copy-pastries

3 ms·

1. 1.▲
   by copy-pastries 6y ago

   Thanks! Yes, I work as an HPC researcher, but not for Cerebras. I just watched the presentation of this paper at the SC20 Supercomputing conference last week. I probably should have said, according to the specs, a CS-1 does have loads (TBs?
2. 2.▲
   by copy-pastries 6y ago

   These are great questions :-) If I remember correctly from the paper, the authors used quite small grids, but they were actually 3D, like your question. They gave each "slice" (one full dimension of the tensorx1x1) to a differen
3. 3.▲
   by copy-pastries 6y ago

   Yes, it's pretty much exactly what you said. Algorithms like those for parallel solvers for systems of equations will divide the problem up into blocks of cells for parallel processing. But typically each cell's next state calcula