To access the full text documents, please follow this link: http://hdl.handle.net/2099.1/4483

Benchmarks en diferents microprocessadors
Caubet Gomà, Josep
Koren, Bengt
This report is the presentation of a Master Project Degree, part of Master of Science in Telecommunication Engineering at The Royal Institute of Technology, Stockholm, Sweden and the Technical University of Catalonia, Barcelona, Spain. The project was conducted between February and July of 2006 at Imsys Technologies AB “Imsys” in Upplands Väsby, Sweden. The objective of this thesis is to obtain results in benchmarking of different microprocessors with an emphasis of Imsys proprietary architecture and comparison between an older versus a new processor design and with a special reference to power efficient computing. The organization of the report in chapters coincides with the steps that have been carried out to obtain the results. The previous work and background information is described in the section Background followed by the section of Board Features, where the microprocessors are explained in more detail. The chapter of Benchmarks shows the tests carried out and the different benchmark types that have been carried out using different programming languages (C, Asm and Java), including some generally well known benchmarks (Dhrystone and Whetstone). A specific TCP/IP test program was also included in the test suite. Finally, after having carried out the planning and the execution of the benchmarks to fulfil the thesis, a conclusion with an analysis of the obtained results is made.
The objective of this thesis is to obtain results of benchmarking between different boards; RCM3720 of Rabbit Semiconductor and SNAP and IM3000 of Imsys Technologies. Three different computer languages have been used: C, Java, and assembler. A set of algorithms have been run on different microprocessors and in one case the processor was simulated using a processor simulator (The IM3000 True cycle simulator). The tests that have been carried out contain important different aspects of a microprocessor: the execution time of different algorithms, the execution of popular benchmarks (Whetstone and Dhrystone) and TCP/IP performances in SNAP. The analysis of the obtained results concludes that IM3000, programmed in assembler is the quickest board. The SNAP, programmed in Java, is the slowest. IM3000 triplicates the SNAP speed. RCM3720 is around 5-10 times execution speed slower than IM3000. The transfer of data TCP/IP in SNAP is around 4,7 Mbps (both ways).
Àrees temàtiques de la UPC::Informàtica::Arquitectura de computadors
Best Practices benchmarking report
Microprocessors
Benchmarking
Microprocessadors
Benchmarking
Microprocessadors
info:eu-repo/semantics/bachelorThesis
Universitat Politècnica de Catalunya
         

Show full item record

Related documents

Other documents of the same author

 

Coordination

 

Supporters