Summer Intern Poster Session

Project: Dependence-Aware Scheduling Using Assistant Threads.
IBM T.J Watson Yorktown Site, Summer Intern Poster Presentation.
Content undisclosed.
Labels: America, IBM Intern, Research
Legend Since 1984
Cruising between Fantasy and Reality...

Labels: America, IBM Intern, Research
PPoPP: Principles and Practice of Parallel Programming
HPCA, High Performance Computing Architecture 即高性能计算体系结构。




电脑屏幕上持续着脚本语言的输出,“Processing benchmarks ...” 我的心情也紧张到了顶点……
开学前夕,一直按部就班的Research一时间陷入了泥潭。由于底层的simulator被发现了一个bug,导致我所有的数据统计工作都要重新进行。错误虽然小,不过足以改变所有的结论。之前还“凑合”的实验结果在更正错误之后准确率下降了10个百分点,变得有些不堪入目了。有些郁闷……
Labels: Research
After the 5-day trip to Los Angeles and the snow skiing and the nice party in Johnben's house, it is finally time to resume my research.
CSCI4011 Formal Language and Automaton Theory是一门极端严肃,极端讲求逻辑推理证明的课。课本的名字叫Introduction to Computation Theory,可以说,涵盖的内容是当今计算机科技中最最最基础的定理和概念。这些概念和定理奠定了计算机科学的基石。正因为如此,尽管4011被喻为本科生课程中难度最大的课,但它仍然被列在必修的列表中。
也许正印了“天将降大任”那句古训,本人首次TA就被“不幸”安排到这门课上。还好本科时上过这门课,有一定的理论基础。但是刚开始那段时间,面对完全陌生的课本,陌生的美国授课流程,陌生的学生的面孔,就像一下子被丢进了大海中,挣扎在漂浮与溺水之间。“天将降大任于世人也,必先苦其心志,劳其筋骨,饿其体肤,空乏其身,行胡乱其所为,然后动心忍性,增益其所不能”。近一个学期的磨砺,逐步掌握了课程内容,语言能力大大提高,和学生之间也熟悉起来,上课的时候不但能充分讲解概念和习题,还能游刃有余地施展各种授课技巧,必要的时候引用现实生活中的例子予以阐述,大量的和学生互动,再不然说点笑话活跃气氛…… 短短的一个学期,对自己的进展,连我自己都感到难以置信!这就是所谓的“增益其所不能”吧。
今天是本学期倒数第三次习题课,压着还略有咳嗽的嗓子,在讲台上如行云流水般进行了50分钟的逻辑推理后,竟然有种意犹未尽的感觉。I feel, I am enjoying teaching! (我想,我是在享受授课的过程吧。)
CSCI5103 Operating System 第二次期中考试揭晓,又拿到一个最高分。今天碰见教课的教授,他拍着我的肩膀说: "Good for you. You deserve it!" (“这是你应得的”)
关于Research,参与的一个项目的论文已经正式提交到本领域最有声望的会议上,等待明年一月的评审结果。如果有幸被接受,那么,2008的夏天,北京见!
生活充满了未知,充满了无限的可能,充满了等待发掘的精彩…… 真正的勇士,时刻整装待发,因为,战斗,才刚刚开始!
出来之后才感受到Linux的强大。系里面的网络都是基于Linux系统的,采用网络认证。一个有效的帐户,可以登录网络中的任何一台机器,从世界上任何一个地方。美国的IP资源就是丰富,没有内网外网这一说,也没有什么NAT。每台机器都有顶级域名、全球唯一的IP。
How to build faster microprocessor is always a hot research spot. The era of achieving speedup by simply increasing clock frequency has long been over, in the recent decades people placed more and more focus on achieving parallelism, both instruction-level and thread-level, to get more speedup.
历史总是在适当的时候重复自己。一个轮回之后,感觉自己再次陷入了循环的起点——黑暗时代again。 每天都感到时间是多么的不够用,好不容易忙完了上周的作业、TA的工作,紧接着新的一轮的任务接踵而至。
As a supporting material, Antonia wrote the study plan for me, from which, although not very formal, I can see which area I will work in for the following 5-6 years.
"Moore's law has contributed to the ability to integrate exponentially more functionality on a chip with higher performance, and was counted upon as a self-fulfilling prophecy that the semiconductor industry would implement. More recently, such advances in the semiconductor technology have enabled the integration of many cores onto a single die, and thus placed a new challenge in front of all computer system designers. Namely, how can we make efficient use of such abundant resources; and how can we fully utilize many-core processors? While both the hardware and software designers hesitate to take full responsibility of these tasks, one promising solution emerges---building a virtualization layer that operates below the operating system level to automatically manage on-chip resources at runtime. Virtualization layer that operates at this level is commonly referred to as the Virtual Machine Monitor (VMM). My group will investigate issues related to such VMMs from both the hardware and the software perspective.
At the maturity of this project, my team will provide a paradigm for designing a new generation of many-core processors that is capable of hiding the hardware complexity behind the VMM layer and present the software designers with a simple and reliable interface."
What an ambitious proposal! I am so interested and excited to take part of it.
Labels: Research