i have fixed my xoom and updated to latest beta.

i have fixed my xoom and updated to latest beta. im about to start a 5 hours print. One thing i noticed and that might be a good feature is to be able to print from the controller sd.

will report back when the print is done, im loadling it to gdrive :smiley:

ok, make sure to enable the status bar in the options, because the non-statusbar temperature label has a memory leak and can cause garbage collector runs which lead to stuttering.
SD Card printing is on the list for future feature, but my printer does not have a sd card yet (for testing).

cool, gdrive is not working fine on the tablet, it dies. Its a 30mb gcode file, so, i enabled the 30mb bumps heap memory thing and its loading quite fast!

30Mb gcode is huge. The largest I test so far was ~15MB. I’m curious if it works.

Yeah, is insane, so its loading, at 28% it slowed down, i guess it will take like 20 mins to load the file

I’m concerned that it won’t work if it already slows down at 28%.
Maybe I need to allow bumping up more
heap memory than just 30MB.

39% and loading :slight_smile:

44.9%

49

51, looks its getting slower.

54

oh man. much too slow. I need to test with 30MB files as well …

63%, i will send you the gdrive link of this file

sounds good. thanks

wow. 1100057 lines in ~500 layers.

I did some optimizations and now it loads the 30MB file in ~250 seconds. Heap is growing up to 145MB :frowning:
I will upload a new beta version with the fix later today.

cool, im printing now (no from the xoom), but if i get the new beta today, i will leave it printing tonight. Its a 4-5 hours print.

I have uploaded the new version 1.34 to Google Play. As always it can take some hours until you see the update.

Fix description: In addition to the 30MB allocation to bump up the heap size, I do 5MB allocations for every 40000 gcode lines. This prevents android from doing thousands of tiny allocations which is slow.

@Nicolas_Arias did you try the new version v1.37 ? My Galaxy tab loads your large gcode file in ~300 seconds. (Still some room for improvement)