Adding to this, the 2000 entries in the display list is not the only limitation I am wondering about.
I did some profiling and found that VERTEX2F is a huge bottleneck because of the odd positions the X corrdinate has to be shifted in.
This is a major pain when working with calculated coordinates.
Especially with 8 Bit controllers which kind of are the designated target for the EVE series.
But this also has to have some impact on 32 bit controllers.
I would apreciate a command that at the cost of one bit resolution has the X coordinate not shifted.
Or heck, make that two bits shorter and 1 pixel resolution by default, that would be a range of +/- 4096 to set.
I also would not mind patching VERTEX2F for a faster operation of the FT81x but FTDI/Bridgetek would need to supply such a patch as the inner workings of the FT8xx are only partly explained.
What also is very slow is PNG support in FT81x.
I clocked a simple test picture with 3867 bytes to take 53ms for processing after sending the data.
While a JPG with 3903 only needed 480µs to process.
That means JPG is 110 faster than PNG.
Maybe this is the reason why PNG is missing from the table at
http://brtchip.com/bt81x/ ?
Also I am wondering why the BT815/816 datasheet has not been updated so far with a release due next month.
The draft version really has nothing new so far, at least not in any detail.
Even more I am wondering why the FT81x programmers guide is not getting any updates despite the numerous bugs it contains and of which I reported some over a year ago.
And am I the only one wondering why the next device is named BT816 instead of BT814? :-)