4 ms·
There has been a resurgence due to the proliferation of more special-purpose processors, like Google TPU or Tesla FSD. I anticipate this trend to continue due t
by mehrdada 7y ago
There has been a resurgence due to the proliferation of more special-purpose processors, like Google TPU or Tesla FSD. I anticipate this trend to continue due to Moore's Law slowdown. However, if you're not really into this, there's no reason to not pursue some other subset as on a relative basis, higher levels of the stack likely will have many more opportunities than infrastructure.
- logcrater 7y agoIs there a way I can integrate Machine Learning to Compiler stuff? I've been meddling with ml for a while and it's pretty interesting too.
- deepnotderp 7y agoCertainly! We do so :)
- logcrater 7y agoCan you elaborate on that please? Like how do they merge together? I am really interested in those two fields
- pjmlp 7y agoCheck what LLVM is doing with the MLIR project. https://blog.tensorflow.org/2019/04/mlir-new-intermediate-representation.html https://blog.tensorflow.org/2019/04/mlir-new-intermediate-re... https://llvm.org/devmtg/2019-04/talks.html https://llvm.org/devmtg/2019-04/talks.html
- sanxiyn 7y agoYou may enjoy the paper "Learning to Optimize Tensor Programs": https://arxiv.org/abs/1805.08166 https://arxiv.org/abs/1805.08166
- deepnotderp 7y agoThis is as good an overview as any: https://arxiv.org/pdf/1805.03441.pdf https://arxiv.org/pdf/1805.03441.pdf In general we can afford to treat the problem as a search problem since our kernels are small enough. Here there is a lot of stuff that ML aided stochastic search is amazing at compared to normal compilers and even in some cases human hand optimized assembly (!).