We present Sequoia, a programming language designed to facilitate the development of memory hierarchy aware parallel programs that remain portable across modern machines featuring different memory hierarchy configurations. Sequoia abstractly exposes hierarchical memory in the programming model and provides language mechanisms to describe communication vertically through the machine and to localize computation to particular memory locations within it. We have implemented a complete programming system, including a compiler and runtime systems for Cell processor-based blade systems and distributed memory clusters, and demonstrate efficient performance running Sequoia programs on both of these platforms.

No abstract

Wide SIMD-based GPUs have evolved into a promising platform for running general purpose workloads. Current programmable GPUs allow even code with irregular control to execute well on their SIMD pipelines. To do this, each SIMD lane is considered to execute a logical thread where hardware ensures that control flow is accurate by automatically applying masked execution. The masked execution, however, often degrades performance because the issue slots of masked lanes are wasted. This degradation can be mitigated by dynamically compacting multiple unmasked threads into a single SIMD unit. This paper proposes a fundamentally new approach to branch compaction that avoids the unnecessary synchronization required by previous techniques and that only stalls threads that are likely to benefit from compaction. Our technique is based on the compaction-adequacy predictor (CAPRI). CAPRI dynamically identifies the compactioneffectiveness of a branch and only stalls threads that are predicted to benefit from compaction. We utilize a simple single-level branch-predictor inspired structure and show that this simple configuration attains a prediction accuracy of 99.8% and 86.6% for non-divergent and divergent workloads, respectively. Our performance evaluation demonstrates that CAPRI consistently outperforms both the baseline design that never attempts compaction and prior work that stalls upon all divergent branches.

Data-parallel memory systems must maintain a large number of outstanding memory references to fully use increasing DRAM bandwidth in the presence of rising latencies. Additionally, throughput is increasingly sensitive to the reference patterns due to the rising latency of issuing DRAM commands, switching between reads and writes, and precharging/activating internal DRAM banks. We study the design space of data-parallel memory systems in light of these trends of increasing concurrency, latency, and sensitivity to access patterns. We perform a detailed performance analysis of scientific and multimedia applications and microbenchmarks, varying DRAM parameters and the memorysystem configuration. We identify the interference between concurrent read and write memory-access threads, and bank conflicts, both within a single thread and across multiple threads, as the most critical factors affecting performance. We then develop hardware techniques to minimize throughput degradation. We advocate either relying on multiple concurrent accesses from a single memory-reference thread only, while sacrificing load-balance, or introducing new hardware to maintain both locality of reference and load-balance between multiple DRAM channels with multiple threads. We show that a low-cost configuration with only 16 channelbuffer entries achieves over 80% of peak throughput in most cases.

There has recently been much interest in stream processing, both in industry (e.g., Cell, NVIDIA G80, ATI R580) and academia (e.g., Stanford Merrimac, MIT RAW), with stream programs becoming increasingly popular for both media and more general-purpose computing. Although a special style of programming called stream programming is needed to target these stream architectures, huge performance benefits can be achieved.In this paper, we minimally add architectural features to commodity general-purpose processors (e.g., Intel/AMD) to efficiently support the stream execution model. We design the extensions to reuse existing components of the generalpurpose processor hardware as much as possible by investigating low-cost modifications to the CPU caches, hardware prefetcher, and the execution core. With a less than 1% increase in die area along with judicious use of a software runtime system, we can efficiently support stream programming on traditional processor cores. We evaluate our techniques by running scientific applications on a cycle-level simulation system. The results show that our system executes stream programs as efficiently as possible, limited only by the ALU performance and the memory bandwidth needed to feed the ALUs.