Use luster as a benchmark (Lua VM)

[zaoqi-unsafe]

No description provided.

[nbp]

At the moment, no. As of today the cranelift backend integration is not yet complete (milestone 0.0.1)

Milestone 0.0.2 is expected to provide interpreter loop unrolling, which would be the beginning any benchmarking effort. At the moment, I am expecting to reach this milestone by the end of 2019.

Looking at run_vm function of luster, I do not see anything which would be too hard to add if missing once the 0.0.2 milestone is reached.

I will rename this issue to be looked at once more for the interpreter loop unrolling, as a potential benchmark reference.

[cztomsik]

can we help somehow?

[nbp]

@cztomsik Thanks for the proposal. I about to release the version 0.0.1, hopefully by the end of the month.

Then I would have to look at a way to replace the current macro written in the library and replace it by a procedural macro. Without procedural macro, this project will have problems to handle elisions for lifetime parameters.

[nbp]

@cztomsik I will try to split up tasks to be done for the version 0.0.2 better than I did so far. Thanks for reminding me.

[cztomsik]

@nbp cool :-)