4 ms·

Speaking of python performance, I recently benchmarked "numpy vs js" matrix multiplication performance, and was surprised to find js significantly outperforming

Speaking of python performance, I recently benchmarked "numpy vs js" matrix multiplication performance, and was surprised to find js significantly outperforming numpy. For multiplying two 512x512 matrices: python numpy: ~3.30ms numpy with numba: ~2.90ms node tfjs: ~1.00ms gpu.js: ~4.00ms ndarray: ~118.00ms vanilla loop: ~138.00ms mathjs: ~1876.00ms browser tfjs webgpu: ~.16ms tfjs webgl: ~.76ms tfjs wasm: ~2.51ms gpu.js: ~6.00ms tfjs cpu: ~244.65ms mathjs: ~3469.00ms c accelerate.h: ~.06ms Source here: https://github.com/raphaelrk/matrix-mul-test https://github.com/raphaelrk/matrix-mul-test

netheril96 4y ago
Numpy runs on CPU. Tfjs runs on GPU. Not a fair comparison.
- raphaelrk 4y ago
  You're right, it's not a fair comparison -- I think it's still interesting though, since numpy is the standard people would reach for, which made me think it would be the fastest / use the GPU. I expect a python library that uses the GPU would be just as fast as the others.
  - qeternity 4y ago
    People don’t reach for numpy because it’s the absolute fastest…
    - raphaelrk 4y ago
      What I meant was I expected numpy to be faster than the js libraries I was testing, simply because people use it so much more, for real "scientific computing" work. And indeed it is very fast given it only uses the CPU, but that still leaves its matrix multiplication as ~100x slower than what my mac is capable of.
  - psyklic 4y ago
    It adds a great deal of complexity (and often user frustration) for a Python package to support both CPU and GPGPU.
  - dhasso 4y ago
    For that, you can use cupy[0], PyTorch[1] or Tensorflow[2]. They all mimic the numpy's API with the possibility to use your GPU. [0] https://cupy.dev/ https://cupy.dev/ [1] https://pytorch.org/ https://pytorch.org/ [2] https://www.tensorflow.org/ https://www.tensorflow.org/
sampo 4y ago
TensorFlow.js matrices are immutable, which puts more restrictions on your programming style that standard Numpy. You cat get immutable, GPU-enhanced matrices for Python, too.

netheril96 4y ago

Numpy runs on CPU. Tfjs runs on GPU. Not a fair comparison.

raphaelrk 4y ago
You're right, it's not a fair comparison -- I think it's still interesting though, since numpy is the standard people would reach for, which made me think it would be the fastest / use the GPU. I expect a python library that uses the GPU would be just as fast as the others.
- qeternity 4y ago
  People don’t reach for numpy because it’s the absolute fastest…
  - raphaelrk 4y ago
    What I meant was I expected numpy to be faster than the js libraries I was testing, simply because people use it so much more, for real "scientific computing" work. And indeed it is very fast given it only uses the CPU, but that still leaves its matrix multiplication as ~100x slower than what my mac is capable of.
- psyklic 4y ago
  It adds a great deal of complexity (and often user frustration) for a Python package to support both CPU and GPGPU.
- dhasso 4y ago
  For that, you can use cupy[0], PyTorch[1] or Tensorflow[2]. They all mimic the numpy's API with the possibility to use your GPU. [0] https://cupy.dev/ https://cupy.dev/ [1] https://pytorch.org/ https://pytorch.org/ [2] https://www.tensorflow.org/ https://www.tensorflow.org/

sampo 4y ago

TensorFlow.js matrices are immutable, which puts more restrictions on your programming style that standard Numpy. You cat get immutable, GPU-enhanced matrices for Python, too.