Sending big data arrays through SPI non-blocking... DMA?

Hello guys,
i’m completely new to EMX, wanna use it for my bachelor thesis - lacking time to trial & error…

I have a many-channel DAC to control via SPI, and wanna send it big data packets every 100us as precise as possible. The data packets are in memory and only need to be forwarded to SPI, but it shouldn’t block the cpu for too long as it has other work to do.

Is there a way, to setup DMA for SSP in RLP, so i create a Task, call the procedure, it forwards the buffer pointer to SSP and exits leaving it to send data while NETMF runs? Or is it even possible to do that in managed code keeping precision?

Example or link with info would be great,
thank you

EDIT:
Especially, if I use SD card from NETMF (which I will), I guess it uses DMA to. How do I not interfere with it, setting the registers?