ASSESSOR’S NAME: STUDENT’S DETAILS
First name: Last name: Email address: Lab Group:
Date of assessment:
SUB-SECTION
MAX. SCORE
NOTE FOR EACH COLUMN BELOW:
I) FULL OR MAXIMUM SCORE IS GIVEN IF THE CRITERIA IS MET.
II) PARTIAL SCORE IS GIVEN IF THE CRITERIA IS NOT FULLY MET.
III) ZERO SCORE IS GIVEN IF THE CRITERIA IS NOT MET. IV) PENALTIES WILL BE APPLIED AGAINST THE TOTAL MARKS OBTAINED ACROSS THE ASSIGNMENT.
Part A: PROGRAMMING TASK
1.0 -Simula ng the EV charging node
A single MPI process simulates one EV charging node. For instance, in a 3 ¡Á 3 Cartesian grid layout, there will be 9 MPI processes. Each node will have a (x, y) coordinate (e.g., Node 0 – (0,0), Node 1 – (0,1), Node 2 – (0,2), … Node 8 (2, 2)). Simula on should work for dynamic value of m x n nodes and threshold se ngs. At startup, the program allows the user to specify the grid size (m x n) and threshold values.
Each EV charging node has k number of in use or free charging ports simulated with POSIX threads (you may opt to use OpenMP as an alterna ve to POSIX thread). Each thread periodically updates its value, indica ng its availability.
The informa on in (b) is stored in a shared array, which can also be accessed by all threads within the same node
If all ports (or almost all ports) are in full use, the node will prompt for neighbour node data. To reiterate, the neighbour nodes refer to immediate top, bo om, right and le adjacent nodes (if exist).
If the received data from the neighbours shows a vacancy on the respec ve neighbour nodes, the repor ng node shall indicate the nearest available neighbour nodes. If the received data also show that all of their ports are heavily u lized, then the repor ng node will alert the base sta on that the node and its quadrant are all being used up.
The report sent to the base sta on should contain as much informa on as possible about the alert. You should demonstrate efficiency when repor ng an alert message to the base sta on. In this context, you should minimize the number of calls to the MPI Send (or ISend) func ons by a node to the base sta on when repor ng an alert condi on.
The node should receive a message from the base sta on about the nearest available EV charging nodes apart from its neighbour nodes.
Each node repeats parts (a) to (g) un l upon receiving a termina on message from the base sta on. Once the node receives a termina on message, the node cleans up and exits.
Note: External user interrup on (e.g., CTRL + C, or sending a ‘kill -9
2.0 – Simula ng the base sta on
A single MPI process simulates the base sta on node
The base sta on node periodically listens for incoming reports from the EV charging nodes
Upon receiving a report from an EV charging node, the base sta on checks for the nearest neighbour nodes that are available based on the neighbouring nodes of the repor ng node. For instance, if the repor ng node is Node 0, the neighbour nodes are Node 1 and 3. The base sta on will check whether there¡¯re any reports received from Node 1 and 3 (i.e., Node 2, 4 and 6):
– If no report is received from the nodes within a predefined period, the base sta on will suggest the available nodes to the repor ng node.
– If there are reports received from all of the nodes within a predefined period, the base sta on will send a message no fying the repor ng node that there are no available nodes nearby.
The base sta on writes (or logs) the key performance metrics to an output file.
The base sta on has an op on to send a termina on message to the EV charging nodes to properly shutdown for maintenance.
Note: External user interrup on (e.g., CTRL + C) are not allowed and should not be needed.
The base sta on uses a thread (i.e., POSIX or OPENMP) to send or receive MPI messages from the EV charging nodes. This thread is created by the base sta on and terminates properly at the end of the program.
Part B: REPORT
Methodology
Thorough illustra ons of network architecture along with a clear descrip on of these illustra ons. Compelling arguments are presented to jus fy the selected architecture with proper cita ons (in acceptable academic format, e.g., IEEE) to published papers in literature.
Note: Do not copy and paste the sample Figure in the assignment specifica ons into the report. You should draw your own diagram. No marks will be given for copying diagrams from the specifica on.
Thorough diagrams/flow-charts/pseudo-code capturing the EV network algorithm in details, including the algorithms on all the charging sta ons and the base sta on free of errors. Proper explana ons are required to explain the diagrams/flow-charts/pseudo-cdoes in the report.
Note: Proper technical diagrams and/or C style pseudo code formats must be applied here to be eligible for marks. Mathema cal style pseudocode is also acceptable. No marks will be given for any unclear/unreadable pseudocode formats.
Results tabula on
Tabulated results which indicates details of the applied simula on scenario. The results must include number of a empted runs, number of reported messages, and a summary of events generated. Includes screenshots of message logs. Addi onal charts are included to illustrate the number of triggered events over a period of me.
Note: Minimum 2 runs with different scenarios are expected.
Includes results when running the program on a local computer
Includes results when running the program using a larger grid size on a cluster compu ng setup (i.e., CAAS)
浙大学霸代写 加微信 cstutorcs
ASSESSOR’S NAME: STUDENT’S DETAILS
First name: Last name: Email address: Lab Group:
Date of assessment:
SUB-SECTION
MAX. SCORE
NOTE FOR EACH COLUMN BELOW:
I) FULL OR MAXIMUM SCORE IS GIVEN IF THE CRITERIA IS MET.
II) PARTIAL SCORE IS GIVEN IF THE CRITERIA IS NOT FULLY MET.
III) ZERO SCORE IS GIVEN IF THE CRITERIA IS NOT MET. IV) PENALTIES WILL BE APPLIED AGAINST THE TOTAL MARKS OBTAINED ACROSS THE ASSIGNMENT.
Analysis and discussion
Explana ons covering results in depth (based on significant amount of observa ons) are expected. Derived inferences must be explained and compared with hypothesis. Analysis are required on the communica on me between charging nodes and the base sta on, based on the exchanged messages and known issues. You are also required to include all possible causes of the known issues.
Includes analysis and performance comparison between running the program on a single computer and across a cluster compu ng setup (e.g., CAAS).
Part C: INTERVIEW + PENALTIES
Code demonstra on and Q&A
Describing the algorithm design, implementa on, results and analysis.
Deduct 0% of Part A:
The student has clearly prepared and understands the code. The student can answer ques ons correctly and concisely with li le to no promp ng.
Deduct up to 10% of Part A:
The student is reasonably well prepared and can consistently provide answers that are mostly correct.The student may lack confidence or speed in answering
Deduct up to 25% of Part A:
The student may have prepared the code and can give answers that are par ally correct but he/she clearly can’t engage in a serious discussion of the parallel algorithm/code
Deduct up to 50% of Part A:
The student may have par ally prepared the code before and can give some very basic answers. However, the student clearly can’t engage in a serious discussion of the code and demonstrates a poor understanding of the parallel algorithm/code.
Deduct up to 100% of Part A:
The student has not prepared and cannot answer most/all basic ques ons, or the student shows any indica on of not even seeing/recognizing the code before.
Code layout and compila on
Code structure & Comments
Deduct 0.5 marks: Some form of proper code structure and code is properly indented, basic comments included.
Deduct 1 mark: Poorly structured code, no proper code indenta on and li le or no comments.
MAKEFILE usage
Deduct 0.5 mark: No MAKEFILE is used.
Job script usage for CAAS
Deduct 0.5 mark: Incorrect usage of job script Deduct 1 mark: No job script
Report – Grammar & Format
Correct spelling, grammar, and punctua on; complete sentences; correct use of capitaliza on. Proper forma ng of the report.
Deduct 0.5 mark: Few spelling and grammar errors; correct punctua on; complete sentences.
Deduct 1 to 2 marks: No ceable spelling and grammar errors; most sentences have punctua on and are complete; uses upper- and lowercase. Report not properly forma ed.
Deduct 2 to 3 marks: Many spelling, grammar, and punctua on errors; sentence fragments; incorrect use of capitaliza on. Poorly forma ed report
Report – High similarity with other reports (using TurnItIn)
< 25% in content similarity is iden fied with another student's report or an external source.
Deduct 0.5 mark: Between 25% and 40% in content similarity is iden fied with another student's report or an external source.
Deduct 1 mark: Between 40% and 55% in content similarity is iden fied with another student's report or an external source.
Deduct 1.5 marks: Between 55% and 70% in content similarity is iden fied with another student's report or an external source.
Deduct 2 marks: > 70% similarity iden fied with another student’s report or an external source (Marks may be withheld pending further inves ga on).
Report – Reference sec on
Reference sec on is present and references are properly forma ed in an appropriate cita on format (IEEE or APA). The references were cited in the report.
Deduct 0.5 mark: Reference sec on is present but references are not properly forma ed in an appropriate cita on format (IEEE or APA). References not properly cited in the report.
Deduct 1 mark: No reference sec on
Late submission (Individual)
Deduct 2.5 marks (or 10% of total assignment marks) per late submission day.
Deduct marks:
e.g., -5 marks if the submission is made 2 days a er the due date.
Total Assignment Marks:
Addi onal comments by Assessor:
Programming Help, Add QQ: 749389476