Quote:
Originally Posted by AmiBoy
Are you planning to workaround the chipset bottleneck somehow?
|
That's a claude question really. Plus I'm not really sure what bottleneck you're referring to.
the TF1200 has a kind of artificial bottleneck of the buffee being on the DIP64 16-bit data bus. the Pistorm32 is using the full bandwidth of the A1200 32-bit bus. Albeit to the FPGA which is then doing its own protocol between it and the Pi CM4 over the GPIO pins, so there is a bottleneck there in a similar but slightly different way to the TF1200/Buffee combo, but the bus transactions are being communicated from the FPGA to the Pi at upwards of 200MHz (that's the speed on current Pistorm/Pi3A combo. I have no idea how fast the CM4 will be running its bus protocol. I doubt even Claude does just yet...