LinearSolve.jl: Because A\b is Not Good Enough | Chris Rackauckas | JuliaCon 2022

[@franksinatra2530]

Chris is an absolute legend. Very good talk.

[@frederik3982]

Love the energy! Thanks for the great talk

[@vahidhosseinzadeh4630]

very fun and useful presentation, thanks

[@abstractnonsense8344]

Thanks. Great as always.

[@kamilziemian995]

Great and funny presentation. I only wish that I would be a little bit slower. Maybe it was written in Julia and that the reason why is so fast?

[@alvaro_gavilan_rojas]

From someone who use inv(A)*b, this is shocking. Great talk, thank you!

[@BariumIon]

Thank you, great work!

[@Caspar__]

Thank you for the talk : )

[@AJ-et3vf]

Awesome video! Thank you!

[@jamesthesnake12]

way to go chris

[@user-wr4yl7tx3w]

this was really well and concisely explained.

[@kamilziemian995]

I wonder what people responsible for linear algebra in Julia standard library would think about LinearSolve.jl?

[@magno5157]

Ask them. Are you pleased?

[@kamilziemian995]

@@magno5157 Maybe I will ask them. But, will they find time to answer?

[@micahchurch5733]

The \ is just an operator overload right? Or something similar perhaps?

[@hexane360]

Similar. In Julia '\' is just a function (that can be called like '\(a, b)' or 'a \ b'), and in Julia all functions use 'multiple dispatch'. This means that the method used is based on the type of all arguments passed, not just the first (as with operator overloading in most languages). In Julia 1.8.0, with just Base loaded, 146 methods (like 'overloads' in other languages) are defined on the function '\'. These include things like specialized algorithms for sparse matrices, algorithms for special matrices and operations on other types like bit vectors.

[@BariumIon]

Thank you, great work!