-
Notifications
You must be signed in to change notification settings - Fork 198
Open
Labels
performancePerformance-related improvements or regressionsPerformance-related improvements or regressionsresearchOpen questions requiring either research or requirements gatheringOpen questions requiring either research or requirements gathering
Description
When we want to run a CASM program, there are a couple of steps that need to be done always, regardless of how the program will be executed. For example:
- Loading instructions.
- Decoding instructions.
- Compiling hints.
This steps take time. Currently, there is no way of caching these values between runs. That is, they have to be computed every time I want to run the same program.
-
The first step is to find which values can be cached between runs. There may be some other values not mentioned here.
-
The next step is to measure/estimate how much would caching this value improve performance. We should note how much keeping this value cached costs.
-
Finally, we need to propose a new API that allows the user for retrieving/specifying this precomputed values.
Metadata
Metadata
Assignees
Labels
performancePerformance-related improvements or regressionsPerformance-related improvements or regressionsresearchOpen questions requiring either research or requirements gatheringOpen questions requiring either research or requirements gathering