-
During simulation (and before analysis), every included signal/model computes the LISA response functions for that signal independently. However, many of the underlying calculations are the same acros…
-
## Current Situation / Problem you want to solve
The proposal in this issue concerns the functions `criterion_plot`, `profile_plot` and `convergence_plot`.
- The `criterion_plot` uses the number…
-
It would be a nice improvement if lazy values were supported (like `std::cell::LazyCell`).
These variables would evaluate their value only when used, at which point they'd turn into some other `Val…
-
## Use Case
I'm a puppet partner and PE customers often start to evaluate bolt and eventually switch to PE plans. Then they are surprised that the (expensive) "enterprise" solution has less feature…
-
GradBench currently evaluates each function only once on each workload. This makes the runtime measurements highly sensitive to noise, and also unrealistically pessimistic, as caches will be cold. Th…
-
When using kriging, the most expensive task is the calibration of the parameters of the covariance function, done by log-likelihood maximization. When the optimization is done using a gradient-based a…
-
Noted by pylinter that using `"%"` formatting is more efficient than using f-strings (related to evaluation of variables delayed with `%` formatting, so especially in the case of costly functions, thi…
-
I've compiled a set of optimizations that are appropriate for the gnovm design and architecture.
Some of these need to be done in a particular order.
For example, Dead Code Elimination needs to be don…
-
Consider the following example.
The default function argument `(arg = expensive())` is evaluated for each function call, making the apparently trivial function calls take a lot of time.
In this exampl…
-
### Problem Statement
I have a query where on a limited number of cases an expensive lookup needs to take place.
At the moment `CASE`, `COALESCE`, and `IF` all result on the lookup taking place for …