Posted on January 13, 2015 at 20:32I can't seem to find more information on this, even looking through the EVAL code...
here it goes:
I am designing a simple system to learn how to use the STM32F439XX and perhaps use it in future projects, but I am not sure if it is fast enough since the AHB can only support concurrent accesses to 1 external memory.
I am hoping to attach 128k of memory as a frame buffer at the beginning of this section of memory. I already know there isn't enough time to render and fill (think a simple game like super mario world or something along those lines) the frame.
That said, I can attach 256k and double buffer by swapping the location of the buffer between frames, that gives me more than enough time to render each frame.
The problem occurs when I realized that I can't concurrently read from external memory for the frame buffer on FMC mapped SRAM/DRAM at the same time as writing to it. Does anyone know if the contention will be a big problem? I have never done something quite like this before, so I'm not sure on the details of sharing a bus (I know it's a round robin algorithm).
The pixel clock is 9 Mhz and system clock is whatever the minimum it needs to be to achieve ~60 fps (for debugging puruposes at the moment let's just say the maximum 180 Mhz). I think the external memory will run at ~50 Mhz depending on which I buy.
Any suggestions are greatly appreciated.
Brandon
#lcd #tft #discovery #stm32f4 #stm32
