Catherine
Catherine
Right, that sounds fine to me. The details seem a little complicated but it seems like the problem we have on our hands is a little complicated too, so that's...
Have you had a chance to revisit this? In the meantime I completely redesigned the demultiplexer (it's now part of `HardwareAssembly`) and raised buffer limits. I'm not sure if the...
> As an aside, I'm excited by the changes to the interface. There's stuff I've been interested in doing that will be more straightforward now. I'm also looking forward to...
@Sazzach Have you been able to give this PR another look?
I'll take a look shortly, but to answer this: > One question I have is how does the `add_inout_pipe` argument `out_buffer_size` work? I tried setting it to limit the buffer...
@Sazzach I'll give it another look once I'm more awake, but my main comment so far is that the code is quite hard to follow. There are a lot of...
Some examples of what I think is difficult to review and maintain: 1. Right now, the existing stimuli consist of a 3-stream chain: source (pipe or LFSR), optional rate limiter,...
> I'm down to: > > * 1. make a PR for adding an interface class to the benchmark applet > * 2. in a later PR add in modes...
The overall structure looks good at a quick glance, and the functionality as a whole is very desirable. I assume you'll address a variety of minor issues before you un-draft...
The implementation of the protocol seems like it would suffer in performance due to the heavy use of sub-coroutines to read individual fields rather than packing/unpacking them in bulk, but...