Binary Search: A new implementation that is up to 25% faster

Re:Binary Search: A new implementation that is up to 25% faster by asah on Monday July 20, 2020 @05:44AM (#23896258)

Random idea: a replacement function library that accepts hardware tuning parameters like this and a corresponding tool which sets them once during OS installation/upgrade and provides a config file or environment variable setup, so apps can share the setup...

Bonus if this library also detected/leveraged specialized hardware (GPU, AVX, etc) and then leveraged it appropriately.

Obviously, these parameters would need optional per-app-invocation overrides, e.g. so random apps didn't hog shared resources, e.g. for benchmarking, etc.

(I'm guessing this exists but not widely deployed...?)

[ Parent | on HN ]

Re:Binary Search: A new implementation that is up to 25% faster by jhayward on Monday July 20, 2020 @11:01AM (#23898481)

This was somewhat the approach used in the ATLAS[1] linear algebra library. In the ATLAS case, systems were characterized through testing and a set of parameters generated. Then at installation time code was generated to match the system type using the given parameters.

[1] https://en.wikipedia.org/wiki/Automatically_Tuned_Linear_Alg...

[ Parent | on HN ]