You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The eDSL currently does a quick and dirty job when it comes to name generation.
It maintains a counter starting at 0, and every time it needs to makes a new name for a cell, it does something like {the_op_in_question}_{counter++}. This is why end up with things like reg_0, add_1, mult_2, etc.
The true solution is something akin to what Calyx itself does:
@rachitnigam, I'm curious why name_hash isn't used additionally to store whether the name has been defined, this task delegated to a separate HashSet? It doesn't seem hard to remove generated_names and keep the same functionality. That being said, it's not important.
The eDSL currently does a quick and dirty job when it comes to name generation.
It maintains a counter starting at 0, and every time it needs to makes a new name for a cell, it does something like
{the_op_in_question}_{counter++}
. This is why end up with things likereg_0
,add_1
,mult_2
, etc.The true solution is something akin to what Calyx itself does:
calyx/calyx-utils/src/namegenerator.rs
Lines 7 to 10 in 6f895a1
It actually tracks existing names and avoids clashes. Let's eventually upgrade the eDSL to something like this.
Originally posted by @anshumanmohan in #2152 (comment)
The text was updated successfully, but these errors were encountered: