Searcharxiv⌕ Search

arXiv subjects

Onur Mutlu

Publications and source records attributed to Onur Mutlu.

At least 127 records · Page 7Linked to original sources
arXiv2024

GateKeeper-GPU: Fast and Accurate Pre-Alignment Filtering in Short Read Mapping

At the last step of short read mapping, the candidate locations of the reads on the reference genome are verified to compute their differences from the corresponding reference segments using sequence alignment algorithms. Calculating the similarities and differences between two sequences is still computationally expensive since approximate string matching techniques traditionally inherit dynamic programming algorithms with quadratic time and space complexity. We introduce GateKeeper-GPU, a fast and accurate pre-alignment filter that efficiently reduces the need for expensive sequence alignment. GateKeeper-GPU provides two main contributions: first, improving the filtering accuracy of GateKeeper (a lightweight pre-alignment filter), and second, exploiting the massive parallelism provided by the large number of GPU threads of modern GPUs to examine numerous sequence pairs rapidly and concurrently. By reducing the work, GateKeeper-GPU provides an acceleration of 2.9x to sequence alignment and up to 1.4x speedup to the end-to-end execution time of a comprehensive read mapper (mrFAST). GateKeeper-GPU is available at https://github.com/BilkentCompGen/GateKeeper-GPU.

q-bio.GN↗
arXiv2024

RUBICON: A Framework for Designing Efficient Deep Learning-Based Genomic Basecallers

Nanopore sequencing generates noisy electrical signals that need to be converted into a standard string of DNA nucleotide bases using a computational step called basecalling. The accuracy and speed of basecalling have critical implications for all later steps in genome analysis. Many researchers adopt complex deep learning-based models to perform basecalling without considering the compute demands of such models, which leads to slow, inefficient, and memory-hungry basecallers. Therefore, there is a need to reduce the computation and memory cost of basecalling while maintaining accuracy. Our goal is to develop a comprehensive framework for creating deep learning-based basecallers that provide high efficiency and performance. We introduce RUBICON, a framework to develop hardware-optimized basecallers. RUBICON consists of two novel machine-learning techniques that are specifically designed for basecalling. First, we introduce the first quantization-aware basecalling neural architecture search (QABAS) framework to specialize the basecalling neural network architecture for a given hardware acceleration platform while jointly exploring and finding the best bit-width precision for each neural network layer. Second, we develop SkipClip, the first technique to remove the skip connections present in modern basecallers to greatly reduce resource and storage requirements without any loss in basecalling accuracy. We demonstrate the benefits of RUBICON by developing RUBICALL, the first hardware-optimized basecaller that performs fast and accurate basecalling. Compared to the fastest state-of-the-art basecaller, RUBICALL provides a 3.96x speedup with 2.97% higher accuracy. We show that RUBICON helps researchers develop hardware-optimized basecallers that are superior to expert-designed models.

cs.AR↗
arXiv2024

Rethinking the Producer-Consumer Relationship in Modern DRAM-Based Systems

Generational improvements to commodity DRAM throughout half a century have long solidified its prevalence as main memory across the computing industry. However, overcoming today's DRAM technology scaling challenges requires new solutions driven by both DRAM producers and consumers. In this paper, we observe that the separation of concerns between producers and consumers specified by industry-wide DRAM standards is becoming a liability to progress in addressing scaling-related concerns. To understand the problem, we study four key directions for overcoming DRAM scaling challenges using system-memory cooperation: (i) improving memory access latencies; (ii) reducing DRAM refresh overheads; (iii) securely defending against the RowHammer vulnerability; and (iv) addressing worsening memory errors. We find that the single most important barrier to advancement in all four cases is the consumer's lack of insight into DRAM reliability. Based on an analysis of DRAM reliability testing, we recommend revising the separation of concerns to incorporate limited information transparency between producers and consumers. Finally, we propose adopting this revision in a two-step plan, starting with immediate information release through crowdsourcing and publication and culminating in widespread modifications to DRAM standards.

cs.AR↗
arXiv2024

Victima: Drastically Increasing Address Translation Reach by Leveraging Underutilized Cache Resources

Address translation is a performance bottleneck in data-intensive workloads due to large datasets and irregular access patterns that lead to frequent high-latency page table walks (PTWs). PTWs can be reduced by using (i) large hardware TLBs or (ii) large software-managed TLBs. Unfortunately, both solutions have significant drawbacks: increased access latency, power and area (for hardware TLBs), and costly memory accesses, the need for large contiguous memory blocks, and complex OS modifications (for software-managed TLBs). We present Victima, a new software-transparent mechanism that drastically increases the translation reach of the processor by leveraging the underutilized resources of the cache hierarchy. The key idea of Victima is to repurpose L2 cache blocks to store clusters of TLB entries, thereby providing an additional low-latency and high-capacity component that backs up the last-level TLB and thus reduces PTWs. Victima has two main components. First, a PTW cost predictor (PTW-CP) identifies costly-to-translate addresses based on the frequency and cost of the PTWs they lead to. Second, a TLB-aware cache replacement policy prioritizes keeping TLB entries in the cache hierarchy by considering (i) the translation pressure (e.g., last-level TLB miss rate) and (ii) the reuse characteristics of the TLB entries. Our evaluation results show that in native (virtualized) execution environments Victima improves average end-to-end application performance by 7.4% (28.7%) over the baseline four-level radix-tree-based page table design and by 6.2% (20.1%) over a state-of-the-art software-managed TLB, across 11 diverse data-intensive workloads. Victima (i) is effective in both native and virtualized environments, (ii) is completely transparent to application and system software, and (iii) incurs very small area and power overheads on a modern high-end CPU.

cs.AR↗
arXiv2023

GenPIP: In-Memory Acceleration of Genome Analysis via Tight Integration of Basecalling and Read Mapping

Nanopore sequencing is a widely-used high-throughput genome sequencing technology that can sequence long fragments of a genome into raw electrical signals at low cost. Nanopore sequencing requires two computationally-costly processing steps for accurate downstream genome analysis. The first step, basecalling, translates the raw electrical signals into nucleotide bases (i.e., A, C, G, T). The second step, read mapping, finds the correct location of a read in a reference genome. In existing genome analysis pipelines, basecalling and read mapping are executed separately. We observe in this work that such separate execution of the two most time-consuming steps inherently leads to (1) significant data movement and (2) redundant computations on the data, slowing down the genome analysis pipeline. This paper proposes GenPIP, an in-memory genome analysis accelerator that tightly integrates basecalling and read mapping. GenPIP improves the performance of the genome analysis pipeline with two key mechanisms: (1) in-memory fine-grained collaborative execution of the major genome analysis steps in parallel; (2) a new technique for early-rejection of low-quality and unmapped reads to timely stop the execution of genome analysis for such reads, reducing inefficient computation. Our experiments show that, for the execution of the genome analysis pipeline, GenPIP provides 41.6X (8.4X) speedup and 32.8X (20.8X) energy savings with negligible accuracy loss compared to the state-of-the-art software genome analysis tools executed on a state-of-the-art CPU (GPU). Compared to a design that combines state-of-the-art in-memory basecalling and read mapping accelerators, GenPIP provides 1.39X speedup and 1.37X energy savings.

cs.AR↗
arXiv2023

Ramulator 2.0: A Modern, Modular, and Extensible DRAM Simulator

We present Ramulator 2.0, a highly modular and extensible DRAM simulator that enables rapid and agile implementation and evaluation of design changes in the memory controller and DRAM to meet the increasing research effort in improving the performance, security, and reliability of memory systems. Ramulator 2.0 abstracts and models key components in a DRAM-based memory system and their interactions into shared interfaces and independent implementations. Doing so enables easy modification and extension of the modeled functions of the memory controller and DRAM in Ramulator 2.0. The DRAM specification syntax of Ramulator 2.0 is concise and human-readable, facilitating easy modifications and extensions. Ramulator 2.0 implements a library of reusable templated lambda functions to model the functionalities of DRAM commands to simplify the implementation of new DRAM standards, including DDR5, LPDDR5, HBM3, and GDDR6. We showcase Ramulator 2.0's modularity and extensibility by implementing and evaluating a wide variety of RowHammer mitigation techniques that require different memory controller design changes. These techniques are added modularly as separate implementations without changing any code in the baseline memory controller implementation. Ramulator 2.0 is rigorously validated and maintains a fast simulation speed compared to existing cycle-accurate DRAM simulators. Ramulator 2.0 is open-sourced under the permissive MIT license at https://github.com/CMU-SAFARI/ramulator2

cs.AR↗
arXiv2023

Swordfish: A Framework for Evaluating Deep Neural Network-based Basecalling using Computation-In-Memory with Non-Ideal Memristors

Basecalling, an essential step in many genome analysis studies, relies on large Deep Neural Networks (DNNs) to achieve high accuracy. Unfortunately, these DNNs are computationally slow and inefficient, leading to considerable delays and resource constraints in the sequence analysis process. A Computation-In-Memory (CIM) architecture using memristors can significantly accelerate the performance of DNNs. However, inherent device non-idealities and architectural limitations of such designs can greatly degrade the basecalling accuracy, which is critical for accurate genome analysis. To facilitate the adoption of memristor-based CIM designs for basecalling, it is important to (1) conduct a comprehensive analysis of potential CIM architectures and (2) develop effective strategies for mitigating the possible adverse effects of inherent device non-idealities and architectural limitations. This paper proposes Swordfish, a novel hardware/software co-design framework that can effectively address the two aforementioned issues. Swordfish incorporates seven circuit and device restrictions or non-idealities from characterized real memristor-based chips. Swordfish leverages various hardware/software co-design solutions to mitigate the basecalling accuracy loss due to such non-idealities. To demonstrate the effectiveness of Swordfish, we take Bonito, the state-of-the-art (i.e., accurate and fast), open-source basecaller as a case study. Our experimental results using Sword-fish show that a CIM architecture can realistically accelerate Bonito for a wide range of real datasets by an average of 25.7x, with an accuracy loss of 6.01%.

cs.AR↗
arXiv2023

MetaStore: High-Performance Metagenomic Analysis via In-Storage Computing

Metagenomics has led to significant advancements in many fields. Metagenomic analysis commonly involves the key tasks of determining the species present in a sample and their relative abundances. These tasks require searching large metagenomic databases containing information on different species' genomes. Metagenomic analysis suffers from significant data movement overhead due to moving large amounts of low-reuse data from the storage system to the rest of the system. In-storage processing can be a fundamental solution for reducing data movement overhead. However, designing an in-storage processing system for metagenomics is challenging because none of the existing approaches can be directly implemented in storage effectively due to the hardware limitations of modern SSDs. We propose MetaStore, the first in-storage processing system designed to significantly reduce the data movement overhead of end-to-end metagenomic analysis. MetaStore is enabled by our lightweight and cooperative design that effectively leverages and orchestrates processing inside and outside the storage system. Through our detailed analysis of the end-to-end metagenomic analysis pipeline and careful hardware/software co-design, we address in-storage processing challenges for metagenomics via specialized and efficient 1) task partitioning, 2) data/computation flow coordination, 3) storage technology-aware algorithmic optimizations, 4) light-weight in-storage accelerators, and 5) data mapping. Our evaluation shows that MetaStore outperforms the state-of-the-art performance- and accuracy-optimized software metagenomic tools by 2.7-37.2$\times$ and 6.9-100.2$\times$, respectively, while matching the accuracy of the accuracy-optimized tool. MetaStore achieves 1.5-5.1$\times$ speedup compared to the state-of-the-art metagenomic hardware-accelerated tool, while achieving significantly higher accuracy.

cs.AR↗
arXiv2023

Sibyl: Adaptive and Extensible Data Placement in Hybrid Storage Systems Using Online Reinforcement Learning

Hybrid storage systems (HSS) use multiple different storage devices to provide high and scalable storage capacity at high performance. Recent research proposes various techniques that aim to accurately identify performance-critical data to place it in a "best-fit" storage device. Unfortunately, most of these techniques are rigid, which (1) limits their adaptivity to perform well for a wide range of workloads and storage device configurations, and (2) makes it difficult for designers to extend these techniques to different storage system configurations (e.g., with a different number or different types of storage devices) than the configuration they are designed for. We introduce Sibyl, the first technique that uses reinforcement learning for data placement in hybrid storage systems. Sibyl observes different features of the running workload as well as the storage devices to make system-aware data placement decisions. For every decision it makes, Sibyl receives a reward from the system that it uses to evaluate the long-term performance impact of its decision and continuously optimizes its data placement policy online. We implement Sibyl on real systems with various HSS configurations. Our results show that Sibyl provides 21.6%/19.9% performance improvement in a performance-oriented/cost-oriented HSS configuration compared to the best previous data placement technique. Our evaluation using an HSS configuration with three different storage devices shows that Sibyl outperforms the state-of-the-art data placement policy by 23.9%-48.2%, while significantly reducing the system architect's burden in designing a data placement mechanism that can simultaneously incorporate three storage devices. We show that Sibyl achieves 80% of the performance of an oracle policy that has complete knowledge of future access patterns while incurring a very modest storage overhead of only 124.4 KiB.

cs.AR↗
arXiv2023

An In-Memory Architecture for High-Performance Long-Read Pre-Alignment Filtering

With the recent move towards sequencing of accurate long reads, finding solutions that support efficient analysis of these reads becomes more necessary. The long execution time required for sequence alignment of long reads negatively affects genomic studies relying on sequence alignment. Although pre-alignment filtering as an extra step before alignment was recently introduced to mitigate sequence alignment for short reads, these filters do not work as efficiently for long reads. Moreover, even with efficient pre-alignment filters, the overall end-to-end (i.e., filtering + original alignment) execution time of alignment for long reads remains high, while the filtering step is now a major portion of the end-to-end execution time. Our paper makes three contributions. First, it identifies data movement of sequences between memory units and computing units as the main source of inefficiency for pre-alignment filters of long reads. This is because although filters reject many of these long sequencing pairs before they get to the alignment stage, they still require a huge cost regarding time and energy consumption for the large data transferred between memory and processor. Second, this paper introduces an adaptation of a short-read pre-alignment filtering algorithm suitable for long reads. We call this LongGeneGuardian. Finally, it presents Filter-Fuse as an architecture that supports LongGeneGuardian inside the memory. FilterFuse exploits the Computation-In-Memory computing paradigm, eliminating the cost of data movement in LongGeneGuardian. Our evaluations show that FilterFuse improves the execution time of filtering by 120.47x for long reads compared to State-of-the-Art (SoTA) filter, SneakySnake. FilterFuse also improves the end-to-end execution time of sequence alignment by up to 49.14x and 5207.63x compared to SneakySnake with SoTA aligner and only SoTA aligner, respectively.

cs.AR↗
arXiv2023

ApHMM: Accelerating Profile Hidden Markov Models for Fast and Energy-Efficient Genome Analysis

Profile hidden Markov models (pHMMs) are widely employed in various bioinformatics applications to identify similarities between biological sequences, such as DNA or protein sequences. In pHMMs, sequences are represented as graph structures. These probabilities are subsequently used to compute the similarity score between a sequence and a pHMM graph. The Baum-Welch algorithm, a prevalent and highly accurate method, utilizes these probabilities to optimize and compute similarity scores. However, the Baum-Welch algorithm is computationally intensive, and existing solutions offer either software-only or hardware-only approaches with fixed pHMM designs. We identify an urgent need for a flexible, high-performance, and energy-efficient HW/SW co-design to address the major inefficiencies in the Baum-Welch algorithm for pHMMs. We introduce ApHMM, the first flexible acceleration framework designed to significantly reduce both computational and energy overheads associated with the Baum-Welch algorithm for pHMMs. ApHMM tackles the major inefficiencies in the Baum-Welch algorithm by 1) designing flexible hardware to accommodate various pHMM designs, 2) exploiting predictable data dependency patterns through on-chip memory with memoization techniques, 3) rapidly filtering out negligible computations using a hardware-based filter, and 4) minimizing redundant computations. ApHMM achieves substantial speedups of 15.55x - 260.03x, 1.83x - 5.34x, and 27.97x when compared to CPU, GPU, and FPGA implementations of the Baum-Welch algorithm, respectively. ApHMM outperforms state-of-the-art CPU implementations in three key bioinformatics applications: 1) error correction, 2) protein family search, and 3) multiple sequence alignment, by 1.29x - 59.94x, 1.03x - 1.75x, and 1.03x - 1.95x, respectively, while improving their energy efficiency by 64.24x - 115.46x, 1.75x, 1.96x.

cs.AR↗
arXiv2023

Utopia: Fast and Efficient Address Translation via Hybrid Restrictive & Flexible Virtual-to-Physical Address Mappings

Conventional virtual memory (VM) frameworks enable a virtual address to flexibly map to any physical address. This flexibility necessitates large data structures to store virtual-to-physical mappings, which leads to high address translation latency and large translation-induced interference in the memory hierarchy. On the other hand, restricting the address mapping so that a virtual address can only map to a specific set of physical addresses can significantly reduce address translation overheads by using compact and efficient translation structures. However, restricting the address mapping flexibility across the entire main memory severely limits data sharing across different processes and increases data accesses to the swap space of the storage device, even in the presence of free memory. We propose Utopia, a new hybrid virtual-to-physical address mapping scheme that allows both flexible and restrictive hash-based address mapping schemes to harmoniously co-exist in the system. The key idea of Utopia is to manage physical memory using two types of physical memory segments: restrictive and flexible segments. A restrictive segment uses a restrictive, hash-based address mapping scheme that maps virtual addresses to only a specific set of physical addresses and enables faster address translation using compact translation structures. A flexible segment employs the conventional fully-flexible address mapping scheme. By mapping data to a restrictive segment, Utopia enables faster address translation with lower translation-induced interference. Utopia improves performance by 24% in a single-core system over the baseline system, whereas the best prior state-of-the-art contiguity-aware translation scheme improves performance by 13%.

cs.AR↗
arXiv2023

Evaluating Homomorphic Operations on a Real-World Processing-In-Memory System

Computing on encrypted data is a promising approach to reduce data security and privacy risks, with homomorphic encryption serving as a facilitator in achieving this goal. In this work, we accelerate homomorphic operations using the Processing-in- Memory (PIM) paradigm to mitigate the large memory capacity and frequent data movement requirements. Using a real-world PIM system, we accelerate the Brakerski-Fan-Vercauteren (BFV) scheme for homomorphic addition and multiplication. We evaluate the PIM implementations of these homomorphic operations with statistical workloads (arithmetic mean, variance, linear regression) and compare to CPU and GPU implementations. Our results demonstrate 50-100x speedup with a real PIM system (UPMEM) over the CPU and 2-15x over the GPU in vector addition. For vector multiplication, the real PIM system outperforms the CPU by 40-50x. However, it lags 10-15x behind the GPU due to the lack of native sufficiently wide multiplication support in the evaluated first-generation real PIM system. For mean, variance, and linear regression, the real PIM system performance improvements vary between 30x and 300x over the CPU and between 10x and 30x over the GPU, uncovering real PIM system trade-offs in terms of scalability of homomorphic operations for varying amounts of data. We plan to make our implementation open-source in the future.

cs.CR↗
arXiv2023

SimplePIM: A Software Framework for Productive and Efficient Processing-in-Memory

Data movement between memory and processors is a major bottleneck in modern computing systems. The processing-in-memory (PIM) paradigm aims to alleviate this bottleneck by performing computation inside memory chips. Real PIM hardware (e.g., the UPMEM system) is now available and has demonstrated potential in many applications. However, programming such real PIM hardware remains a challenge for many programmers. This paper presents a new software framework, SimplePIM, to aid programming real PIM systems. The framework processes arrays of arbitrary elements on a PIM device by calling iterator functions from the host and provides primitives for communication among PIM cores and between PIM and the host system. We implement SimplePIM for the UPMEM PIM system and evaluate it on six major applications. Our results show that SimplePIM enables 66.5% to 83.1% reduction in lines of code in PIM programs. The resulting code leads to higher performance (between 10% and 37% speedup) than hand-optimized code in three applications and provides comparable performance in three others. SimplePIM is fully and freely available at https://github.com/CMU-SAFARI/SimplePIM.

cs.AR↗
arXiv2023

GateSeeder: Near-memory CPU-FPGA Acceleration of Short and Long Read Mapping

Motivation: Read mapping is a computationally expensive process and a major bottleneck in genomics analyses. The performance of read mapping is mainly limited by the performance of three key computational steps: Index Querying, Seed Chaining, and Sequence Alignment. The first step is dominated by how fast and frequent it accesses the main memory (i.e., memory-bound), while the latter two steps are dominated by how fast the CPU can compute their computationally-costly dynamic programming algorithms (i.e., compute-bound). Accelerating these three steps by exploiting new algorithms and new hardware devices is essential to accelerate most genome analysis pipelines that widely use read mapping. Given the large body of work on accelerating Sequence Alignment, this work focuses on significantly improving the remaining steps. Results: We introduce GateSeeder, the first CPU-FPGA-based near-memory acceleration of both short and long read mapping. GateSeeder exploits near-memory computation capability provided by modern FPGAs that couple a reconfigurable compute fabric with high-bandwidth memory (HBM) to overcome the memory-bound and compute-bound bottlenecks. GateSeeder also introduces a new lightweight algorithm for finding the potential matching segment pairs. Using real ONT, HiFi, and Illumina sequences, we experimentally demonstrate that GateSeeder outperforms Minimap2, without performing sequence alignment, by up to 40.3x, 4.8x, and 2.3x, respectively. When performing read mapping with sequence alignment, GateSeeder outperforms Minimap2 by 1.15-4.33x (using KSW2) and by 1.97-13.63x (using WFA-GPU). Availability: https://github.com/CMU-SAFARI/GateSeeder

cs.AR↗
arXiv2023

Extending Memory Capacity in Consumer Devices with Emerging Non-Volatile Memory: An Experimental Study

The number and diversity of consumer devices are growing rapidly, alongside their target applications' memory consumption. Unfortunately, DRAM scalability is becoming a limiting factor to the available memory capacity in consumer devices. As a potential solution, manufacturers have introduced emerging non-volatile memories (NVMs) into the market, which can be used to increase the memory capacity of consumer devices by augmenting or replacing DRAM. Since entirely replacing DRAM with NVM in consumer devices imposes large system integration and design challenges, recent works propose extending the total main memory space available to applications by using NVM as swap space for DRAM. However, no prior work analyzes the implications of enabling a real NVM-based swap space in real consumer devices. In this work, we provide the first analysis of the impact of extending the main memory space of consumer devices using off-the-shelf NVMs. We extensively examine system performance and energy consumption when the NVM device is used as swap space for DRAM main memory to effectively extend the main memory capacity. For our analyses, we equip real web-based Chromebook computers with the Intel Optane SSD, which is a state-of-the-art low-latency NVM-based SSD device. We compare the performance and energy consumption of interactive workloads running on our Chromebook with NVM-based swap space, where the Intel Optane SSD capacity is used as swap space to extend main memory capacity, against two state-of-the-art systems: (i) a baseline system with double the amount of DRAM than the system with the NVM-based swap space; and (ii) a system where the Intel Optane SSD is naively replaced with a state-of-the-art (yet slower) off-the-shelf NAND-flash-based SSD, which we use as a swap space of equivalent size as the NVM-based swap space.

cs.AR↗
arXiv2023

Memory-Centric Computing

Memory-centric computing aims to enable computation capability in and near all places where data is generated and stored. As such, it can greatly reduce the large negative performance and energy impact of data access and data movement, by fundamentally avoiding data movement and reducing data access latency & energy. Many recent studies show that memory-centric computing can greatly improve system performance and energy efficiency. Major industrial vendors and startup companies have also recently introduced memory chips that have sophisticated computation capabilities. This talk describes promising ongoing research and development efforts in memory-centric computing. We classify such efforts into two major fundamental categories: 1) processing using memory, which exploits analog operational properties of memory structures to perform massively-parallel operations in memory, and 2) processing near memory, which integrates processing capability in memory controllers, the logic layer of 3D-stacked memory technologies, or memory chips to enable high-bandwidth and low-latency memory access to near-memory logic. We show both types of architectures (and their combination) can enable orders of magnitude improvements in performance and energy consumption of many important workloads, such as graph analytics, databases, machine learning, video processing, climate modeling, genome analysis. We discuss adoption challenges for the memory-centric computing paradigm and conclude with some research & development opportunities.

cs.AR↗
arXiv2023

Casper: Accelerating Stencil Computation using Near-cache Processing

Stencil computation is one of the most used kernels in a wide variety of scientific applications, ranging from large-scale weather prediction to solving partial differential equations. Stencil computations are characterized by three unique properties: (1) low arithmetic intensity, (2) limited temporal data reuse, and (3) regular and predictable data access pattern. As a result, stencil computations are typically bandwidth-bound workloads, which only experience limited benefits from the deep cache hierarchy of modern CPUs. In this work, we propose Casper, a near-cache accelerator consisting of specialized stencil compute units connected to the last-level cache (LLC) of a traditional CPU. Casper is based on two key ideas: (1) avoiding the cost of moving rarely reused data through the cache hierarchy, and (2) exploiting the regularity of the data accesses and the inherent parallelism of the stencil computation to increase the overall performance. With minimal changes in LLC address decoding logic and data placement, Casper performs stencil computations at the peak bandwidth of the LLC. We show that, by tightly coupling lightweight stencil compute units near to LLC, Casper improves the performance of stencil kernels by 1.65x on average, while reducing the energy consumption by 35% compared to a commercial high-performance multi-core processor. Moreover, Casper provides a 37x improvement in performance-per-area compared to a state-of-the-art GPU.

cs.AR↗