11th October 2018 in London at CodeNode

There are 38 other SkillsCasts available from Haskell eXchange 2018

Please log in to watch this conference skillscast.

731775060 640x360

The talk will introduce the concepts of vectorization and show its applications in the context of Haskell. Abhiroop will begin by demonstrating what are SIMD instructions and where they fit in the paradigm of parallel computing.

He will also talk about basic applications which can be parallelized using vectorization. Followed by that he will switch over to presenting the vectorization APIs that he has added to GHC's native code generator and how library authors can leverage those APIs to parallelize their code. If time permits, the talk will briefly go over some Automatic Vectorization algorithms the author has experimented with.

YOU MAY ALSO LIKE:

Thanks to our sponsors

Vectorization in Haskell

Abhiroop Sarkar

Abhiroop is interested in the intersection of compilers and parallel programming and believes that pure functional programming will eventually be the panacea to the challenges of multicore parallelism.

SkillsCast

Please log in to watch this conference skillscast.

731775060 640x360

The talk will introduce the concepts of vectorization and show its applications in the context of Haskell. Abhiroop will begin by demonstrating what are SIMD instructions and where they fit in the paradigm of parallel computing.

He will also talk about basic applications which can be parallelized using vectorization. Followed by that he will switch over to presenting the vectorization APIs that he has added to GHC's native code generator and how library authors can leverage those APIs to parallelize their code. If time permits, the talk will briefly go over some Automatic Vectorization algorithms the author has experimented with.

YOU MAY ALSO LIKE:

Thanks to our sponsors

About the Speaker

Vectorization in Haskell

Abhiroop Sarkar

Abhiroop is interested in the intersection of compilers and parallel programming and believes that pure functional programming will eventually be the panacea to the challenges of multicore parallelism.

Photos