Read dram will automatically request the data in increments. So if you want the first 100MB, for example, then just say fpga.read_dram(100*1024*1024). Even with the block based reading, it takes about 7minutes to read the entire 1GB DIMM. This is a blocking call, so you'll get nothing back for a while until everything's been retrieved.
If you're interested in reading a bit of data somewhere in the middle of the DRAM, you can use the offset. For example, if you want the one-hundred-and-first MB (ie one megabyte at offset 100MB), fpga.read_dram(1*1024*1024,offset=100*1024*1024) Jason On 15 Aug 2011, at 20:01, Soriano, Melissa (335J) wrote: > Hi, > > I installed corr 0.6.7. > > How can I use katcp_wrapper.read_dram to read several megabytes of data > from the DRAM? I am currently using read_dram to read 1 MB at a time, > like this: > data=fpga.read_dram(megabyte) > > How should I use the offset argument? > > Thanks, > Melissa > > On 8/9/11 10:07 AM, "Jason Manley" <[email protected]> wrote: > >> Corr is a separate package. You can find it here: >> http://pypi.python.org/pypi/corr >> >> Corr depends on the Python katcp package and adds (amongst other things) >> functionality to talk to ROACH boards using the KATCP protocol, which is >> a generic. If you're not building packetised correlators, then you >> probably only want the katcp_wrapper.py part of corr. >> >> Jason >
