Forum Discussion
Altera_Forum
Honored Contributor
8 years ago1)
a) You can use shared on chip ram if only FPGA have enough resources to accommodate array size. b) PIO is good, i use it alot, but sometimes i make custom PIO-like component, for example when i needed to export 383 bit wide digit. c) You can use Scatter-Getter DMA for transferring stream data to memory. HPS provides DRAM interface that allows FPGA part to write on HPS memory. If you are using linux, you can write a driver that allocates enough memory for your array and then transfer raw physical address of this chunk of memory to FPGA so it can read from and write to it. 2) connect On-Chip ram to whatever bridge you want and make some experiments first. 3,4) probably it is written somewhere in user manual, but ultimately TimeQuest Analyser will tell You maximum speed and it will always depend on what you connect to Your bridges. I have never needed more than 50 - 100MHz.