we are doing a lot of stats stuff
so the random package and stats package will be our first focus
…from the look of things, we seem to have declarations for most all of
@stdlib/random and for most of @stdlib/stats/*.
was having troubles
Hey @alvinsunyixiao ! Thanks for reaching out. At the moment, no…although this is a current WIP. Much of that work is happening here: https://github.com/stdlib-js/stdlib/tree/develop/lib/node_modules/%40stdlib/blas/ext/base, but we’ve yet to fully roll out our lower-level strided array APIs. Beyond basic arithmetic, are you looking to do anything else?
@kgryte Thank you for the pointer! I am trying to implement an interactive simulation based on vectorized differential equations, so basically looking for
numpy equivalent in javascript. I have tried Tensorflow.js. So far tfjs is great in terms of API but still several times slower than numpy native C++ bindings. Is there any benchmark carrying out comparisons between stdlib-js and numpy?
@alvinsunyixiao Yes, indeed there are benchmarks! For example, https://github.com/stdlib-js/stdlib/blob/b4d11aac1aa74441653f0ba195a184fd5cc51de0/lib/node_modules/%40stdlib/blas/ext/base/dcusum/benchmark/python/numpy/benchmark.py
Unfortuntely, atm, these benchmarks are scattered across the code base, being localized with their respective package, and need to be run locally. We’ve yet to add tooling for aggregating all the benchmarks across the project.
If there are particular features you need, we can work to prioritize those. Certainly some of the lower level building blocks.
@alvinsunyixiao There are a few APIs for addition/subtraction over strided arrays. These are lower level APIs, but can easily be used by ndarray users. For example, here is a BLAS function for the operation
a*x + y element-wise: https://github.com/stdlib-js/stdlib/tree/8542ec63ba38530cdbe61d8ec86a09ff370c95e7/lib/node_modules/%40stdlib/blas/base/daxpy
To manually call the lower-level strided API when using ndarrays, see https://github.com/stdlib-js/stdlib/blob/8542ec63ba38530cdbe61d8ec86a09ff370c95e7/lib/node_modules/%40stdlib/blas/dswap/lib/main.js#L64 for an example.
We’re still in the process of writing the
ndarray API wrappers, as done in the second link. If these would be useful, we can easily prioritize them this week.
Note that, if you are performing these simulations in Node.js, you can leverage hardware optimizations by instructing
stdlib to compile against a particular BLAS library (e.g., OpenBLAS). This allows for an order of magnitude speed-up against reference C implementations.
any help would be appreciated
thanks
Except for stdlib I was not able to add any other submodules using require
If it is public, would you be able to provide a link to it?
If you mean being able to
require stdlib packages (e.g., @stdlib/math/base/special/erf), what version of @stdlib/stdlib are you using?
@SindujaRajadurai Ah! I know what you did. You’ve installed the wrong package. You did
npm install stdlibrather than
npm install @stdlib/stdlibThis project is published as a scoped package.
let me reinstall it
No problem! Here are the installation instructions for the project: https://www.npmjs.com/package/@stdlib/stdlib#installation
@kgryte Thank you very much for the suggestions! I will give it a try with those low level strided APIs. I am not planning to use a backend, since we already have well written performant simulator in Python, we could have just go for a Python backend. But we are trying to deliver real-time interactive experience, so we decided to bring everything to the browser.
@kgryte Hi! Apologies for going AWOL earlier, but I would love to continue contributing to stdlib. I was thinking of taking up #322 as a refresher to the codebase. Is that okay?
You should be able to use
@stdlib/math/base/special/fibonacci as a template for implementing #322 . Let me know if you encounter any issues!