One line. Many voicesSeek and you shall find

muni.works

Don't Write Kernels, Derive Them

  • Save1 save
  • tang's tensor IR derives high-performance GPU and CPU kernels entirely through algebraic rewrites of two constructs (affine views and monoid comprehensions), matching or beating hand-tuned libraries on some workloads while trailing on others.

    Waiting on you

    Sign upLog in

    read later

    For all the tabs you promised to read.
    Save to read. Read to clear.