Jangwoo Kim

Multi-bit Error Tolerant Caches Using Two-Dimensional Error Coding

In deep sub-micron ICs, growing amounts of on-die memory and scaling effects make embedded memories increasingly vulnerable to reliability and yield problems. As scaling progresses, soft and hard errors in the memory system will increase and single error events are more likely to cause large-scale multi- bit errors. However, conventional memory protection techniques can neither detect nor correct large-scale multi-bit errors without incurring large performance, area, and power overheads. We propose two-dimensional (2D) error coding in embedded memories, a scalable multi-bit error protection technique to improve memory reliability and yield. The key innovation is the use of vertical error coding across words that is used only for error correction in combination with conventional per-word horizontal error coding. We evaluate this scheme in the cache hierarchies of two representative chip multiprocessor designs and show that 2D error coding can correct clustered errors up to 32times32 bits with significantly smaller performance, area, and power overheads than conventional techniques.

SimFlex: a fast, accurate, flexible full-system simulation framework for performance evaluation of server architecture

The new focus on commercial workloads in simulation studies of server systems has caused a drastic increase in the complexity and decrease in the speed of simulation tools. The complexity of a large-scale full-system model makes development of a monolithic simulation tool a prohibitively difficult task. Furthermore, detailed full-system models simulate so slowly that experimental results must be based on simulations of only fractions of a second of execution of the modelled system.This paper presents SIMFLEX, a simulation framework which uses component-based design and rigorous statistical sampling to enable development of complex models and ensure representative measurement results with fast simulation turnaround. The novelty of SIMFLEX lies in its combination of a unique, compile-time approach to component interconnection and a methodology for obtaining accurate results from sampled simulations on a platform capable of evaluating unmodified commercial workloads.

Temporal Streaming of Shared Memory

Coherent read misses in shared-memory multiprocessors account for a substantial fraction of execution time in many important scientific and commercial workloads. We propose temporal streaming, to eliminate coherent read misses by streaming data to a processor in advance of the corresponding memory accesses. Temporal streaming dynamically identifies address sequences to be streamed by exploiting two common phenomena in shared-memory access patterns: (1) temporal address correlation-groups of shared addresses tend to be accessed together and in the same order; and (2) temporal stream locality-recently-accessed address streams are likely to recur. We present a practical design for temporal streaming. We evaluate our design using a combination of trace-driven and cycle-accurate full-system simulation of a cache-coherent distributed shared-memory system. We show that temporal streaming can eliminate 98% of coherent read misses in scientific applications, and between 43% and 60% in database and Web server workloads. Our design yields speedups of 1.07 to 3.29 in scientific applications, and 1.06 to 1.21 in commercial workloads.

Memory coherence activity prediction in commercial workloads

Recent research indicates that prediction-based coherence optimizations offer substantial performance improvements for scientific applications in distributed shared memory multiprocessors. Important commercial applications also show sensitivity to coherence latency, which will become more acute in the future as technology scales. Therefore it is important to investigate prediction of memory coherence activity in the context of commercial workloads.This paper studies a trace-based Downgrade Predictor (DGP) for predicting last stores to shared cache blocks, and a pattern-based Consumer Set Predictor (CSP) for predicting subsequent readers. We evaluate this class of predictors for the first time on commercial applications and demonstrate that our DGP correctly predicts 47%-76% of last stores. Memory sharing patterns in commercial workloads are inherently non-repetitive; hence CSP cannot attain high coverage. We perform an opportunity study of a DGP enhanced through competitive underlying predictors, and in commercial and scientific applications, demonstrate potential to increase coverage up to 14%.

Stealing Webpages Rendered on Your Browser by Exploiting GPU Vulnerabilities

Graphics processing units (GPUs) are important components of modern computing devices for not only graphics rendering, but also efficient parallel computations. However, their security problems are ignored despite their importance and popularity. In this paper, we first perform an in-depth security analysis on GPUs to detect security vulnerabilities. We observe that contemporary, widely-used GPUs, both NVIDIAs and AMDs, do not initialize newly allocated GPU memory pages which may contain sensitive user data. By exploiting such vulnerabilities, we propose attack methods for revealing a victim programs data kept in GPU memory both during its execution and right after its termination. We further show the high applicability of the proposed attacks by applying them to the Chromium and Firefox web browsers which use GPUs for accelerating webpage rendering. We detect that both browsers leave rendered webpage textures in GPU memory, so that we can infer which web pages a victim user has visited by analyzing the remaining textures. The accuracy of our advanced inference attack that uses both pixel sequence matching and RGB histogram matching is up to 95.4%.

A fully associative, tagless DRAM cache

This paper introduces a tagless cache architecture for large in-package DRAM caches. The conventional die-stacked DRAM cache has both a TLB and a cache tag array, which are responsible for virtual-to-physical and physical-to-cache address translation, respectively. We propose to align the granularity of caching with OS page size and take a unified approach to address translation and cache tag management. To this end, we introduce cache-map TLB (cTLB), which stores virtual-to-cache, instead of virtual-to-physical, address mappings. At a TLB miss, the TLB miss handler allocates the requested block into the cache if it is not cached yet, and updates both the page table and cTLB with the virtual-to-cache address mapping. Assuming the availability of large in-package DRAM caches, this ensures that an access to the memory region within the TLB reach always hits in the cache with low hit latency since a TLB access immediately returns the exact location of the requested block in the cache, hence saving a tag-checking operation. The remaining cache space is used as victim cache for memory pages that are recently evicted from cTLB. By completely eliminating data structures for cache tag management, from either on-die SRAM or inpackage DRAM, the proposed DRAM cache achieves best scalability and hit latency, while maintaining high hit rate of a fully associative cache. Our evaluation with 3D Through-Silicon Via (TSV)-based in-package DRAM demonstrates that the proposed cache improves the IPC and energy efficiency by 30.9% and 39.5%, respectively, compared to the baseline with no DRAM cache. These numbers translate to 4.3% and 23.8% improvements over an impractical SRAM-tag cache requiring megabytes of on-die SRAM storage, due to low hit latency and zero energy waste for cache tags.

TRUSS: a reliable, scalable server architecture

Traditional techniques that mainframes use to increase reliability -special hardware or custom software - are incompatible with commodity server requirements. The Total Reliability Using Scalable Servers (TRUSS) architecture, developed at Carnegie Mellon, aims to bring reliability to commodity servers. TRUSS features a distributed shared-memory (DSM) multiprocessor that incorporates computation and memory storage redundancy to detect and recover from any single point of transient or permanent failure. Because its underlying DSM architecture presents the familiar shared-memory programming model, TRUSS requires no changes to existing applications and only minor modifications to the operating system to support error recovery.

Store-ordered streaming of shared memory

Coherence misses in shared-memory multiprocessors account for a substantial fraction of execution time in many important scientific and commercial workloads. Memory streaming provides a promising solution to the coherence miss bottleneck because it improves memory level parallelism and lookahead while using on-chip resources efficiently. We observe that the order in which shared data are consumed by one processor is correlated to the order in which they were produced by another. We investigate this phenomenon and demonstrate that it can be exploited to send store-ordered streams (SORDS) of shared data from producers to consumers, thereby eliminating coherent read misses. Using a trace-driven analysis of all user and OS memory references in a cache-coherent distributed shared-memory multiprocessor, we show that SORDS-based memory streaming can eliminate between 36% and 100% of all coherent read misses in scientific workloads and between 23% and 48% in online transaction processing workloads.

Microbank: architecting through-silicon interposer-based main memory systems

Through-Silicon Interposer (TSI) has recently been proposed to provide high memory bandwidth and improve energy efficiency of the main memory system. However, the impact of TSI on main memory system architecture has not been well explored. While TSI improves the I/O energy efficiency, we show that it results in an unbalanced memory system design in terms of energy efficiency as the core DRAM dominates overall energy consumption. To balance and enhance the energy efficiency of a TSI-based memory system, we propose μbank, a novel DRAM device organization in which each bank is partitioned into multiple smaller banks (or μbanks) that operate independently like conventional banks with minimal area overhead. The μbank organization significantly increases the amount of bank-level parallelism to improve the performance and energy efficiency of the TSI-based memory system. The massive number of μbanks reduces bank conflicts, hence simplifying the memory system design. We evaluated a sophisticated prediction-based DRAM page-management policy, which can improve performance by up to 20.5% in a conventional memory system without μbanks. However, a μbank-based design does not require such a complex page-management policy and a simple open-page policy is often sufficient -- achieving within 5% of a perfect predictor. Our proposed μbank-based memory system improves the IPC and system energy-delay product by 1.62× and 4.80×, respectively, for memory-intensive SPEC 2006 benchmarks on average, over the baseline DDR3-based memory system.

GPUdmm: A high-performance and memory-oblivious GPU architecture using dynamic memory management

GPU programmers suffer from programmer-managed GPU memory because both performance and programmability heavily depend on GPU memory allocation and CPU-GPU data transfer mechanisms. To improve performance and programmability, programmers should be able to place only the data frequently accessed by GPU on GPU memory while overlapping CPU-GPU data transfers and GPU executions as much as possible. However, current GPU architectures and programming models blindly place entire data on GPU memory, requiring a significantly large GPU memory size. Otherwise, they must trigger unnecessary CPU-GPU data transfers due to an insufficient GPU memory size. In this paper, we propose GPUdmm, a novel GPU architecture to enable high-performance and memory-oblivious GPU programming. First, GPUdmm uses GPU memory as a cache of CPU memory to provide programmers a view of the CPU memory-sized programming space. Second, GPUdmm achieves high performance by exploiting data locality and dynamically transferring data between CPU and GPU memories while effectively overlapping CPU-GPU data transfers and GPU executions. Third, GPUdmm can further reduce unnecessary CPU-GPU data transfers by exploiting simple programmer hints. Our carefully designed and validated experiments (e.g., PCIe/DMA timing) against representative benchmarks show that GPUdmm can achieve up to five times higher performance for the same GPU memory size, or reduce the GPU memory size requirement by up to 75% while maintaining the same performance.