Showing 1–18 of 18 results for author: Georganas, E

DECA: A Near-Core LLM Decompression Accelerator Supporting Out-of-Order Invocation

Authors: Gerasimos Gerogiannis, Stijn Eyerman, Evangelos Georganas, Wim Heirman, Josep Torrellas

Abstract: To alleviate the memory bandwidth bottleneck in Large Language Model (LLM) inference workloads, weight matrices are stored in memory in quantized and sparsified formats. Hence, before tiles of these matrices can be processed by in-core generalized matrix multiplication (GeMM) hardware engines, they need to be dequantized and de-sparsified. This is currently performed in software with vector operat… ▽ More To alleviate the memory bandwidth bottleneck in Large Language Model (LLM) inference workloads, weight matrices are stored in memory in quantized and sparsified formats. Hence, before tiles of these matrices can be processed by in-core generalized matrix multiplication (GeMM) hardware engines, they need to be dequantized and de-sparsified. This is currently performed in software with vector operations. Unfortunately, this approach delivers only modest performance. Moreover, it is hard to understand how to improve the system, as the overall GeMM performance depends on the interaction between memory resources, vector units, and hardware matrix engines. To improve the performance of LLM inference in advanced platforms equipped with in-core GeMM engines and HBM, this paper makes three main contributions. First, it develops an analytical performance model with a 3D visual representation that provides insights into how memory resources, vector units, and hardware matrix engines interact to deliver compressed GeMM performance. Second, it proposes DECA, a new near-core ML-model decompression accelerator. DECA offloads tile de-sparsification and dequantization from the CPU, producing ready-to-use tiles for in-core GeMM engines. Third, it introduces a new ISA extension that enables out-of-order invocation of the near-core accelerator. With this extension, accelerator and core computations can interleave and overlap with high-performance. Our evaluation shows that, in a simulated 56-core Xeon 4 server with HBM, DECA accelerates the execution of compressed GeMMs by up to 4x over the use of optimized Intel software kernels. Further, DECA reduces the next-token generation time of Llama2-70B and OPT-66B by 1.6x-2.6x. △ Less

Submitted 25 May, 2025; originally announced May 2025.

ML-SpecQD: Multi-Level Speculative Decoding with Quantized Drafts

Authors: Evangelos Georganas, Dhiraj Kalamkar, Alexander Kozlov, Alexander Heinecke

Abstract: Speculative decoding (SD) has emerged as a method to accelerate LLM inference without sacrificing any accuracy over the 16-bit model inference. In a typical SD setup, the idea is to use a full-precision, small, fast model as "draft" to generate the next few tokens and use the "target" large model to verify the draft-generated tokens. The efficacy of this method heavily relies on the acceptance rat… ▽ More Speculative decoding (SD) has emerged as a method to accelerate LLM inference without sacrificing any accuracy over the 16-bit model inference. In a typical SD setup, the idea is to use a full-precision, small, fast model as "draft" to generate the next few tokens and use the "target" large model to verify the draft-generated tokens. The efficacy of this method heavily relies on the acceptance ratio of the draft-generated tokens and the relative token throughput of the draft versus the target model. Nevertheless, an efficient SD pipeline requires pre-training and aligning the draft model to the target model, making it impractical for LLM inference in a plug-and-play fashion. In this work, we propose using MXFP4 models as drafts in a plug-and-play fashion since the MXFP4 Weight-Only-Quantization (WOQ) merely direct-casts the BF16 target model weights to MXFP4. In practice, our plug-and-play solution gives speedups up to 2x over the BF16 baseline. Then we pursue an opportunity for further acceleration: the MXFP4 draft token generation itself can be accelerated via speculative decoding by using yet another smaller draft. We call our method ML-SpecQD: Multi-Level Speculative Decoding with Quantized Drafts since it recursively applies speculation for accelerating the draft-token generation. Combining Multi-Level Speculative Decoding with MXFP4 Quantized Drafts we outperform state-of-the-art speculative decoding, yielding speedups up to 2.72x over the BF16 baseline. △ Less

Submitted 17 March, 2025; originally announced March 2025.

Towards a high-performance AI compiler with upstream MLIR

Authors: Renato Golin, Lorenzo Chelini, Adam Siemieniuk, Kavitha Madhu, Niranjan Hasabnis, Hans Pabst, Evangelos Georganas, Alexander Heinecke

Abstract: This work proposes a compilation flow using open-source compiler passes to build a framework to achieve ninja performance from a generic linear algebra high-level abstraction. We demonstrate this flow with a proof-of-concept MLIR project that uses input IR in Linalg-on-Tensor from TensorFlow and PyTorch, performs cache-level optimizations and lowering to micro-kernels for efficient vectorization,… ▽ More This work proposes a compilation flow using open-source compiler passes to build a framework to achieve ninja performance from a generic linear algebra high-level abstraction. We demonstrate this flow with a proof-of-concept MLIR project that uses input IR in Linalg-on-Tensor from TensorFlow and PyTorch, performs cache-level optimizations and lowering to micro-kernels for efficient vectorization, achieving over 90% of the performance of ninja-written equivalent programs. The contributions of this work include: (1) Packing primitives on the tensor dialect and passes for cache-aware distribution of tensors (single and multi-core) and type-aware instructions (VNNI, BFDOT, BFMMLA), including propagation of shapes across the entire function; (2) A linear algebra pipeline, including tile, fuse and bufferization strategies to get model-level IR into hardware friendly tile calls; (3) A mechanism for micro-kernel lowering to an open source library that supports various CPUs. △ Less

Submitted 15 April, 2024; originally announced April 2024.

Comments: 13 pages, 8 figures, presented at CGO C4ML 2024 & MLIR Workshop EuroLLVM 2024

Harnessing Deep Learning and HPC Kernels via High-Level Loop and Tensor Abstractions on CPU Architectures

Authors: Evangelos Georganas, Dhiraj Kalamkar, Kirill Voronin, Abhisek Kundu, Antonio Noack, Hans Pabst, Alexander Breuer, Alexander Heinecke

Abstract: During the past decade, Deep Learning (DL) algorithms, programming systems and hardware have converged with the High Performance Computing (HPC) counterparts. Nevertheless, the programming methodology of DL and HPC systems is stagnant, relying on highly-optimized, yet platform-specific and inflexible vendor-optimized libraries. Such libraries provide close-to-peak performance on specific platforms… ▽ More During the past decade, Deep Learning (DL) algorithms, programming systems and hardware have converged with the High Performance Computing (HPC) counterparts. Nevertheless, the programming methodology of DL and HPC systems is stagnant, relying on highly-optimized, yet platform-specific and inflexible vendor-optimized libraries. Such libraries provide close-to-peak performance on specific platforms, kernels and shapes thereof that vendors have dedicated optimizations efforts, while they underperform in the remaining use-cases, yielding non-portable codes with performance glass-jaws. This work introduces a framework to develop efficient, portable DL and HPC kernels for modern CPU architectures. We decompose the kernel development in two steps: 1) Expressing the computational core using Tensor Processing Primitives (TPPs): a compact, versatile set of 2D-tensor operators, 2) Expressing the logical loops around TPPs in a high-level, declarative fashion whereas the exact instantiation (ordering, tiling, parallelization) is determined via simple knobs. We demonstrate the efficacy of our approach using standalone kernels and end-to-end workloads that outperform state-of-the-art implementations on diverse CPU platforms. △ Less

Submitted 15 March, 2024; v1 submitted 25 April, 2023; originally announced April 2023.

FPGA-based AI Smart NICs for Scalable Distributed AI Training Systems

Authors: Rui Ma, Evangelos Georganas, Alexander Heinecke, Andrew Boutros, Eriko Nurvitadhi

Abstract: Rapid advances in artificial intelligence (AI) technology have led to significant accuracy improvements in a myriad of application domains at the cost of larger and more compute-intensive models. Training such models on massive amounts of data typically requires scaling to many compute nodes and relies heavily on collective communication algorithms, such as all-reduce, to exchange the weight gradi… ▽ More Rapid advances in artificial intelligence (AI) technology have led to significant accuracy improvements in a myriad of application domains at the cost of larger and more compute-intensive models. Training such models on massive amounts of data typically requires scaling to many compute nodes and relies heavily on collective communication algorithms, such as all-reduce, to exchange the weight gradients between different nodes. The overhead of these collective communication operations in a distributed AI training system can bottleneck its performance, with more pronounced effects as the number of nodes increases. In this paper, we first characterize the all-reduce operation overhead by profiling distributed AI training. Then, we propose a new smart network interface card (NIC) for distributed AI training systems using field-programmable gate arrays (FPGAs) to accelerate all-reduce operations and optimize network bandwidth utilization via data compression. The AI smart NIC frees up the system's compute resources to perform the more compute-intensive tensor operations and increases the overall node-to-node communication efficiency. We perform real measurements on a prototype distributed AI training system comprised of 6 compute nodes to evaluate the performance gains of our proposed FPGA-based AI smart NIC compared to a baseline system with regular NICs. We also use these measurements to validate an analytical model that we formulate to predict performance when scaling to larger systems. Our proposed FPGA-based AI smart NIC enhances overall training performance by 1.6x at 6 nodes, with an estimated 2.5x performance improvement at 32 nodes, compared to the baseline system using conventional NICs. △ Less

Submitted 22 April, 2022; originally announced April 2022.

Comments: 5 pages, 4 figures

Efficient and Generic 1D Dilated Convolution Layer for Deep Learning

Authors: Narendra Chaudhary, Sanchit Misra, Dhiraj Kalamkar, Alexander Heinecke, Evangelos Georganas, Barukh Ziv, Menachem Adelman, Bharat Kaul

Abstract: Convolutional neural networks (CNNs) have found many applications in tasks involving two-dimensional (2D) data, such as image classification and image processing. Therefore, 2D convolution layers have been heavily optimized on CPUs and GPUs. However, in many applications - for example genomics and speech recognition, the data can be one-dimensional (1D). Such applications can benefit from optimize… ▽ More Convolutional neural networks (CNNs) have found many applications in tasks involving two-dimensional (2D) data, such as image classification and image processing. Therefore, 2D convolution layers have been heavily optimized on CPUs and GPUs. However, in many applications - for example genomics and speech recognition, the data can be one-dimensional (1D). Such applications can benefit from optimized 1D convolution layers. In this work, we introduce our efficient implementation of a generic 1D convolution layer covering a wide range of parameters. It is optimized for x86 CPU architectures, in particular, for architectures containing Intel AVX-512 and AVX-512 BFloat16 instructions. We use the LIBXSMM library's batch-reduce General Matrix Multiplication (BRGEMM) kernel for FP32 and BFloat16 precision. We demonstrate that our implementation can achieve up to 80% efficiency on Intel Xeon Cascade Lake and Cooper Lake CPUs. Additionally, we show the generalization capability of our BRGEMM based approach by achieving high efficiency across a range of parameters. We consistently achieve higher efficiency than the 1D convolution layer with Intel oneDNN library backend for varying input tensor widths, filter widths, number of channels, filters, and dilation parameters. Finally, we demonstrate the performance of our optimized 1D convolution layer by utilizing it in the end-to-end neural network training with real genomics datasets and achieve up to 6.86x speedup over the oneDNN library-based implementation on Cascade Lake CPUs. We also demonstrate the scaling with 16 sockets of Cascade/Cooper Lake CPUs and achieve significant speedup over eight V100 GPUs using a similar power envelop. In the end-to-end training, we get a speedup of 1.41x on Cascade Lake with FP32, 1.57x on Cooper Lake with FP32, and 2.27x on Cooper Lake with BFloat16 over eight V100 GPUs with FP32. △ Less

Submitted 16 April, 2021; originally announced April 2021.

DistGNN: Scalable Distributed Training for Large-Scale Graph Neural Networks

Authors: Vasimuddin Md, Sanchit Misra, Guixiang Ma, Ramanarayan Mohanty, Evangelos Georganas, Alexander Heinecke, Dhiraj Kalamkar, Nesreen K. Ahmed, Sasikanth Avancha

Abstract: Full-batch training on Graph Neural Networks (GNN) to learn the structure of large graphs is a critical problem that needs to scale to hundreds of compute nodes to be feasible. It is challenging due to large memory capacity and bandwidth requirements on a single compute node and high communication volumes across multiple nodes. In this paper, we present DistGNN that optimizes the well-known Deep G… ▽ More Full-batch training on Graph Neural Networks (GNN) to learn the structure of large graphs is a critical problem that needs to scale to hundreds of compute nodes to be feasible. It is challenging due to large memory capacity and bandwidth requirements on a single compute node and high communication volumes across multiple nodes. In this paper, we present DistGNN that optimizes the well-known Deep Graph Library (DGL) for full-batch training on CPU clusters via an efficient shared memory implementation, communication reduction using a minimum vertex-cut graph partitioning algorithm and communication avoidance using a family of delayed-update algorithms. Our results on four common GNN benchmark datasets: Reddit, OGB-Products, OGB-Papers and Proteins, show up to 3.7x speed-up using a single CPU socket and up to 97x speed-up using 128 CPU sockets, respectively, over baseline DGL implementations running on a single CPU socket △ Less

Submitted 16 April, 2021; v1 submitted 14 April, 2021; originally announced April 2021.

Tensor Processing Primitives: A Programming Abstraction for Efficiency and Portability in Deep Learning & HPC Workloads

Authors: Evangelos Georganas, Dhiraj Kalamkar, Sasikanth Avancha, Menachem Adelman, Deepti Aggarwal, Cristina Anderson, Alexander Breuer, Jeremy Bruestle, Narendra Chaudhary, Abhisek Kundu, Denise Kutnick, Frank Laub, Vasimuddin Md, Sanchit Misra, Ramanarayan Mohanty, Hans Pabst, Brian Retford, Barukh Ziv, Alexander Heinecke

Abstract: During the past decade, novel Deep Learning (DL) algorithms, workloads and hardware have been developed to tackle a wide range of problems. Despite the advances in workload and hardware ecosystems, the programming methodology of DL systems is stagnant. DL workloads leverage either highly-optimized, yet platform-specific and inflexible kernels from DL libraries, or in the case of novel operators, r… ▽ More During the past decade, novel Deep Learning (DL) algorithms, workloads and hardware have been developed to tackle a wide range of problems. Despite the advances in workload and hardware ecosystems, the programming methodology of DL systems is stagnant. DL workloads leverage either highly-optimized, yet platform-specific and inflexible kernels from DL libraries, or in the case of novel operators, reference implementations are built via DL framework primitives with underwhelming performance. This work introduces the Tensor Processing Primitives (TPP), a programming abstraction striving for efficient, portable implementation of DL workloads with high-productivity. TPPs define a compact, yet versatile set of 2D-tensor operators (or a virtual Tensor ISA), which subsequently can be utilized as building-blocks to construct complex operators on high-dimensional tensors. The TPP specification is platform-agnostic, thus code expressed via TPPs is portable, whereas the TPP implementation is highly-optimized and platform-specific. We demonstrate the efficacy and viability of our approach using standalone kernels and end-to-end DL & HPC workloads expressed entirely via TPPs that outperform state-of-the-art implementations on multiple platforms. △ Less

Submitted 30 November, 2021; v1 submitted 12 April, 2021; originally announced April 2021.

Optimizing Deep Learning Recommender Systems' Training On CPU Cluster Architectures

Authors: Dhiraj Kalamkar, Evangelos Georganas, Sudarshan Srinivasan, Jianping Chen, Mikhail Shiryaev, Alexander Heinecke

Abstract: During the last two years, the goal of many researchers has been to squeeze the last bit of performance out of HPC system for AI tasks. Often this discussion is held in the context of how fast ResNet50 can be trained. Unfortunately, ResNet50 is no longer a representative workload in 2020. Thus, we focus on Recommender Systems which account for most of the AI cycles in cloud computing centers. More… ▽ More During the last two years, the goal of many researchers has been to squeeze the last bit of performance out of HPC system for AI tasks. Often this discussion is held in the context of how fast ResNet50 can be trained. Unfortunately, ResNet50 is no longer a representative workload in 2020. Thus, we focus on Recommender Systems which account for most of the AI cycles in cloud computing centers. More specifically, we focus on Facebook's DLRM benchmark. By enabling it to run on latest CPU hardware and software tailored for HPC, we are able to achieve more than two-orders of magnitude improvement in performance (110x) on a single socket compared to the reference CPU implementation, and high scaling efficiency up to 64 sockets, while fitting ultra-large datasets. This paper discusses the optimization techniques for the various operators in DLRM and which component of the systems are stressed by these different operators. The presented techniques are applicable to a broader set of DL workloads that pose the same scaling challenges/characteristics as DLRM. △ Less

Submitted 10 May, 2020; originally announced May 2020.

The Parallelism Motifs of Genomic Data Analysis

Authors: Katherine Yelick, Aydin Buluc, Muaaz Awan, Ariful Azad, Benjamin Brock, Rob Egan, Saliya Ekanayake, Marquita Ellis, Evangelos Georganas, Giulia Guidi, Steven Hofmeyr, Oguz Selvitopi, Cristina Teodoropol, Leonid Oliker

Abstract: Genomic data sets are growing dramatically as the cost of sequencing continues to decline and small sequencing devices become available. Enormous community databases store and share this data with the research community, but some of these genomic data analysis problems require large scale computational platforms to meet both the memory and computational requirements. These applications differ from… ▽ More Genomic data sets are growing dramatically as the cost of sequencing continues to decline and small sequencing devices become available. Enormous community databases store and share this data with the research community, but some of these genomic data analysis problems require large scale computational platforms to meet both the memory and computational requirements. These applications differ from scientific simulations that dominate the workload on high end parallel systems today and place different requirements on programming support, software libraries, and parallel architectural design. For example, they involve irregular communication patterns such as asynchronous updates to shared data structures. We consider several problems in high performance genomics analysis, including alignment, profiling, clustering, and assembly for both single genomes and metagenomes. We identify some of the common computational patterns or motifs that help inform parallelization strategies and compare our motifs to some of the established lists, arguing that at least two key patterns, sorting and hashing, are missing. △ Less

Submitted 20 January, 2020; originally announced January 2020.

High-Performance Deep Learning via a Single Building Block

Authors: Evangelos Georganas, Kunal Banerjee, Dhiraj Kalamkar, Sasikanth Avancha, Anand Venkat, Michael Anderson, Greg Henry, Hans Pabst, Alexander Heinecke

Abstract: Deep learning (DL) is one of the most prominent branches of machine learning. Due to the immense computational cost of DL workloads, industry and academia have developed DL libraries with highly-specialized kernels for each workload/architecture, leading to numerous, complex code-bases that strive for performance, yet they are hard to maintain and do not generalize. In this work, we introduce the… ▽ More Deep learning (DL) is one of the most prominent branches of machine learning. Due to the immense computational cost of DL workloads, industry and academia have developed DL libraries with highly-specialized kernels for each workload/architecture, leading to numerous, complex code-bases that strive for performance, yet they are hard to maintain and do not generalize. In this work, we introduce the batch-reduce GEMM kernel and show how the most popular DL algorithms can be formulated with this kernel as the basic building-block. Consequently, the DL library-development degenerates to mere (potentially automatic) tuning of loops around this sole optimized kernel. By exploiting our new kernel we implement Recurrent Neural Networks, Convolution Neural Networks and Multilayer Perceptron training and inference primitives in just 3K lines of high-level code. Our primitives outperform vendor-optimized libraries on multi-node CPU clusters, and we also provide proof-of-concept CNN kernels targeting GPUs. Finally, we demonstrate that the batch-reduce GEMM kernel within a tensor compiler yields high-performance CNN primitives, further amplifying the viability of our approach. △ Less

Submitted 17 June, 2019; v1 submitted 14 June, 2019; originally announced June 2019.

A Study of BFLOAT16 for Deep Learning Training

Authors: Dhiraj Kalamkar, Dheevatsa Mudigere, Naveen Mellempudi, Dipankar Das, Kunal Banerjee, Sasikanth Avancha, Dharma Teja Vooturi, Nataraj Jammalamadaka, Jianyu Huang, Hector Yuen, Jiyan Yang, Jongsoo Park, Alexander Heinecke, Evangelos Georganas, Sudarshan Srinivasan, Abhisek Kundu, Misha Smelyanskiy, Bharat Kaul, Pradeep Dubey

Abstract: This paper presents the first comprehensive empirical study demonstrating the efficacy of the Brain Floating Point (BFLOAT16) half-precision format for Deep Learning training across image classification, speech recognition, language modeling, generative networks and industrial recommendation systems. BFLOAT16 is attractive for Deep Learning training for two reasons: the range of values it can repr… ▽ More This paper presents the first comprehensive empirical study demonstrating the efficacy of the Brain Floating Point (BFLOAT16) half-precision format for Deep Learning training across image classification, speech recognition, language modeling, generative networks and industrial recommendation systems. BFLOAT16 is attractive for Deep Learning training for two reasons: the range of values it can represent is the same as that of IEEE 754 floating-point format (FP32) and conversion to/from FP32 is simple. Maintaining the same range as FP32 is important to ensure that no hyper-parameter tuning is required for convergence; e.g., IEEE 754 compliant half-precision floating point (FP16) requires hyper-parameter tuning. In this paper, we discuss the flow of tensors and various key operations in mixed precision training, and delve into details of operations, such as the rounding modes for converting FP32 tensors to BFLOAT16. We have implemented a method to emulate BFLOAT16 operations in Tensorflow, Caffe2, IntelCaffe, and Neon for our experiments. Our results show that deep learning training using BFLOAT16 tensors achieves the same state-of-the-art (SOTA) results across domains as FP32 tensors in the same number of iterations and with no changes to hyper-parameters. △ Less

Submitted 13 June, 2019; v1 submitted 29 May, 2019; originally announced May 2019.

ISA Mapper: A Compute and Hardware Agnostic Deep Learning Compiler

Authors: Matthew Sotoudeh, Anand Venkat, Michael Anderson, Evangelos Georganas, Alexander Heinecke, Jason Knight

Abstract: Domain specific accelerators present new challenges and opportunities for code generation onto novel instruction sets, communication fabrics, and memory architectures. In this paper we introduce an intermediate representation (IR) which enables both deep learning computational kernels and hardware capabilities to be described in the same IR. We then formulate and apply instruction mapping to det… ▽ More Domain specific accelerators present new challenges and opportunities for code generation onto novel instruction sets, communication fabrics, and memory architectures. In this paper we introduce an intermediate representation (IR) which enables both deep learning computational kernels and hardware capabilities to be described in the same IR. We then formulate and apply instruction mapping to determine the possible ways a computation can be performed on a hardware system. Next, our scheduler chooses a specific mapping and determines the data movement and computation order. In order to manage the large search space of mappings and schedules, we developed a flexible framework that allows heuristics, cost models, and potentially machine learning to facilitate this search problem. With this system, we demonstrate the automated extraction of matrix multiplication kernels out of recent deep learning kernels such as depthwise-separable convolution. In addition, we demonstrate two to five times better performance on DeepBench sized GEMMs and GRU RNN execution when compared to state-of-the-art (SOTA) implementations on new hardware and up to 85% of the performance for SOTA implementations on existing hardware. △ Less

Submitted 12 October, 2018; originally announced October 2018.

Extreme Scale De Novo Metagenome Assembly

Authors: Evangelos Georganas, Rob Egan, Steven Hofmeyr, Eugene Goltsman, Bill Arndt, Andrew Tritt, Aydin Buluc, Leonid Oliker, Katherine Yelick

Abstract: Metagenome assembly is the process of transforming a set of short, overlapping, and potentially erroneous DNA segments from environmental samples into the accurate representation of the underlying microbiomes's genomes. State-of-the-art tools require big shared memory machines and cannot handle contemporary metagenome datasets that exceed Terabytes in size. In this paper, we introduce the MetaHipM… ▽ More Metagenome assembly is the process of transforming a set of short, overlapping, and potentially erroneous DNA segments from environmental samples into the accurate representation of the underlying microbiomes's genomes. State-of-the-art tools require big shared memory machines and cannot handle contemporary metagenome datasets that exceed Terabytes in size. In this paper, we introduce the MetaHipMer pipeline, a high-quality and high-performance metagenome assembler that employs an iterative de Bruijn graph approach. MetaHipMer leverages a specialized scaffolding algorithm that produces long scaffolds and accommodates the idiosyncrasies of metagenomes. MetaHipMer is end-to-end parallelized using the Unified Parallel C language and therefore can run seamlessly on shared and distributed-memory systems. Experimental results show that MetaHipMer matches or outperforms the state-of-the-art tools in terms of accuracy. Moreover, MetaHipMer scales efficiently to large concurrencies and is able to assemble previously intractable grand challenge metagenomes. We demonstrate the unprecedented capability of MetaHipMer by computing the first full assembly of the Twitchell Wetlands dataset, consisting of 7.5 billion reads - size 2.6 TBytes. △ Less

Submitted 19 September, 2018; originally announced September 2018.

Comments: Accepted to SC18

Anatomy Of High-Performance Deep Learning Convolutions On SIMD Architectures

Authors: Evangelos Georganas, Sasikanth Avancha, Kunal Banerjee, Dhiraj Kalamkar, Greg Henry, Hans Pabst, Alexander Heinecke

Abstract: Convolution layers are prevalent in many classes of deep neural networks, including Convolutional Neural Networks (CNNs) which provide state-of-the-art results for tasks like image recognition, neural machine translation and speech recognition. The computationally expensive nature of a convolution operation has led to the proliferation of implementations including matrix-matrix multiplication form… ▽ More Convolution layers are prevalent in many classes of deep neural networks, including Convolutional Neural Networks (CNNs) which provide state-of-the-art results for tasks like image recognition, neural machine translation and speech recognition. The computationally expensive nature of a convolution operation has led to the proliferation of implementations including matrix-matrix multiplication formulation, and direct convolution primarily targeting GPUs. In this paper, we introduce direct convolution kernels for x86 architectures, in particular for Xeon and XeonPhi systems, which are implemented via a dynamic compilation approach. Our JIT-based implementation shows close to theoretical peak performance, depending on the setting and the CPU architecture at hand. We additionally demonstrate how these JIT-optimized kernels can be integrated into a lightweight multi-node graph execution model. This illustrates that single- and multi-node runs yield high efficiencies and high image-throughputs when executing state-of-the-art image recognition tasks on CPUs. △ Less

Submitted 20 August, 2018; v1 submitted 16 August, 2018; originally announced August 2018.

Comments: Accepted to SC18

Mixed Precision Training of Convolutional Neural Networks using Integer Operations

Authors: Dipankar Das, Naveen Mellempudi, Dheevatsa Mudigere, Dhiraj Kalamkar, Sasikanth Avancha, Kunal Banerjee, Srinivas Sridharan, Karthik Vaidyanathan, Bharat Kaul, Evangelos Georganas, Alexander Heinecke, Pradeep Dubey, Jesus Corbal, Nikita Shustrov, Roma Dubtsov, Evarist Fomenko, Vadim Pirogov

Abstract: The state-of-the-art (SOTA) for mixed precision training is dominated by variants of low precision floating point operations, and in particular, FP16 accumulating into FP32 Micikevicius et al. (2017). On the other hand, while a lot of research has also happened in the domain of low and mixed-precision Integer training, these works either present results for non-SOTA networks (for instance only Ale… ▽ More The state-of-the-art (SOTA) for mixed precision training is dominated by variants of low precision floating point operations, and in particular, FP16 accumulating into FP32 Micikevicius et al. (2017). On the other hand, while a lot of research has also happened in the domain of low and mixed-precision Integer training, these works either present results for non-SOTA networks (for instance only AlexNet for ImageNet-1K), or relatively small datasets (like CIFAR-10). In this work, we train state-of-the-art visual understanding neural networks on the ImageNet-1K dataset, with Integer operations on General Purpose (GP) hardware. In particular, we focus on Integer Fused-Multiply-and-Accumulate (FMA) operations which take two pairs of INT16 operands and accumulate results into an INT32 output.We propose a shared exponent representation of tensors and develop a Dynamic Fixed Point (DFP) scheme suitable for common neural network operations. The nuances of developing an efficient integer convolution kernel is examined, including methods to handle overflow of the INT32 accumulator. We implement CNN training for ResNet-50, GoogLeNet-v1, VGG-16 and AlexNet; and these networks achieve or exceed SOTA accuracy within the same number of iterations as their FP32 counterparts without any change in hyper-parameters and with a 1.8X improvement in end-to-end training throughput. To the best of our knowledge these results represent the first INT16 training results on GP hardware for ImageNet-1K dataset using SOTA CNNs and achieve highest reported accuracy using half-precision △ Less

Submitted 23 February, 2018; v1 submitted 3 February, 2018; originally announced February 2018.

Comments: Published as a conference paper at ICLR 2018

Extreme-Scale De Novo Genome Assembly

Authors: Evangelos Georganas, Steven Hofmeyr, Rob Egan, Aydin Buluc, Leonid Oliker, Daniel Rokhsar, Katherine Yelick

Abstract: De novo whole genome assembly reconstructs genomic sequence from short, overlapping, and potentially erroneous DNA segments and is one of the most important computations in modern genomics. This work presents HipMER, a high-quality end-to-end de novo assembler designed for extreme scale analysis, via efficient parallelization of the Meraculous code. Genome assembly software has many components, ea… ▽ More De novo whole genome assembly reconstructs genomic sequence from short, overlapping, and potentially erroneous DNA segments and is one of the most important computations in modern genomics. This work presents HipMER, a high-quality end-to-end de novo assembler designed for extreme scale analysis, via efficient parallelization of the Meraculous code. Genome assembly software has many components, each of which stresses different components of a computer system. This chapter explains the computational challenges involved in each step of the HipMer pipeline, the key distributed data structures, and communication costs in detail. We present performance results of assembling the human genome and the large hexaploid wheat genome on large supercomputers up to tens of thousands of cores. △ Less

Submitted 31 May, 2017; originally announced May 2017.

Comments: To appear as a chapter in Exascale Scientific Applications: Programming Approaches for Scalability, Performance, and Portability, Straatsma, Antypas, Williams (editors), CRC Press, 2017

Constructing Performance Models for Dense Linear Algebra Algorithms on Cray XE Systems

Authors: Jorge González-Domínguez, Evangelos Georganas, Yili Zheng, María J. Martín

Abstract: Hiding or minimizing the communication cost is key in order to obtain good performance on large-scale systems. While communication overlapping attempts to hide communications cost, 2.5D communication avoiding algorithms improve performance scalability by reducing the volume of data transfers at the cost of extra memory usage. Both approaches can be used together or separately and the best choice d… ▽ More Hiding or minimizing the communication cost is key in order to obtain good performance on large-scale systems. While communication overlapping attempts to hide communications cost, 2.5D communication avoiding algorithms improve performance scalability by reducing the volume of data transfers at the cost of extra memory usage. Both approaches can be used together or separately and the best choice depends on the machine, the algorithm and the problem size. Thus, the development of performance models is crucial to determine the best option for each scenario. In this paper, we present a methodology for constructing performance models for parallel numerical routines on Cray XE systems. Our models use portable benchmarks that measure computational cost and network characteristics, as well as performance degradation caused by simultaneous accesses to the network. We validate our methodology by constructing the performance models for the 2D and 2.5D approaches, with and without overlapping, of two matrix multiplication algorithms (Cannon's and SUMMA), triangular solve (TRSM) and Cholesky. We compare the estimations provided by these models with the experimental results using up to 24,576 cores of a Cray XE6 system and predict the performance of the algorithms on larger systems. Results prove that the estimations significantly improve when taking into account network contention. △ Less

Submitted 6 February, 2014; originally announced February 2014.