Stefan Krastanov
Stefan Krastanov
Julia is not fast when working with containers over abstract types. In particular, the heap and other fields in the `Simulation` type are currently abstractly typed or have abstract `eltype`s....
There is an old PR that can help with that https://github.com/JuliaDynamics/ConcurrentSim.jl/pull/71
There are a few dependents on this library that probably should be tested to avoid accidentally breaking them. Some of them are private research projects of @BenLauwens , some of...
In #212 we introduced the `TableDecoder` among other decoders. This is a lookup-table decoder for ECC. Currently, it supports tables of up-to-one error. It would be valuable to have a...
The results of `pfmeasurements` and more generally the way we store measurement results in the Pauli frame object is not local. We should not change the order of the indices...
In #212 we introduced compactification for `ClassicalXOR` which seems to cause more allocations (see the benchmark comments in that issue). Fix those and add explicit tests confirming we have low...
`evaluate_decoder` should have an option to run at least the Pauli frame simulations on a GPU
In #212 we provided a workaround for concrete `ClassicalXOR` types. While the workaround is a bit ugly, it makes it easy to now also provide GPU kernels for `ClassicalXOR` --...
`evaluate_decoder` might evaluate a couple of times during setup, but there should be no new allocations happening when we enter the sampling&decodeing loop. This would require removing allocations from: -...