We will try to get disk I/O as fast as possible and evaluate the effects of caches and the cost of system calls.

INSTRUCTIONS TO CANDIDATES

ANSWER ALL QUESTIONS

Goal

The project is centered around performance.

We will try to get disk I/O as fast as possible and evaluate the effects of caches and the cost of system calls. In the end you should have a good understanding of what sits in the way between your process requesting data from disk and receiving it.

 

Breakdown

 

1. Basics

Write a program that can read and write a file from disk using the standard C/C++

 

library's open ,  read , write , and

Add parameter for the file name;

 

functions.

 

 

Add parameter for how big the file should be (for writing);

Add a parameter to specify how much to read with a single call (block size); Way to execute:

2. Measurement

When measuring things it helps if they run for "reasonable" time. It is hard to measure things that run too fast as you need high-precision clocks and a lot of other things can affect the measurement. It is also annoying to wait for a long time for an experiment, especially if you have to do many experiments. For this reason you should make sure that your experiments take "reasonable" time. I recommend something between 5 and 15 seconds.

Write a program to find a file size which can be read in "reasonable" time.

Input: block size Output: file size

 

 

 

3. Raw Performance

Because you would be looking at different file sizes when using different block sizes, it makes sense to use units for performance that are independent of the size. Use MiB/s (this is megabytes per second).

Make your program output the performance it achieved in MiB/s

 

Make a graph that shows its performance as you change the block size. block size on the x axis and performance on the y axis.

We will not mandate a way to run this... anything that works for you to produce the

graphs.

 

4. Caching

Once you have a file of "reasonable" size created, reboot your machine. Call your program to read the file and note the performance.

Call your program to read the file again (immediately after) and note the performance.

Ideally you should observe the second read be much faster, assuming the file can fit in your physical memory. This is the effect of caching.

 

 

Experiment with clearing the caches and not.

Make a graph that shows performance for cached and non-cached reads for various block sizes.

We will not mandate a way to run this... anything that works for you to produce the graphs.

 

5. System Calls

If you have very, very small block size (e.g. 1 byte), most of the time would be spent trapping into the kernel. We talked in class how system calls are expensive. Let's quantify it!

Measure performance MiB/s when using block size of 1 byte

 

Measure performance in B/s. This is how many system calls you can do per second.

 

Try with other system calls that arguably do even less real work (e.g. lseek)

We will not mandate a way to run this... anything that works for you to produce the graphs.

 

6. Raw Performance

Try to optimize your program as much as you can to run as fast as it could.

Find a good enough block size? Use multiple threads?

Report both cached and non-cached performance numbers.

 

 

Attachments:

Related Questions

. Introgramming & Unix Fall 2018, CRN 44882, Oakland University Homework Assignment 6 - Using Arrays and Functions in C

DescriptionIn this final assignment, the students will demonstrate their ability to apply two ma

. The standard path finding involves finding the (shortest) path from an origin to a destination, typically on a map. This is an

Path finding involves finding a path from A to B. Typically we want the path to have certain properties,such as being the shortest or to avoid going t

. Develop a program to emulate a purchase transaction at a retail store. This program will have two classes, a LineItem class and a Transaction class. The LineItem class will represent an individual

Develop a program to emulate a purchase transaction at a retail store. Thisprogram will have two classes, a LineItem class and a Transaction class. Th

. SeaPort Project series For this set of projects for the course, we wish to simulate some of the aspects of a number of Sea Ports. Here are the classes and their instance variables we wish to define:

1 Project 1 Introduction - the SeaPort Project series For this set of projects for the course, we wish to simulate some of the aspects of a number of

. Project 2 Introduction - the SeaPort Project series For this set of projects for the course, we wish to simulate some of the aspects of a number of Sea Ports. Here are the classes and their instance variables we wish to define:

1 Project 2 Introduction - the SeaPort Project series For this set of projects for the course, we wish to simulate some of the aspects of a number of