brrcrites / cs220 Goto Github PK

Dear sir,

Thank you for providing the sample output and the visualization. The "feature" indicates the location of each components and the connection between them. Still there are couple of question here:

(1) Since the mixer in visualize.json contains two port for inputing, we can route one for each input and build this easily. But when it comes to columba.json, there exists many-to-one connection.
Should we create a pseudo bar component as a source/sink item, and use this pseudo component to describe connection in "Feature"?
Or we can ignore the bar and allow such many-to-one in "Feature"?

When I check the paper, the switch component seems to fit our need (as a midpoint to route many-to-mant flow). Should we use it?

(2) I think the all flow channel should be at horizon direction?

(3) According to our previous mail, we should set parameter d as 50. But the sample doesn't look like that way. Also, do we have any constraint to keep component (or channel) at least certain distance from boundary?
Just want to know why setting the input_1 at (5,5). The objective function can obviously be improved by shifting(-5,-5) for every item

Sorry for giving so many question, thank you for any response.

my question is what happen when there are 3 or more parallel module [y—stack], how does the middle module get its control layer.

For example
{control_}
[module_]
[module_] <——————— how does this get it control layer
[module_]
{control_}

can we assume the control bbox that are already present will route up/down to all module in the y,-y direction

OR
should it be like this
For example
{control_}
[module_]
<-------flow line---[module_] <——————— how does this get it control layer
[module_] | |
{control_} | |
{control_}