no code implementations • 14 Apr 2024 • Tian Jin, Wanzin Yazar, Zifei Xu, Sayeh Sharify, Xin Wang
We demonstrate that using this custom CUDA kernel improves the throughput of LLM inference by 28%.
no code implementations • 11 Jul 2023 • Zihao Deng, Xin Wang, Sayeh Sharify, Michael Orshansky
Quantization assigning the same bit-width to all layers leads to large accuracy degradation at low precision and is wasteful at high precision settings.
no code implementations • 10 May 2018 • Sayeh Sharify, Mostafa Mahmoud, Alberto Delmas Lascorz, Milos Nikolic, Andreas Moshovos
A Laconic configuration that uses a 1K-wire weight memory interface, outperforms the 2K-wire conventional accelerator by 15. 4x and is 1. 95x more energy efficient.
no code implementations • 17 Apr 2018 • Alberto Delmas, Sayeh Sharify, Patrick Judd, Kevin Siu, Milos Nikolic, Andreas Moshovos
The per group precisions are selected statically for the weights and dynamically by hardware for the activations.
no code implementations • 9 Mar 2018 • Alberto Delmas, Patrick Judd, Dylan Malone Stuart, Zissis Poulos, Mostafa Mahmoud, Sayeh Sharify, Milos Nikolic, Andreas Moshovos
We show that, during inference with Convolutional Neural Networks (CNNs), more than 2x to $8x ineffectual work can be exposed if instead of targeting those weights and activations that are zero, we target different combinations of value stream properties.
no code implementations • 27 Jul 2017 • Alberto Delmas, Sayeh Sharify, Patrick Judd, Andreas Moshovos
Experiments on image classification CNNs show that on average across all networks studied, TRT outperforms a state-of-the-art bit-parallel accelerator by 1:90x without any loss in accuracy while it is 1:17x more energy efficient.
no code implementations • 23 Jun 2017 • Sayeh Sharify, Alberto Delmas Lascorz, Kevin Siu, Patrick Judd, Andreas Moshovos
LM can trade-off accuracy for additional improvements in execution performance and energy efficiency and compares favorably to an accelerator that targeted only activation precisions.
no code implementations • 1 Jun 2017 • Alberto Delmas, Patrick Judd, Sayeh Sharify, Andreas Moshovos
Stripes is a Deep Neural Network (DNN) accelerator that uses bit-serial computation to offer performance that is proportional to the fixed-point precision of the activation values.
no code implementations • 29 Apr 2017 • Patrick Judd, Alberto Delmas, Sayeh Sharify, Andreas Moshovos
We also present a modified organization that detects the activations that are deemed as ineffectual while fetching them from memory.