Mustafa Ali

Tanvi Sharma, Mustafa Ali, Indranil Chakraborty et al.

Matrix multiplication is the dominant computation during Machine Learning (ML) inference. To efficiently perform such multiplication operations, Compute-in-memory (CiM) paradigms have emerged as a highly energy efficient solution. However, integrating compute in memory poses key questions, such as 1) What type of CiM to use: Given a multitude of CiM design characteristics, determining their suitability from architecture perspective is needed. 2) When to use CiM: ML inference includes workloads with a variety of memory and compute requirements, making it difficult to identify when CiM is more beneficial than standard processing cores. 3) Where to integrate CiM: Each memory level has different bandwidth and capacity, creating different data reuse opportunities for CiM integration. To answer such questions regarding on-chip CiM integration for accelerating ML workloads, we use an analytical architecture-evaluation methodology with tailored mapping algorithm. The mapping algorithm aims to achieve highest weight reuse and reduced data movements for a given CiM prototype and workload. Our analysis considers the integration of CiM prototypes into the cache levels of a tensor-core-like architecture, and shows that CiM integrated memory improves energy efficiency by up to 3.4x and throughput by up to 15.6x compared to established baseline with INT-8 precision. We believe the proposed work provides insights into what type of CiM to use, and when and where to optimally integrate it in the cache hierarchy for efficient matrix multiplication.

Jiawei Zhao, Steve Dai, Rangharajan Venkatesan et al.

Representing deep neural networks (DNNs) in low-precision is a promising approach to enable efficient acceleration and memory reduction. Previous methods that train DNNs in low-precision typically keep a copy of weights in high-precision during the weight updates. Directly training with low-precision weights leads to accuracy degradation due to complex interactions between the low-precision number systems and the learning algorithms. To address this issue, we develop a co-designed low-precision training framework, termed LNS-Madam, in which we jointly design a logarithmic number system (LNS) and a multiplicative weight update algorithm (Madam). We prove that LNS-Madam results in low quantization error during weight updates, leading to stable performance even if the precision is limited. We further propose a hardware design of LNS-Madam that resolves practical challenges in implementing an efficient datapath for LNS computations. Our implementation effectively reduces energy overhead incurred by LNS-to-integer conversion and partial sum accumulation. Experimental results show that LNS-Madam achieves comparable accuracy to full-precision counterparts with only 8 bits on popular computer vision and natural language tasks. Compared to FP32 and FP8, LNS-Madam reduces the energy consumption by over 90% and 55%, respectively.

Sourjya Roy, Mustafa Ali, Anand Raghunathan

Deep Neural Networks (DNNs) have transformed the field of machine learning and are widely deployed in many applications involving image, video, speech and natural language processing. The increasing compute demands of DNNs have been widely addressed through Graphics Processing Units (GPUs) and specialized accelerators. However, as model sizes grow, these von Neumann architectures require very high memory bandwidth to keep the processing elements utilized as a majority of the data resides in the main memory. Processing in memory has been proposed as a promising solution for the memory wall bottleneck for ML workloads. In this work, we propose a new DRAM-based processing-in-memory (PIM) multiplication primitive coupled with intra-bank accumulation to accelerate matrix vector operations in ML workloads. The proposed multiplication primitive adds < 1% area overhead and does not require any change in the DRAM peripherals. Therefore, the proposed multiplication can be easily adopted in commodity DRAM chips. Subsequently, we design a DRAM-based PIM architecture, data mapping scheme and dataflow for executing DNNs within DRAM. System evaluations performed on networks like AlexNet, VGG16 and ResNet18 show that the proposed architecture, mapping, and data flow can provide up to 19.5x speedup over an NVIDIA Titan Xp GPU highlighting the need to overcome the memory bottleneck in future generations of DNN hardware.

Mustafa Ali, Akhilesh Jaiswal, Sangamesh Kodge et al.

`In-memory computing' is being widely explored as a novel computing paradigm to mitigate the well known memory bottleneck. This emerging paradigm aims at embedding some aspects of computations inside the memory array, thereby avoiding frequent and expensive movement of data between the compute unit and the storage memory. In-memory computing with respect to Silicon memories has been widely explored on various memory bit-cells. Embedding computation inside the 6 transistor (6T) SRAM array is of special interest since it is the most widely used on-chip memory. In this paper, we present a novel in-memory multiplication followed by accumulation operation capable of performing parallel dot products within 6T SRAM without any changes to the standard bitcell. We, further, study the effect of circuit non-idealities and process variations on the accuracy of the LeNet-5 and VGG neural network architectures against the MNIST and CIFAR-10 datasets, respectively. The proposed in-memory dot-product mechanism achieves 88.8% and 99% accuracy for the CIFAR-10 and MNIST, respectively. Compared to the standard von Neumann system, the proposed system is 6.24x better in energy consumption and 9.42x better in delay.