Eureka! NXP battle, an update
-
I still don't know how to boot this NXP evaluation board off of SD but I can now happily use the generous SDRAM, plus any of the SRAM banks I like for my program code and/or data. No more issues like earlier. I don't even need the SDRAM anymore because I separated everything out into the SRAM partitions. Organization! woo I'm working on getting the LVGL benchmark to run 5 separate instances on 5 separate screens. It's not easy because it's C, and so everything is static, global, and hopelessly procedural. Porting that code to exist in multiple instances is a bit of a bear, particularly since I don't understand it yet. Oh well. I'll figure it out. I contribute to LVGL, how hard can it be? BUT BUT BUT I was getting blazing framerates this time around. Last time I was happy to get 16-19. Now I'm getting 45. Sometimes a lot more. (sometimes less if I bog the CPU), but I haven't tried across all 5 displays. I imagine I'll lose plenty of frames because the CPU is going to be tied up, but if I can get the I/O to be async like before it should perform well enough for what we need. I'm convinced that this mix had something to do with my successes. clippy vs the black hole - [Atmospheric Drum and Bass Mix] - YouTube[^]
Check out my IoT graphics library here: https://honeythecodewitch.com/gfx And my IoT UI/User Experience library here: https://honeythecodewitch.com/uix