Detailed dependency analysis and I/O profiling
Resolve application file and network dependencies for migration.
Capture application environment and I/O patterns for debugging deployment and performance issues.
Profile application I/O for storage sizing and tuning.
Why profile I/O?
Even highly-optimised applications can have bad I/O patterns in deployment so it is a good idea to profile all applications, even when you can’t change the source code.
Problems such as storing files in the wrong place or trawling the PATH variable happen all the time and are easy to fix.
Fixing problems in third party tools at Arm
This chart shows reads and writes from a build at ARM. It is using shared storage for temp files by mistake, instead of local storage. It caused significant slow down, but was easy to fix.
“Why does this application work for you and not for me?”
Deep dive into application arguments, environment, libraries and dependencies to debug deployment and performance issues. Breeze gives you everything you always wanted from strace.
HPC user education
Make every engineer an I/O expert with an I/O healthcheck report.
The Ellexus Healthcheck reports give a straightforward I/O overview and performance recommendations based on common I/O issues.
Tool and system optimisation
“Why is my application slow today?”
Automate performance and regression testing with detailed I/O profiling. Export file and dependencies for application correctness and migration.
We changed the environment and the problem was solved! We solved the problem in 45 minutes. I’m happy with that!
After changing my script and rerunning I observed about 87% improvement in incremental build (16 sec to 2 sec)!
Using Breeze, we could immediately see that it was not a dependencies issue so both Ansys and our customer saved time.
Hard-coded paths caused a massive slowdown where jobs were taking 3x longer. Using Breeze, we quickly found the hard-coded path and later confirmed the fix.
With the tools from Ellexus we were able to identify why we were hitting I/O bottlenecks in areas where we expected full CPU utilisation.