undefined | Better HN

0 pointswaynecochran2y ago0 comments

The document said it outputs the exact same values as the conventional method. There is no accuracy trade off here.

0 comments

The paper cited is about hardware, where there is no accuracy tradeoff because you control the numerical precision completely and use fixed point. In a software implementation, neither is true. There is no chance that you will get the exact same values out of this method that you do out of other FP matmuls.

p1esk2y ago

For floating point? Are you sure?

waynecochranOP2y ago

Opening statement of README

    This repository contains the source code for ML hardware architectures that 
    require nearly half the number of multiplier units to achieve the same 
    performance, by executing alternative inner-product algorithms that trade 
    nearly half the multiplications for cheap low-bitwidth additions, while still 
    producing identical output as the conventional inner product.

p1esk2y ago

I just looked at the paper: the answer is no, floating point is not supported.

j / k navigate · click thread line to collapse

0 comments

pclmulqdq2y ago

p1esk2y ago

For floating point? Are you sure?

waynecochranOP2y ago

Opening statement of README

    This repository contains the source code for ML hardware architectures that 
    require nearly half the number of multiplier units to achieve the same 
    performance, by executing alternative inner-product algorithms that trade 
    nearly half the multiplications for cheap low-bitwidth additions, while still 
    producing identical output as the conventional inner product.

p1esk2y ago

I just looked at the paper: the answer is no, floating point is not supported.

j / k navigate · click thread line to collapse