More in

Big brother, little brother

1 min read

Is the relationship between dsps and fpgas about to change? By Philip Ling

If you take a step back and think about it, electronics can be broadly segmented into three sections – memory, analogue and digital. Within these three discrete sections there are, naturally, subsections, which unlike their parent groups experience a degree of crossover. The level of technology crossover between these sections has largely been defined and regulated by how comfortably the application area they best fit sits alongside neighbouring application areas. For instance, a high end microprocessor wouldn’t normally be found in an Ethernet modem and it’s unlikely that sram would be used as main memory in a digital camera. The reason for this is that, while it’s technically possible, commercially it wouldn’t make sense to use a powerful but expensive microprocessor as a modem, or fast but volatile memory as long term storage. It does make commercial sense, though, to use a microcontroller and flash memory respectively in these two examples. This self regulating harmony exists across the three main segments because of these application boundaries and commercial pressures, but what happens when an exception proves the rule; when the application boundaries for two technologies occupy almost the same space and the commercial pressure between them is in constant ebb and flow?