Chapter 5
Link to the solution:
5.1 Compare and contrast sampling and tracing profiles.
5.2 Why is it important to profile an application with a problem set that is representative of those used in production?
5.3 Why is it useful to obtain a profile which includes loop trip counts (consider -hprofile_generate or similar)?
5.4 What are hardware performance counters and what kinds of information do they provide?
5.5 Construct a case study: Select an application and problem set for further examination.
a. Obtain a sampling profile of the application.
b. Obtain a tracing profile of the application.
c. Obtain a profile of the application including a call-tree.
d. Obtain a profile of the application including loop statistics.
e. Obtain a profile of the application with hardware counter data.
f. From the above profiles, identify:
i. Degree of load-imbalance.
ii. Important routines and looping structures.
iii. Locations of any significant MPI communication.
5.1 Compare and contrast sampling and tracing profiles.
5.2 Why is it important to profile an application with a problem set that is representative of those used in production?
5.3 Why is it useful to obtain a profile which includes loop trip counts (consider -hprofile_generate or similar)?
5.4 What are hardware performance counters and what kinds of information do they provide?
5.5 Construct a case study: Select an application and problem set for further examination.
a. Obtain a sampling profile of the application.
b. Obtain a tracing profile of the application.
c. Obtain a profile of the application including a call-tree.
d. Obtain a profile of the application including loop statistics.
e. Obtain a profile of the application with hardware counter data.
f. From the above profiles, identify:
i. Degree of load-imbalance.
ii. Important routines and looping structures.
iii. Locations of any significant MPI communication.
5.1 Compare and contrast sampling and tracing profiles.
5.2 Why is it important to profile an application with a problem set that is representative of those used in production?
5.3 Why is it useful to obtain a profile which includes loop trip counts (consider -hprofile_generate or similar)?
5.4 What are hardware performance counters and what kinds of information do they provide?
5.5 Construct a case study: Select an application and problem set for further examination.
a. Obtain a sampling profile of the application.
b. Obtain a tracing profile of the application.
c. Obtain a profile of the application including a call-tree.
d. Obtain a profile of the application including loop statistics.
e. Obtain a profile of the application with hardware counter data.
f. From the above profiles, identify:
i. Degree of load-imbalance.
ii. Important routines and looping structures.
iii. Locations of any significant MPI communication.
a. Obtain a sampling profile of the application.
b. Obtain a tracing profile of the application.
c. Obtain a profile of the application including a call-tree.
d. Obtain a profile of the application including loop statistics.
e. Obtain a profile of the application with hardware counter data.
f. From the above profiles, identify:
i. Degree of load-imbalance.
ii. Important routines and looping structures.
iii. Locations of any significant MPI communication.