A new golden age for computer architecture

TL;DR: This paper explores Domain-Specific Deep Learning Architectures for GPU Computer Vision through a "brainstorming" approach on selected hands-on topics in the area through tools, frameworks and data pipelines commonly used to train and deploy DNNs in GPUs and Domain- Specific Architectures (DSAs).

TL;DR: IXIAM as discussed by the authors is a cost-effective HW-SW framework to control a wide variety of accelerators in a standard way, and directly from the cores, including reservation, work offloading, data transfer, and synchronization.

TL;DR: ForMS as mentioned in this paper is a fine-grained ReRAM-based DNN accelerator with polarized weights, which enforce exactly what is assumed in the in-situ computation, ensuring that all weights in the same column of a crossbar have the same sign.

TL;DR: Integrated circuits will lead to such wonders as home computers or at least terminals connected to a central computer, automatic controls for automobiles, and personal portable communications equipment as mentioned in this paper. But the biggest potential lies in the production of large systems.

TL;DR: The future of integrated electronics is the future of electronics itself, and the advantages of integration will bring about a proliferation of electronics, pushing this science into many new areas.

TL;DR: Integrated circuits will lead to such wonders as home computers or at least terminals connected to a central computer, automatic controls for automobiles, and personal portable communications equipment as discussed by the authors. But the biggest potential lies in the production of large systems.

TL;DR: This paper evaluates a custom ASIC-called a Tensor Processing Unit (TPU)-deployed in datacenters since 2015 that accelerates the inference phase of neural networks (NN) and compares it to a server-class Intel Haswell CPU and an Nvidia K80 GPU, which are contemporaries deployed in the samedatacenters.