Abstract

early dissemination of its contents. In view of the transfer of copyright to the outside publisher, its distribution outside of IBM prior to publication should be limited to peer communications and specific requests. After outside publication, requests should be filled only by reprints or legally obtained copies of the article (e.g., payment of royalties).

No abstract

Abstract. In Clustered Instruction-level Parallel (ILP) processors, the function units are partitioned and resources such as register file and cache are either partitioned or replicated and then grouped together into onchip clusters. We present a novel partitioned register file architecture for clustered ILP processors which exploits the temporal locality of references to remote registers in a cluster and combines multiple inter-cluster communication operations into a single broadcast operation using a new sendb instruction. Our scheme makes use of a small Caching Register Buffer (CRB) attached to the traditional partitioned local register file, which is used to store copies of remote registers. We present an efficient code generation algorithm to schedule sendb operations on-the-fly. Detailed experimental results show that a windowed CRB with just 4 entries provides the same performance as that of a partitioned register file with infinite non-architected register space for keeping remote registers.

Abstract. Information hiding has been studied in many security applications such as authentication, copyright management and digital forensics. In this work, we introduce a new application where successful information hiding in compiled program binaries could bring system-wide performance improvements. Our goal is to enhance computer system performance by providing additional information to the processor, without changing the instruction set architecture. We first analyze the statistics of typical programs to demonstrate the feasibility of hiding data in them. We then propose several techniques to hide a large amount of data in the operand fields with very low computation and storage requirements during the extraction process. The data embedding is made reversible to recover the original instructions and to ensure the correct execution of the computer program. Our experiments on the SPEC CPU2000 benchmark programs show that up to 110K bits of information can be embedded in large programs with as little as 3K bits of additional run-time memory in the form of a simple look-up table.