Co-optimizing memory-level parallelism and cache-level parallelism

Abstract

Minimizing cache misses has been the traditional goal in optimizing cache performance using compiler based techniques. However, continuously increasing dataset sizes combined with large numbers of cache banks and memory banks connected using on-chip networks in emerging manycores/accelerators makes cache hitśmiss latency optimization as important as cache miss rate minimization. In this paper, we propose compiler support that optimizes both the latencies of last-level cache (LLC) hits and the latencies of LLC misses. Our approach tries to achieve this goal by improving the parallelism exhibited by LLC hits and LLC misses. More speciically, it tries to maximize both cache-level parallelism (CLP) and memory-level parallelism (MLP). This paper presents diferent incarnations of our approach, and evaluates them using a set of 12 multithreaded applications. Our results indicate that (i) optimizing MLP irst and CLP later brings, on average, 11.31% performance improvement over an approach that already minimizes the number of LLC misses, and (ii) optimizing CLP irst and MLP later brings 9.43% performance improvement. In comparison, balancing MLP and CLP brings 17.32% performance improvement on average. CCS Concepts • Computer systems organization → Multicore architectures.

Performance optimization of IoT based biological systems using deep learning

“…This agent receives the incoming request information and statistically analyzes the history of requested objects with respect to specific time intervals. The perspective of this agent is to know (1) which object is demanded from which requesting terminal (2) what type of computation objects are required (3) how much time an object needs for completing its computation (4) what other constituents and supporting objects are required to the object which is being processed (5) in which time interval objects are highly demanded (6) which object is requested before or after the object which is being processed. This agent understands such characteristics of the objects and proposes an execution plan for fulfilling upcoming needs.…”

“…Percent Gain = ((TMCA -TMPA) / TMCA) * 100 (4) Here TMCA stands for Total Migrations with Common Approach and TMPA stands for Total Migrations with Proposed Approach (i.e., TMCA and TMPA show the total number of required migration for both approaches.). Table 10 shows the migration plan that is computed by the Percent Gain formula on the sample dataset given in Table 3.…”

“…Currently, many approaches are available for optimizing system performance at these three levels [4]. Caching techniques share a common objective of improving cache hits while ensuring cache consistency in place.…”