[ Pobierz całość w formacie PDF ]
.In the rare event that a problem occurs, it isfixed first thing the next week.Then the functional teams begin theirnext two-week iteration.The Modules There are also cross-functional teams whose job it is toput together modules that assemble infrastructure and tools to test spe-cific silicon configurations; you might think of modules as apps run-ning on the platform.The module teams work in two-week iterations,offset by a week from the platform teams.Thus, module teams start aweek after the platform teams complete a technology drop, giving thelatest drop time to be validated and for any problems to be fixed.The module teams have only a week to assemble (or modify) a mod-ule; the first weekend of the iteration is spent testing each module withshort runs of silicon chips, and the early part of the second week isspent fixing any problems.As soon as all modules are solid and freeof defects, the complete testing system is integrated and ready for itsfull-scale, 48-hour validation.Because of the extensive module-testing 54 THE LEAN MINDSETprocedures, it is rare that module defects are found during final vali-dation.However, issues do arise that are exposed by the volume test-ing, and they are addressed at the beginning of the next week beforenew module development begins.Integration and Validation The integration team works in two-week cycles, with the bulk of their work focused on weekend integra-tion and validation.On the first weekend they integrate the technologydrop from the platform teams and also validate the new modules withshort runs of silicon.On the second weekend they do a full-scale inte-gration and run the 48-hour validation testing, with the robust robotarm doing its job of placing tens of thousands of chips in sockets soevery aspect of the test can be validated.If any issues are found, theintegration team works with the appropriate teams early the followingweek to resolve them.A root cause analysis of each problem is con-ducted and measures are put in place to prevent a recurrence.After all,the integration team is not supposed to find problems; it is supposedto not find problems.ResultsThings went extremely well when first silicon arrived.Capacity to de-liver had indeed improved by a factor of three over two years earlier;in fact, it had improved much more when compared to the pre-agiledays.The test success rate, already very good with agile development,increased another notch.But something happened during the two-year cycle that caught PDEby surprise.Market conditions had prompted the addition of moreproducts to the new-generation silicon than had originally been an-ticipated.So tripling productivity was not quite enough.By addingpeople and working longer hours, the PDE teams were able to keep upwith the unexpectedly large workload, helped greatly by their rapidcadence and the rigorous validation regime.Expanding across IntelIt took many years of continuous improvement for the Oregon PDEdepartment to reach its current level, and there is no doubt that moreimprovement will follow.But PDE is just one small group of engi-neers in a very large company.The next step is to spread the learningthrough both high-level support and on-the-ground coaching.Intel isin the process of establishing both company-level and site-level agile CHAPTER 2 ENERGIZED WORKERS 55competency centers to coordinate communities of practice and provideexperienced coaches to support and accelerate adoption of the im-proved processes Intel needs as it continues its quest to keep up withMoore s Law and a rapidly changing market.Otto: You would think that PDE has figured out mostof the improvement that s possible.How can theykeep on improving during the next cycle?M&T: We don t know what PDE will do for the nextcycle, but since Intel has been doing the seemingly impossible fordecades, we assume they will figure something out.Intel is a com-pany where you find the kind of engineers who are eager to bechallenged and invent better ways of doing things.The Science of ExpertiseThink about musicians and athletes: How do they become virtuosos orchampions? They practice.Decades of research, initiated by the workof Anders Ericsson, has led to consistent findings: Expertise is devel-oped through hours upon hours of deliberate practice [ Pobierz całość w formacie PDF ]