Breakin is Advanced Clustering’s stress-test and hardware diagnostics tool. Our engineers developed Breakin because no other available product — commercial or open-source — could pinpoint hardware issues and component failures as well as they wanted.
The Breakin application proved to be so powerful at finding hardware issues that it is now used by several major component manufacturers in their validation and factory testing procedures.
All of the clusters we build spend a minimum of 24 hours running the Breakin stress testing. Breakin performs the following tests:
Full processor and memory test: Using HPL, the supercomputing industry standard performance benchmark, the tests fully utilize all processor cores and physical memory installed in the system. Activating all cores and hardware at 100% utilization uncovers any power- and thermal-related design issues.
Hard drive stress testing: During Breakin’s operation, all of the hard drives in the system are fully exercised and checked to find any faults.
Temperature sensor reporting: With the support of the Linux hardware health monitoring package lm_sensors, any supported temperature sensor is reported.
ECC error reporting: Using the EDAC Linux kernel patches, Breakin reports ECC errors.
MCE reporting: Breakin reports any MCE (Machine Check Exception) events generated by the system, which often can point to serious hardware issues.
Although we use our Breakin software for testing individual machines in your cluster, that’s not the end of our testing process. After every system passes our hardware Breakin test, it moves onto a cluster-wide test, which is performed after the system is installed with the final operating system, drivers and software configuration.
Request More Information
Breakin Features and Benefits
Extensible
Breakin has a modular testing architecture. New stress-test scripts can be added to your custom breakin install.
Remote Access
Breakin runs an SSH server, so you can remotely monitor the system while running the tests.
CPU and Memory Test
Using applications using the most intensive math libraries, breakin fully tests all processor cores and physical memory installed in the system.
Hard Drive Tests
Using a non-destructive test all drives in the system are exercised to find any faults.
Temperature Readings
If your system supports IPMI temperature and fan data is shown to make it easy to check for adequate cooling.
ECC Reporting
With most systems that support ECC memory, Breakin can provide details on any memory errors that occur.
Hardware Requirements
Breakin should support any hardware supported by the Linux kernel. The main tests are optimized for both the AMD and Intel processor architectures using math libraries created by the processor manufacturer. Advanced Clustering uses breakin every day on a variety of systems and architectures:
Breakin has been fully tested and works on any recent Advanced Clustering Pinnacle Servers or Workstations.
Should also work on:
- Socket F dual-core (2200 and 8200 series) and quad-core (2300 and 8300 series)
Socket G34 (6100-6300 series)
- AMD Opteron systems
Socket 940 single- and dual-core (200 and 800 series
Intel Xeon systems - Intel Xeon 5XXX processors (including 5400, 5500, and 5600)
- Intel Xeon 3XXX processors (including 3000, 3200, 3400, 3500, and E3-XXXX)
Intel Core series systems - 2nd Generation Core
- i3/i5/i7 series
- 1st Generation Core i3/i5/i7 series
- Core 2 Duo or Quad
- Many older platforms should also work. (Please let us know if you use breakin on older systems and we will update this list.)