapters are preconfigured badly. You could set the latency to a lower value and it really did speed up things for me by a factor of 10 or so. Or use another USB serial adapter type. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi Is there any way to make the upload faster? My 17kB code takes 33 seconds to load. My next project has significantly more code...
many thanks for providing this, I hope you can speed it up.
The text was updated successfully, but these errors were encountered: