Interviewing the parallel programming idols
Interview with parallel programming expertsBy Michael suessthinking parallel (http://www.thinkingparallel.com) by Liu weipeng (pongba) C ++'s Luo Yun (http://blog.csdn.net/pongba) Original article link (http://www.thinkingparallel.com/2007/03/14/interviewing-the-parallel-programming-idols/) I said, I don't feel like I am a parallel programming expert. There are always so many things waiting for you to learn in this field, so many ways are waiting for you to go, so many exciting new technologies are waiting for you to try; not to mention that there are so many big cows who have a longer time in this field than I do not know-some are even earlier than I was born. I have always benefited from discussing and learning with them. So some time ago I tried to contact some of them by email (they are my parallel programming Daniel, prawns, idols ...). Fortunately, some of them replied to my email. In the mail, I asked each of them the same ten questions: five general parallel programming questions, and the other five questions about
TheirParallel Programming System. I will publish an interview each week from next week. Anyway, let's make a notice first. The following are the experts I interviewed (sorted alphabetically by their parallel programming systems-(e) rlang, (m) PI, (O) penmp, (P) osix threads ,. net-threads): Joe Armstrong for erlangwilliam gropp for mpisanjiv Shah for openmpdavid butenhof for POSIX threadsjoe duy. net-threads I would also like to interview William put and Doug Lea and ask some questions about Java threads, but so far I have not received an email from them. If any of you knows who to interview about Java threads, please leave a comment. Oh, by the way, I almost forgot. The following are my ten questions (PPS stands for parallel programming systems for specific interviewees): As we are currently entering the multi-core era, do you think this means parallel computing?
FinallyIs it going to enter the mainstream? Or is this just a phase, and soon I will be interested in parallel programming (again) only in the high-performance computing field? We often see heated debates
Shared Memory ProgrammingAnd
Message transmissionWhich is the best method for parallel programming. What do you think of this? From your point of view, what do you think is the most exciting progress/innovation in the field of parallel programming now or in the past few years? What is the future of parallel programming? Will there be a silver bullet? Parallel Programming is currently the most frustrating problem because it is still more difficult and less productive than sequential programming. Do you think this situation may change? What are the unique advantages and weaknesses of PPS compared with other parallel programming systems? What improvements do you think should be made? What are the differences if you are given the opportunity to design PPS from the ground up? Do programmers who want to program with PPS have any special suggestions on tools? For example, IDE, editor, debugger, profiler, correctness tools (correctness tool) or something. Give some suggestions to programmers who are new to PPS! For example, how to get started? Books, tutorials, online resources, and questions, etc. What is the worst mistake you encounter in PPS programming ...?
