@article{anderson_scalar_2019,
	title = {Scalar Arithmetic Multiple Data: Customizable Precision for Deep Neural Networks},
	url = {http://arxiv.org/abs/1809.10572},
	doi = {10.1109/ARITH.2019.00018},
	shorttitle = {Scalar Arithmetic Multiple Data},
	abstract = {Quantization of weights and activations in Deep Neural Networks ({DNNs}) is a powerful technique for network compression, and has enjoyed significant attention and success. However, much of the inference-time benefit of quantization is accessible only through the use of customized hardware accelerators or by providing an {FPGA} implementation of quantized arithmetic. Building on prior work, we show how to construct arbitrary bit-precise signed and unsigned integer operations using a software technique which logically {\textbackslash}emph\{embeds\} a vector architecture with custom bit-width lanes in universally available fixed-width scalar arithmetic. We evaluate our approach on a high-end Intel Haswell processor, and an embedded {ARM} processor. Our approach yields very fast implementations of bit-precise custom {DNN} operations, which often match or exceed the performance of operations quantized to the sizes supported in native arithmetic. At the strongest level of quantization, our approach yields a maximum speedup of \${\textbackslash}thicksim6{\textbackslash}times\$ on the Intel platform, and \${\textbackslash}thicksim10{\textbackslash}times\$ on the {ARM} platform versus quantization to native 8-bit integers.},
	pages = {61--68},
	journaltitle = {2019 {IEEE} 26th Symposium on Computer Arithmetic ({ARITH})},
	author = {Anderson, Andrew and Gregg, David},
	urldate = {2019-12-17},
	date = {2019-06},
	eprinttype = {arxiv},
	eprint = {1809.10572},
	keywords = {Computer Science - Computer Vision and Pattern Recognition, Computer Science - Mathematical Software, Computer Science - Performance},
}