Jetpipeline: a Hybrid Pipeline Architecture for Instruction-level Parallelism
نویسندگان
چکیده
High performance processors based on pipeline processing play an important role in scientific and engineering computation. However, it is difficult to gain a satisfactory solution when taking both high degree of flexibility of parallel processing and low hardware complexity into account. This paper propose a hybrid pipeline architecture named Jetpipeline that possesses high degree of flexibility in parallel processing as well as suitable hardware complexity. Jetpipeline incorporates several instruction pipelines together. Multiple ALUs and floating point arithmetic pipelines are respectively used in the execution stages of these instruction pipelines. An instruction set for Jetpipeline is described in detail. Furthermore, a policy for scheduling instructions to instruction pipelines is discussed. The simulation results offer us the potential of Jetpipeline.
منابع مشابه
Software pipelining for Jetpipeline architecture
High performance processors based on pipeline processing play an important role in scientific computation. We have proposed a hybrid pipeline architecture named Jetpipeline in our former work. The concept of Jetpipeline comes from the integration of superscalar, VLIW and vector architectures. Jetpipeline has multiple instruction pipelines, which execute multiple instructions like superscalar ar...
متن کاملInstruction Level Parallelism Loop Unrolling
K – Survey of Instruction Set Architectures related to instruction-, data-, thread-, and requestlevel parallelism necessary for understanding Loop unrolling. ILP, Compiler techniques to increase ILP. Register Renaming, Pipeline Scheduling, Loop Unrolling. Conclusion. CPE 731, ILP. 3. Instruction Level Parallelism. 5 Optimizing Program Performance(Loop Unrolling and Enhancing Parallelism ) Michael.
متن کاملIntegrating Parallelizing Compilation Technology and Processor Architecture for Cost-Effective Concurrent multithreading
As the number of transistors on a single chip continues to grow, it is important to think beyond the traditional approaches of compiler optimizations for deeper pipelines and wider instruction issue units to improve performance. This single-threaded execution model limits these approaches to exploiting only the relatively small amount of instruction-level parallelism available in application pr...
متن کاملSwing Modulo Scheduling
[19] B.R. Rau and C.D. Glaeser. Some scheduling techniques and an easily schedulable horizontal architecture for high performance scientific computing. [22] J. Wang and C. Eisenbeis. Decomposed software pipelining: A new approach to exploit instruction level parallelism for loops programs. In IFIP, January 1993.
متن کاملExplicit Dynamic Scheduling: A Practical Micro-Data ow Architecture
This paper introduces Explicit Dynamic Scheduling (EDS), a practical implementation of dataaow on a chip. By combining RISC design principles with well-known compiler dependence analysis techniques, EDS combines a straightforward hardware design, suitable for high speed implementation, with the performance advantages of dataaow at the instruction level. EDS uniies pipeline and memory latency to...
متن کامل