slugCompiler Optimization for SIMD type Vector Processor

Abstract

Performance of the processor can be enhanced by parallelization of instructions in terms of execution. Here we are applying compiler optimization techniques like loop unrolling, loop peeling for SIMD type Vector Processor. SIMD type vector processor is a high performance computational model which exploits the computational capabilities of both SIMD and vector architecture. SIMD type vector processor works on short vector instructions of vector length four and has four processing units which enables execution of four vector operands simultaneously [1]. To have the model which speeds up the computation in this paper we have created a CDFG (Control Data Flow graph) which gives the direct translation to the hardware. To create a CDFG we have used the tool SUIF2 and MACHSUIF which is a research project provided by Stanford University and Harvard university. To create a CDFG from the C source file first we have to do profiling on it to get the area of source file which have higher run time. We have unrolled the loops to get maximum ILP (Instruction level Parallelism) [7]. To limit the hardware we have unrolled only up to a maximum size of 4 [1]. Then a CFG (Control Flow Graph) is created for the source C file. Then the data flow analysis has been done on each basic block of the CFG by using the bvd class provided with MachSUIF. Then all the DFG (Data Flow graph) of every basic block are combined to get a full CDFG. In this paper We have designed the new hardware i.e. the SIMD type Vector processor based on HPL-PD VLIW architecture ISA which is supported by trimaran by default to accelerate the work.

Authors and Affiliations

Mohammad Suaib, Mohd. Akbar

Keywords

Related Articles

A Review on Digital Watermarking, its features, need and various techniques

The advent of the Internet has resulted in many new opportunities for the creation and delivery of content in digital form. Applications include electronic advertising, real time video and audio delivery, digital reposi...

PCA-Recommend: Increasing Trust on Recommendation models using the Similarity prediction on User rating and Item Rating

The recommendation modelling is challenging issue in the research of collaborative filtering model by integrating the user rating and item rating from the different source with rating sparsity and high dimensional struc...

Analysis of Flue Gas Flow Behavior in Economiser Duct Using Cfd

Energy saving and efficiency are the key issues in power generation system not only from the View point of fuel consumptions, but also for the protection of global environment .flue gas Ducts are the major parts of oil-...

Exogenous Foliar Application of Phenolic Acids on Quality Constituents of Tea

Exogenous foliar application of phenolic acid could alter the biochemical composition and enhance the antioxidant potential there by protect the living organisms from oxidative injuries. Though the reports on foliar app...

Energy Renormalization on the Sierpinski Pentagon Using Electrical Transform

In this paper the energy renormalization factor (resistence of the network) on the fractal Sierpinski Pentagon (Pentagasket) is described by the electrical network interpretation. First the Pentagasket-D5 is obtained by...

Download PDF file
  • EP ID EP17807
  • DOI -
  • Views 329
  • Downloads 12

How To Cite

Mohammad Suaib, Mohd. Akbar (2014). slugCompiler Optimization for SIMD type Vector Processor. International Journal for Research in Applied Science and Engineering Technology (IJRASET), 2(3), -. https://europub.co.uk/articles/-A-17807