List view

• v0.0.3 release

  better parallelization kernels for CPU(with SIMD & BLAS), better Scalar value handling, fixing more bugs in both C/CPP & Python codes. Direct C-API. supports CUDA ops (optional)

  Due by February 1, 2026

  •1/3 issues closed
  33% complete2 open1 closed

• More efficient & faster cpu-matmul using the `transpose -> dot-product`

  need to implement parallelization kernels for faster matmul for current version, since this doesn't use naive matmul instead uses transpose then dot-product, it's faster than naive one by 6-7x but we need to make it even faster

  No due date
  0% complete0 open0 closed