Intel® provides a collection of scripts (collectively known as Intel® Customer Scripts, or CScripts for short), which assist customers with platform debug and validation. Can they be made to run faster?
Debug of Intel-based designs can be made much easier using Intel® Customer Scripts (ICS), which are commonly known as CScripts. Their functionality ranges from basic state capture (register and memory dumps) to error injection/logging and sideband-enabled postmortem access.
One example of their utility is when a feature on a design is not working. CScripts are commonly used to compare the system configuration (set up by BIOS) to that of an Intel reference platform. For example, the output of the CScript sysTopo() function is compared to the same output on a reference platform.
Also, when a system catastrophic error occurs which locks up or “wedges” the processor, CScripts are very effective at identifying the cause of the failure. While sysError() is usually used for non-fatal errors, the sysDump() function performs a system dump even for catastrophic “three-strike” failures – a CATERR/IERR.
CScripts as implemented on SourcePoint are screamingly fast. As it uses JTAG direct register/memory/IO access methods with very little overhead, it performs much more quickly than legacy methods. For example, a complete sysDump() on a two-socket, 56-core system takes around five minutes.
CScripts are a powerful tool to help validate Intel platforms, and CScripts running on SourcePoint have dramatically better performance. SourcePoint combines powerful source-level debug, trace, and fast hardware validation on one platform.
Want to know more about SourcePoint’s CScripts implementation? The eBook Platform Debug using Intel® Customer Scripts (CScripts) provides an excellent introduction.