140 likes | 295 Views
Scaling The Speedup of multi-core chips based on Amdahl’s law. A.V. Bogdanov , Kyaw Zaya. DUBNA, 2012. Speedup in parallelism. Architecture of the machines Operation Systems Memory and processor resources Number of processor Problem size The use of algorithms.
E N D
Scaling The Speedup of multi-core chips based on Amdahl’s law A.V. Bogdanov, KyawZaya DUBNA, 2012
Speedup in parallelism • Architecture of the machines • Operation Systems • Memory and processor resources • Number of processor • Problem size • The use of algorithms
Constant problem size scaling • Independent on number of processors • Using Amdahl's Law as an argument against massively parallel processing is not valid • The serial percentage is not practically obtainable
Where, N - number of processors α - The proportion of successive calculations β, γ - parameters characterizing of the network -.
Types of multi-core chips Symmetric multi-core chip Asymmetric multi-core chip Dynamic multi-core chip
Symmetric multi-core chip Processor Processor Processor Processor I/O Memory Where, (f)= software fraction that is parallelizable n= the total number of BCE (base core equivalents) r= number of BCE on a kernel Perf(r)= performance of r number cores
Asymmetric multi-core chip Private memory Private memory Processor Processor Common Memory I/O
Dynamic multi-core chip Sequential mode Parallel mode
Benchmarking the T-Platforms Cluster The test result of NPB LU Class C
Conclusion • New Technologies and Architectures • Clustering problems • Modifications on new types of processor