Mark Gottscho

Power Variability in Contemporary DRAMs

Technology scaling has led to significant variability in chip performance and power consumption. In this work, we measured and analyzed the power variability in dynamic random access memories (DRAMs). We tested 22 double date rate third generation (DDR3) dual inline memory modules (DIMMs), and found that power usage in DRAMs depends on both operation type (write, read, and idle) as well as data, with write operations consuming more than reads, and 1s in the data generally costing more power than 0s. Temperature had little effect (1-3%) across the C to 50 C range. Variations were up to 12.29% and 16.40% for idle power within a single model and for different models from the same vendor, respectively. In the scope of all tested 1 gigabyte (GB) modules, deviations were up to 21.84% in write power. Our ongoing work addresses memory management methods to leverage such power variations.

Power / Capacity Scaling: Energy Savings With Simple Fault-Tolerant Caches

Complicated approaches to fault-tolerant voltage-scalable (FTVS) SRAM cache architectures can suffer from high overheads. We propose static (SPCS) and dynamic (DPCS) variants of power/capacity scaling, a simple and low-overhead fault-tolerant cache architecture that utilizes insights gained from our 45nm SOI test chip. Our mechanism combines multi-level voltage scaling with power gating of blocks that become faulty at each voltage level. The SPCS policy sets the runtime cache VDD statically such that almost all of the cache blocks are not faulty. The DPCS policy opportunistically reduces the voltage further to save more power than SPCS while limiting the impact on performance caused by additional faulty blocks. Through an analytical evaluation, we show that our approach can achieve lower static power for all effective cache capacities than a recent complex FTVS work. This is due to significantly lower overheads, despite the failure of our approach to match the min-VDD of the competing work at fixed yield. Through architectural simulations, we find that the average energy saved by SPCS is 55%, while DPCS saves an average of 69% of energy with respect to baseline caches at 1 V. Our approach incurs no more than 4% performance and 5% area penalties in the worst case cache configuration.

ViPZonE: OS-level memory variability-driven physical address zoning for energy savings

ITRS predicts that over the next decade, hardware power variation will increase at alarming rates. As a result, designers must build software that can adapt to and exploit these variations to reduce power consumption and improve system performance. This paper presents ViPZonE, a system-level solution that opportunistically exploits DRAM power variation through physical address zoning. ViPZonE is composed of a variability-aware software stack that allows developers to indicate to the OS the expected dominant usage patterns (write or read) as well as level of utilization (high, medium, or low) through high-level APIs. ViPZonEs variability-aware page allocator, implemented in the Linux kernel, is responsible for interpreting these high-level requests for memory and transparently mapping them to physical address zones with different power consumption. Our experimental results across various configurations running PAR-SEC workloads show an average of 13.1% memory power consumption savings at the cost of a modest 1.03% increase in execution time over a typical Linux virtual memory allocator.

Multi-Layer Memory Resiliency

With memories continuing to dominate the area, power, cost and performance of a design, there is a critical need to provision reliable, high-performance memory bandwidth for emerging applications. Memories are susceptible to degradation and failures from a wide range of manufacturing, operational and environmental effects, requiring a multi-layer hardware/software approach that can tolerate, adapt and even opportunistically exploit such effects. The overall memory hierarchy is also highly vulnerable to the adverse effects of variability and operational stress. After reviewing the major memory degradation and failure modes, this paper describes the challenges for dependability across the memory hierarchy, and outlines research efforts to achieve multi-layer memory resilience using a hardware/software approach. Two specific exemplars are used to illustrate multi-layer memory resilience: first we describe static and dynamic policies to achieve energy savings in caches using aggressive voltage scaling combined with disabling faulty blocks; and second we show how software characteristics can be exposed to the architecture in order to mitigate the aging of large register files in GPGPUs. These approaches can further benefit from semantic retention of application intent to enhance memory dependability across multiple abstraction levels, including applications, compilers, run-time systems, and hardware platforms.

DPCS: Dynamic Power/Capacity Scaling for SRAM Caches in the Nanoscale Era

Fault-Tolerant Voltage-Scalable (FTVS) SRAM cache architectures are a promising approach to improve energy efficiency of memories in the presence of nanoscale process variation. Complex FTVS schemes are commonly proposed to achieve very low minimum supply voltages, but these can suffer from high overheads and thus do not always offer the best power/capacity trade-offs. We observe on our 45nm test chips that the “fault inclusion property” can enable lightweight fault maps that support multiple runtime supply voltages. Based on this observation, we propose a simple and low-overhead FTVS cache architecture for power/capacity scaling. Our mechanism combines multilevel voltage scaling with optional architectural support for power gating of blocks as they become faulty at low voltages. A static (SPCS) policy sets the runtime cache VDD once such that a only a few cache blocks may be faulty in order to minimize the impact on performance. We describe a Static Power/Capacity Scaling (SPCS) policy and two alternate Dynamic Power/Capacity Scaling (DPCS) policies that opportunistically reduce the cache voltage even further for more energy savings. This architecture achieves lower static power for all effective cache capacities than a recent more complex FTVS scheme. This is due to significantly lower overheads, despite the inability of our approach to match the min-VDD of the competing work at a fixed target yield. Over a set of SPEC CPU2006 benchmarks on two system configurations, the average total cache (system) energy saved by SPCS is 62% (22%), while the two DPCS policies achieve roughly similar energy reduction, around 79% (26%). On average, the DPCS approaches incur 2.24% performance and 6% area penalties.

Variability-aware memory management for nanoscale computing

As the semiconductor industry continues to push the limits of sub-micron technology, the ITRS expects hardware (e.g., die-to-die, wafer-to-wafer, and chip-to-chip) variations to continue increasing over the next few decades. As a result, it is imperative for designers to build variation-aware software stacks that may adapt and opportunistically exploit said variations to increase system performance/responsiveness as well as minimize power consumption. The memory subsystem is one of the largest components in todays computing system, a main contributor to the overall power consumption of the system, and therefore one of the most vulnerable components to the effects of variations (e.g., power). This paper discusses the concept of variability-aware memory management for nanoscale computing systems. We show how to opportunistically exploit the hardware variations in on-chip and off-chip memory at the system level through the deployment of variation-aware software stacks.

X-Mem: A cross-platform and extensible memory characterization tool for the cloud

Effective use of the memory hierarchy is crucial to cloud computing. Platform memory subsystems must be carefully provisioned and configured to minimize overall cost and energy for cloud providers. For cloud subscribers, the diversity of available platforms complicates comparisons and the optimization of performance. To address these needs, we present X-Mem, a new open-source software tool that characterizes the memory hierarchy for cloud computing.

NSF expedition on variability-aware software: Recent results and contributions

Abstract In this paper we summarize recent results and contributions from the NSF Expedition on Variability-Aware Software, a five year, multi-university effort to tackle the problem of hardware variations and its implications and opportunities in software. The Expedition has made contributions in characterization and online monitoring of variations (particularly in microprocessors and flash memories), proposed new coding techniques for variability-tolerant storage, provided tools and platforms for the development of variability-aware software, and created new runtime support systems for variability-aware task-scheduling and execution.

ViPZonE: Hardware Power Variability-Aware Virtual Memory Management for Energy Savings

Hardware variability is predicted to increase dramatically over the coming years as a consequence of continued technology scaling. In this paper, we apply the Underdesigned and Opportunistic Computing (UnO) paradigm by exposing system-level power variability to software to improve energy efficiency. We present ViPZonE, a memory management solution in conjunction with application annotations that opportunistically performs memory allocations to reduce DRAM energy. ViPZonEs components consist of a physical address space with DIMM-aware zones, a modified page allocation routine, and a new virtual memory system call for dynamic allocations from userspace. We implemented ViPZonE in the Linux kernel with GLIBC API support, running on a real x86-64 testbed with significant access power variation in its DDR3 DIMMs. We demonstrate that on our testbed, ViPZonE can save up to 27.80 percent memory energy, with no more than 4.80 percent performance degradation across a set of PARSEC benchmarks tested with respect to the baseline Linux software. Furthermore, through a hypothetical “what-if” extension, we predict that in future non-volatile memory systems which consume almost no idle power, ViPZonE could yield even greater benefits, demonstrating the ability to exploit memory hardware variability through opportunistic software.

Software-Defined Error-Correcting Codes

Conventional error-correcting codes (ECCs) and system-level fault-tolerance mechanisms are currently treated as separate abstraction layers. This can reduce the overall efficacy of error detection and correction (EDAC) capabilities, impacting the reliability of memories by causing crashes or silent data corruption. To address this shortcoming, we propose Software-Defined ECC (SWD-ECC), a new class of heuristic techniques to recover from detected but uncorrectable errors (DUEs) in memory. It uses available side information to estimate the original message by first filtering and then ranking the possible candidate codewords for a DUE. SWD-ECC does not incur any hardware or software overheads in the cases where DUEs do not occur.As an exemplar for SWD-ECC, we show through offline analysis on SPEC CPU2006 benchmarks how to heuristically recover from 2-bit DUEs in MIPS instruction memory using a common (39,32) single-error-correcting, double-error-detecting (SECDED) code. We first apply coding theory to compute all of the candidate codewords for a given DUE. Second, we filter out the candidates that are not legal MIPS instructions, increasing the chance of successful recovery. Finally, we choose a valid candidate whose logical operation (e.g., add or load) occurs most frequently in the application binary image. Our results show that on average, 34% of all possible 2-bit DUEs in the evaluated set of instructions can be successfully recovered using this heuristic recovery strategy. If a DUE affects the bit fields used for instruction decoding, we are able to recover correctly up to 99% of the time. We believe these results to be a significant achievement compared to an otherwise-guaranteed crash which can be undesirable in many systems and applications. Moreover, there is room for future improvement of this result with more sophisticated uses of side information. We look forward to future work in this area.