antonblanchard/will-it-scale

will-it-scale

Overview

Will It Scale takes a testcase and runs it from 1 through to n parallel copies to see if the testcase will scale. It builds both a process and threads based test in order to see any differences between the two.

We rely on hwloc for a platform independent way of laying out tasks on cores. It can be found at www.open-mpi.org/projects/hwloc/.

Care is taken to try and reduce run to run variability. By using hwloc we ensure each task is on its own core and won't get bounced around by the scheduler. The wrapper script (runtest.py) turns off address space randomisation which can cause huge differences in pagetable related benchmarks (one run may fit within one pte page, the next may span two). There is a warmup period before which an average is taken. The averaging period can be changed with the -s option, which by default is 5 seconds.

Testcase design

Each test has two required components. A testcase description:

char *testcase_description = "Context switch via pipes";

and a testcase() which is passed a pointer to an iteration count that the testcase should increment. This testcase is run whatever number of times the user specifies on the command line via the -t option:

void testcase(unsigned long long *iterations)

A (not very useful) example:

#include <sys/types.h>
#include <unistd.h>

char *testcase_description = "getppid";

void testcase(unsigned long long *iterations)
{
    	while (1) {
            	getppid();
            	(*iterations)++;
    	}
}

If you need to setup something globally such as a single file for all parallel testcases to operate on, there are two functions:

void testcase_prepare(unsigned long nr_tasks)
void testcase_cleanup(void)

Finally if you need a new task such as when you want to write a context switch benchmark between two tasks, you can use:

void new_task(void *(func)(void *), void *arg)

This takes care of creating a new process or a new thread depending on which version of the test is being run.

Quick start:

make
./runalltests

Postprocessing and graphing

The graphing scripts use plotly, a client side javascript graphing package.

To generate html files for generated results:

./postprocess.py

Then load the generated html file in the browser.

The graphs show number of tasks run on the x axis vs performance (in operations per second) on the left y axis. We also plot the amount of idle time against the right y axis. In the ideal case we should have an X pattern, with operations per second increasing and idle time decreasing.