Philip Howard

At a wild guess - it's been so long since I looked into this - the existing `get_palette` only works after displaying a PNG? Ah, it looks like it uses...

I've retitled this PR to avoid it being confused with the RP2350 MCU support one 😆

I seem to be having trouble with PSRAM when using split heap, a crude test which uses RAM until a garbage collection is forced will crash/hang on that collection when...

@sfe-SparkFro I've got a new commit from @MichaelBell that, rather sensibly, shifts the PSRAM enable config from the header to a CMake variable, so that the CMakeLists can respond to...

No changes to the garbage collection at all other than https://github.com/micropython/micropython/pull/15620/commits/18e7d019a9f40782857fbaa7219886d12fb4976b to avoid the larger RAM area overflowing the type. The sheer size of PSRAM certainly exacerbates GC timings… but...

AIUI the garbage collection is potentially bounded by the highest allocation, if you're using split heap could the abysmal slowness be the point at which an allocation crosses the boundary...

IIRC there are two bits per word allocated as the GC “bitmap” which is used to mark block heads and allocated RAM. So for 8MB PSRAM that is (if my...

> So there was never any problem with "suddenly" dipping into the PSRAM To be fair this was wild and clueless speculation on my part, PSRAM performance seems to be...

I have pushed two new commits addressing enabling split heap by default, and raising the GC stack size.

> Does this PR still need to be in a draft state? @sfe-SparkFro that's a good question. It *works* so I guess the next stage is finding out if it's...