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
A simple question: did you checked a performance and memory overhead produced by this library (i.e. in comparison with other libs like Sigil? I'd like to evaluate it in situation when code will be generated at runtime, potentially per user request. In this case memory usage and IL generation speed matters.
The text was updated successfully, but these errors were encountered:
There's definitely some overhead because the computation expression results in a lot of little closures being allocated. They're very short lived so should all be gen 0 collections, but the big downside is that the first time a given block of IL-generation code runs it has a lot (milliseconds) of overhead, which goes away on subsequent iterations (to generate additional types/methods using the same pattern of instructions). I'm guessing this one-time overhead is the JIT compiling each little closure in the cil block.
I haven't checked against Sigil (or anything else) but I would bet it's measurably faster than LicenseToCIL.
A simple question: did you checked a performance and memory overhead produced by this library (i.e. in comparison with other libs like Sigil? I'd like to evaluate it in situation when code will be generated at runtime, potentially per user request. In this case memory usage and IL generation speed matters.
The text was updated successfully, but these errors were encountered: