System memory content editor

In order for the memory content editor to work, it needs to read and write to the memory. This means it needs an address port and data in and data out ports. So you can only use this option with single-port RAMs, which I understand can be limiting. If you're dual-port RAM is run off the same clock(or related clocks), you might be able to change it to use one port and time multiplex your address and data(basically running twice as fast and muxing all the control, address and data). If you have extra memory available, putting signaltap into the design is usually easier, since you don't have this limitation(you're just pumping the data into another memory).

Thanks for the info. I'll add a signal tap as the 2 port memory is across different clock domains.

Note that you could also create two single port rams that are written to concurrently with your dp ram. Since the memory content editor can access these, you can do the same thing. The nice thing is your output will look more like memory contents, where signaltap will show it in a waveform format(unless there's a simple trick I don't know about). For example, if you're memory is 4K deep and your signaltap is, say 8K deep, it is unlikely that you'll capture at a time when all 4K locations have been written to over 8K cycles, so you won't actually see everything in there. Shadowing your ram with single-port in-system content memory editors will resolve that.

Thanks again Rysc. I like the idea of adding a single port ram for debug purposes. As you say, it will look more like memory contents which will be easier (in this application) than looking at waveforms.

hi

i want to check a memory, i checked the read port in this way that i write something in an mif file and read it with memory and it was ok