To be offered Autumn, 2015
We will give a functional course on adequately utilizing PCs to tackle logical registering issues in building and the physical sciences.
The course is suitable for understudies with small processing foundation as it begins with a prologue to UNIX environment and to programming. The second 50% of the course is devoted to hands-on parallel programming.
Fortran90 is utilized to outline the ideas in light of the fact that it is a low edge dialect and understudies can begin composing parallel code after generally short guideline time. On the off chance that C was utilized, a great deal additional time would need to be spent on the programming presentation and code investigating taking center off the parallel part.
Review
The significance of investigative processing has been quickly expanding in every aspect of science and building. It is key for understudies to figure out how to compose productive projects that exploit the (frequently considerable) computational force accessible to them.
Customarily, educating around there has concentrated on arranged dialects (C, Fortran) and less demanding to-utilize instruments, for example, Matlab. Be that as it may, nowadays most portable PCs use multicore processors, and a solid establishment in parallel registering is required also. This course mixes the conventional and "cutting edge" approach and acquaints understudies with Fortran 90, Python, parallel processing with OpenMP and MPI.
The course will comprise of 2 addresses and 1 lab session for every week (3 hours all out). There will likewise be 4 programming assignments, and a programming venture. Before the end of the course, understudies will be arranged to handle research issues utilizing the devices of cutting edge superior experimental figuring in an educated, viable, and proficient way.
Syllabus
L1
Course Overview. Beginning - working with UNIX and the order line
L2
Programming variant control with GIT and Bitbucket
L3 - L6
Programming and logical figuring with Python
L7 L10
Prologue to Fortran 90 - programming and utilizing prevalent libraries. Makefiles for multi-record codes.
L11
Prologue to parallel figuring: Methods, devices, and execution
L12 - L14
OpenMP with Fortran90 for parallel programming of shared-memory PCs
L15 - L17
MPI with Fortran90 for programming on dispersed memory machines, for example, bunches
L18 - L19
Parallel figuring samples: quadrature, Monte Carlo systems
L20
Course audit
Note: Good programming rehearse - arranging, unit testing, investigating, approval - to be coordinated with the above points and the programming assignments.
source:--Imperial College Introducton to HPC
0 comments:
Post a Comment