Fully realizing the potential of acceleration for Deep Neural Networks (DNNs)
requires understanding and leveraging algorithmic properties. This paper builds
upon the algorithmic insight that bitwidth of operations in DNNs can be reduced
without compromising their classification accuracy. However, to prevent
accuracy loss, the bitwidth varies significantly across DNNs and it may even be
adjusted for each layer. Thus, a fixed-bitwidth accelerator would either offer
limited benefits to accommodate the worst-case bitwidth requirements, or lead
to a degradation in final accuracy. To alleviate these deficiencies, this work
introduces dynamic bit-level fusion/decomposition as a new dimension in the
design of DNN accelerators. We explore this dimension by designing Bit Fusion,
a bit-flexible accelerator, that constitutes an array of bit-level processing
elements that dynamically fuse to match the bitwidth of individual DNN layers.
This flexibility in the architecture enables minimizing the computation and the
communication at the finest granularity possible with no loss in accuracy. We
evaluate the benefits of BitFusion using eight real-world feed-forward and
recurrent DNNs. The proposed microarchitecture is implemented in Verilog and
synthesized in 45 nm technology. Using the synthesis results and cycle accurate
simulation, we compare the benefits of Bit Fusion to two state-of-the-art DNN
accelerators, Eyeriss and Stripes. In the same area, frequency, and process
technology, BitFusion offers 3.9x speedup and 5.1x energy savings over Eyeriss.
Compared to Stripes, BitFusion provides 2.6x speedup and 3.9x energy reduction
at 45 nm node when BitFusion area and frequency are set to those of Stripes.
Scaling to GPU technology node of 16 nm, BitFusion almost matches the
performance of a 250-Watt Titan Xp, which uses 8-bit vector instructions, while
BitFusion merely consumes 895 milliwatts of power.

本文提出一种新的 DNN 加速器设计 ——Bit Fusion，通过引入动态位级融合／分解，实现了对 DNN 层位宽进行动态匹配，提高了计算性能且无损精度。通过与现有 DNN 加速器模型的比较实验，Bit Fusion 在面积、频率与工艺技术相同的情况下，相较于 Eyeriss 加速了 3.9 倍并节省了 5.1 倍的能源，相较于 Stripes 加速了 2.6 倍并减少了 3.9 倍的能源。在 16nm 的 GPU 技术节点上，Bit Fusion 的性能远超过使用 8 位向量指令的 250 瓦 Titan Xp，消耗仅为 895 毫瓦。