Tekastu_WEb

Intel

System Level SoC Validation Engineer, Digital Enterprise Group – Tolapai (Low Power IA Embedded ) Worked in Digital Enterprise Group Tolapai, Intel’s embedded system-on-a-chip combining Pentium M processor core, DDR2 memory controllers & I/O controllers, &QuickAssist integrated accelerator for security functions

Lead the focused testing efforts of a memory block within the SoC, developed its test cases, updated its Test

Plan, improved the silicon validation hardware infrastructure, automating the manual tests, regressively debugging the failures with proper documentation & inventory. This led to increase in efficiency by 20% in early bug detection.

  • Focused testing : Developed code in C ,Tcl and assembly to regress the memory block. This block interacted with IA and microengines to perform hashing operations, read/ write register and memory tests, ring tests and push/ pull reflector tests.

  • Test Plan Development : Created test plans, analytically-designed test cases, wrote code in C & executed for focused & concurrency testing

  • Concurrent Testing : Performed concurrent testing by randomizing various parameters during compile & run time. Also generated pseudo random stress tests regressing whole SoC using Python over a Linux based Tcl infrastructure to identify bugs earlier.

  • Automation Infrastructure Development : Wrote code in Python to test automation infrastructure. This would ensure proper communication between host, silicon validation platform, test cards & equipments. It initializes, run & gather results after test execution with logs

  • Wrote code in Python for post processing of results, cleanup, process logs, store results & debug information

  • Chip Bring Up : Worked on chip bring up, ensuring that after power on smooth boot process and that the OS is ported properly

  • Debug : Debugged failures using debug tools (JTAG, Kernel debuggers), communicated status, and resolved with stakeholders. Exercised various DFX features, analyzing scandump/memdump

  • Conducted manual tests until boards become stable and then put them into automated execution for regressions