Friday, March 28, 2014

Great chat with Yunji Chen at ASPLOS

It was the break before the last session, I grabbed my suitcase and was heading to the last session. Then I saw Yunji Chen, who also did his undergraduate at USTC. Since he was on one of the best papers, I approached him and said "hi and congrats" to him. The chat lasted way longer than I expected, and I learnt way more than I expected.

He views doing research just like playing video games. It was not until Yunji got his Ph.D. then he realized that he probably should spend less time on video games and more on research. He dares to dream big, and the ASPLOS awarded paper stems from the dream of building real intelligence. He is lighten up on research, and he is also helping his students, many of them also played lots of video games as undergraduate, be lighten up.

After one and half hours, I had to interrupt to checkout and we wrapped up. The good thing is that I have already been lighten up by then, and begin to dream about having my papers in ASPLOS 2018 which hopefully will go ShanYa as proposed by Yunji.

Monday, March 24, 2014

My Exciting Adventures in Utah


ASPLOS 2014 was a wonderful experience for a myriad of reasons! I had the opportunity to meet with a diverse set of people (both technically and culturally), curl the stone, discuss exciting research topics, meet old friends, observe superb presentations, and have my ideas hacked and improved. I find that each of these made the conference an outstanding experience.

The presentations were well done, insightful, and thought provoking. With authors of the various presentations, I discussed several interesting ideas on topics ranging from randomly throwing away instructions for performance, telling the operating system to keep its hand off my app memory, and even build my iOS app once and run on Android. These are just a few of the memorable discussions I was privy to while attending ASPLOS.

A last aspect of my ASPLOS adventure to comment on was the opportunity to mingle my ideas with a fresh crop of interesting research directions and perspectives. This exposure led to an exuberating set of new ideas, which sadly I will have to wait upon to explore because there are so many fun things to look into. It is truly a great time to be a systems+compilers security researcher.

I would like to say thank you to the ASPLOS chairs, organizational committees, reviewers, and (I imagine) numerous volunteers who made this an exciting and memorable time; Also, I would like to send out a great thanks to all the other participants at ASPLOS who shared in spirited debates and discussions.
This was my first time attending ASPLOS. Overall it was a great experience. The organization, the quality of the presentations, the food and the view of SLC were all great. Sadly I missed the excursion :(

I was excited by the broadness of the topics and enjoyed the chance to speak with researcher from both industry and academia. As to the sessions, I think the lightning session was a great idea, it really helped me decide which session to go. The debate panel session was also inspiring and insightful. Among the main sessions, I enjoyed most the heterogeneous computing session the virtualization session. And every other sessions have one or more interesting papers.

I'm looking forward to attending ASPLOS'15 in Istanbul!

Thursday, March 20, 2014

ASPLOS 2014

This is the first time I attend the ASPLOS conference and I had a great time. One of my favorite parts of the conference was the poster session because it allowed me to interact with other attendees that were interested in our work. Since I am a second author on the paper that was accepted at ASPLOS, it gave me the opportunity to talk about my research.

I thought the conference was very well organized and the location was beautiful. The venue was a really great space to chat with others over the different papers that were presented. I also enjoyed curling since it was my first time and the instructors were great!

I am very thankful for the travel grant that allowed me to attend and I also want to thank the ASPLOS committee for organizing such a great conference.

Wednesday, March 19, 2014

My ASPLOS'14 -- totally awesome

This is my 3rd time ASPLOS -- previously London and Houston.

The view of SLC is breath-taking (see the following photo taken from my room). The organization is also perfect, in particular the Olympics Oval part. Nice souvenir -- a jacket.

The 90-sec lightning session is new this year. This indeed adds some extra workload to presenters -- but it really gave an excellent overview of the program. So please consider keep it, future program chairs! I had to miss multiple great talks due to the dual-track sessions, which might be inevitable. Probably the sessions are better to be organized according to fields (arch, pl, os, ...) rather than topics. 

Tuesday, March 18, 2014

This round of ASPLOS yielded the most interesting conference conversations I can recall recently.  The conference-attached poster session and lightning talks (particularly the latter) were positive additions to this year's ASPLOS.  The extra bit of work for authors was worth it, and resulted in more accidental chances for discussion of my paper than I've had at previous venues.  As nice as single-track sessions are, I did not feel the two tracks to be too difficult.

SLC's mountains made for a pretty setting, and I regret not having time to take advantage of the Utah outdoors.  I was baffled by the seeming emptiness of the immediate city center area, but this seems to accompany convention centers in general.

Despite the great additions to the conference format, my favorite feature of this year's conference was the decision to bring SIGMOO on board with a forward-thinking session on dairy center computation.
I recommend that future ASPLOS organizers repeat this session (or other topics in agricultural support) and milk it for all its interdisciplinary worth.

Experiences at ASPLOS'14

This was my first time to attend ASPLOS and my second time to have a paper at ASPLOS. The conference main program is excellent and the conference is very well organized. The lightning section is a very good idea, especially given the parallel sessions. I wish the poster session could be held even earlier, e.g., during the reception. The WACI session is very interesting and inspiring, particularly Michael Wei's talk on Dirty RAM and Rotten Caches : Saving the World From Useless Updates.

I really like the inter-disciplinary nature of ASPLOS, which encourages and creates an opportunity for computer scientists in different area to collaborate and inspire each other. I think it is the essential ingredient in making larger impact.

Finally, I want to thank the organizing committee for their great effort in making such great event happen and the generous support on travel grants.

Stefan Bucur's thoughts on ASPLOS'14

I am a PhD student at EPFL, who works on scaling automated testing techniques, such as symbolic execution, to large real-world software systems. This year, I got the chance to present at ASPLOS my work on generating symbolic execution engines for interpreted languages.

I think my other peers have already done a great job at covering the event, so I won't reiterate on that. I'll just mention that I was happy with how my talk went and with the great feedback I received from the people I met. As an ASPLOS first-timer, I was pleased to meet a community that fits so well my research area.

I will express here, instead, my own view on the topic debated in a session during the last day of the conference: “Resolved: Specialized architectures, languages, and system software should largely supplant general-purpose alternatives within the next decade.” I really enjoyed the discussions and it was fun to see how this polarized the audience to the point that the chair had to cut on questions, as we were running out of time.

Since I didn't get the chance to express my opinion to the audience there, I'm doing it here: I think both specialization and generalization camps deserve the fine bottles of wine offered as a prize, because both trends are bound to stay in the computer architecture. The reason is simple: any computer system roughly operates at two boundaries: the physical boundary -- the hardware artefacts that support its existence -- and the human interface layer, where I include both end users and developers. These two boundaries create the opposing forces of specialization and generalization.

On the one hand, physics is "heterogenous" by nature. You can't drive a Lamborghini over a corn field -- you'd have to use a tractor; there are different types of vehicles available for the different types of terrain that exist on the planet. In the same way, a good computer architecture will need to work around and take advantage of any physical peculiarity in order to make the computation more efficient.

On the other hand, the human interface layer should be as uniform as possible, because human minds cannot tackle complexity well. Ideally, a developer would be happy to write the system entirely in, say, Python. But in practice, there are no good compilers to turn Python in an optimal hardware+software system, and the language becomes hard to maintain at large scales, so this is the point where the generalization and specialization forces collide.

These two forces collide and reconcile in any system. It only differs the abstraction level where this happens, and history had showed that these points evolve in time for a particular system, driven by technological trends. The question is not whether we should aim for one or the other, but how we can automatically generate optimal hardware and software that translate the high-level human intent. There's a lot of research work ahead of us :)


ASPLOS is great, you should come

This was my first time at ASPLOS, and I had a great time.

I'm very grateful to the organizers and sponsors for the travel grant.  These funds really help non-US folks like me to travel across the ocean and interact with the ASPLOS community, thus opening the door to new collaborations and great projects.

The most striking thing about ASPLOS is how multidisciplinary it is.  Being exposed to many novel ideas covering multiple areas was a powerful experience that stimulated many new ideas (some of which I might even pursue).  Discussing the technical issues with attendees---presenters, students and senior people---was very interesting and exposed a lot of interesting perspectives on many topics.  In short, ASPLOS'14 was everything a good conference should be.

The keynotes and WACI session were excellent, but I especially enjoyed the debate.  Both sides made some thought provoking arguments, and did it in a funny and entertaining way. What's not to like?  Personally, the points that most resonated with me were (1) that scale requires specialization, and cloud computing is shifting computation onto a few very-high-scale systems, (2) that general purpose systems absorb the best specialized techniques, and (3) that innovation historically comes from "abusing" general purpose machines in unexpected ways.  I wonder if we can we have a system that provides a general purpose interface which efficiently uses a specialized implementation...

Organizationally, the conference was very well executed.  Kudos and many thanks to the organizers!  Making everything run so smoothly surely required a lot of hard work and dedication.

In the end, I was left with a taste for more, and looking forward to my next ASPLOS.

ASPLOS 2014 Experiences

This was my first time attending ASPLOS, and based on my experience, I'm very much looking forward to attending future versions. I loved the variety of sessions and papers. It gave a great opportunity to learn about disciplines outside of my own and interact with people from those areas. As an area far outside my own, I really enjoyed the Approximate Computing session with Paraprox and Uncertain<T> (whose name is awesome).

The excursion was an excellent idea and very well executed. Curling was awesome, as was being able to skate on Olympic ice - I'm quite certain I bested Apolo Ohno's lap time. The food served throughout the conference was delicious, and I appreciated the variety that was offered.

Outside of the conference, I enjoyed walking around Salt Lake City and seeing what it has to offer. The city made a great impression on me, and I still can't believe how clean it is (I realize that's a weird thing to point out about a city, but come on! How can so many light coloured buildings and walkways be kept so clean?). The weather was also an invigorating change of pace from the never-ending winter I arrived from.

The ASPLOS experience.

I came to ASPLOS for the first time this year. I enjoyed the sessions which I attended, but I find the most value from conferences in the interactions and connections to be made (while you can read a paper after a conference or look at slides, talking to people face-to-face is a bit more difficult...). I felt that ASPLOS was very conducive of this; and the hotel had plenty of space outside of the conference rooms for people to mingle and chat, and the excursion was at a nice open location with plenty of opportunity to randomly bump into different people.

Of the sessions I attended, I felt that while it was clear that while great effort was put into some presentations, there were a few presentations that were very disappointing and hard to understand. While I am sensitive to the fact that not everyone is a native English speaker, and this problem may have occurred since ASPLOS appeared to be particularly diverse this year, I think that the community at large should engage in discussion on how to improve on the quality of presentations in general. It is difficult to have a productive academic discourse when basic communication is a problem.

Apart from that, I enjoyed that ASPLOS strived to be interdisciplinary, and that there was a healthy mix of students, academics and industry at the conference (in contrast to ISCA, where there were much fewer students). I believe that travel grants really help make this possible, especially for students who do not have a paper at the conference.


This year's ASPLOS was great !!!

This is the second time I have been to ASPLOS. I want to thank and applaud the organizers for organizing such a great conference.

One of the things I like about ASPLOS is its interdisciplinary orientation. Although my research focuses on computer architecture, I always look forward to seeing and following research related to OS and PL.

The quality of the presented work was really high, illustrating the hard work of both authors and the program committee. I really liked the insights of many papers. Although there were two parallel tracks, the lightning session gave a gist of the conference, allowing participants to choose paper sessions to attend. As such, I was able to follow all the papers that were of great interest to me.

I also liked the special sessions (debate and crazy ideas) this year. They were fun, inspiring, and very well organized. Kudos to the organizers and the presenters.

Besides a great research organization, the recreational events were amazing. In particular, the excursion to the Olympic stadium was really cool -- an NBA game would also be great :). I haven't never skated or played curling in my life. Although I was a bit reluctant to try both, I am more than happy that I did. The food was great. The beer was great. Everything was great. All the kudos to the organizers. A big thanks.

Last but not least, I really enjoyed the inspiring and fruitful discussions in the hallway with other junior and senior attendees.
 
I look forward to attending ASPLOS in the future.

Monday, March 17, 2014

ASPLOS'14 Experience

This was my first time at ASPLOS and my first conference as well. I was nervous but also excited to give my first talk in front of so many researches at such a high-standard conference. Thanks for all the useful feedback and comments! 

As a multi-disciplinary conference, ASPLOS offers a great opportunity for gathering ideas from various areas, e.g. architecture, compiler, and os. To learn all these thoughtful and intriguing findings from those experts in that area is really exciting for me.

I really enjoyed the lightning talks, the presentations, and the special sessions("wild and crazy ideas" and "debate”). The lightning talks give great advertisement of the works and make me want to listen to almost every talk, especially talks in Session 1A: Data centers and Session 3B: Heterogeneous computing. Talking with the authors directly makes me better understand their work and the whole area. But I have to say, my favorite is still the special sessions, especially the debate. Before the debate, I was on the side of specialized architecture, but I changed my mind afterwards. 

It is a great experience at ASPLOS’14. Not only I learnt a lot, but also made so many good friends. Thanks a lot for such great organization and opportunity! See you at ASPLOS’15! :)

An ASPLOSion of Ideas

I should start by saying that this was my very first conference, so please forgive me if my observations seem simple to more seasoned conference-goers.  I had a wonderful time at ASPLOS seeing cutting edge research across the system stack, both in areas I was familiar with, as well as those that I know very little about.  I enjoyed that the serious talks were broken up by more light hearted ones, such as the WACI presentations.  Most of all though, I loved thinking about the combination of everything I saw to get an exciting glimpse at what the future may hold for computing.  I see now that this is one of the major benefits of having a conference that brings people from many different research areas together.

I was invited out to speak for the first time at the WACAS co-located event. I found this act of contributing what I've been working on to be extremely rewarding and I hope to do it again sometime with a full conference talk. 

Finally, as an undergraduate trying to figure out whether or not to attend graduate school, I found this conference to be very useful in seeing part of the research world as well as learning from other attendees about differences between institutions.  All in all I had a fantastic trip and I'd like to thank ASPLOS for aiding me in attending.

ASPLOS experience

I have experiences to attend several conferences before, but I can say that ASPLOS was the best! Since I worked on specifically the large-scale and cloud systems area, the conferences that I attended before have focused interests and most people who were publishing and presenting had more concentrated and specified view of research. However, I have seen very active and interdisciplinary conversation between great researchers working on totally different areas in ASPLOS, and it was really exciting!

I had attended most sessions in ASPLOS, and I was surprised of the variety of topics treated by ASPLOS! Throughout from low-level power and architectural layers to operating system and virtualization, there are so many interesting talks and debates among researchers.

Foods and excursion event were more reasons why I was so happy with ASPLOS! High-quality foods and drinks were provided every day for participants, and every participant was able to enjoy many activities such as skating and curling in excursion event. Through these things, we could make a relationship with other researchers and totally enjoy the conference!

ASPLOS 2014

This was my first time attending ASPLOS. I really enjoyed seeing papers from many different areas of computer research - big data, computer architecture, virtualization, etc. This diversity is what makes ASPLOS unique and a great place to network and get ideas.  I was able to get a lot of good feedback on my own work both during the Q&A session after my talk, as well as during the poster session.

One of the highlights of the conference was the WACI session. Although the presentations could have been WACI-er, I thought Michael Wei's talk stole the show and was really in the spirit of WACI. This trip was also my first time in Salt Lake City and I really enjoyed it. It was beautiful this time of year, and the locals were extremely hospitable. The excursion was also unique and a lot of fun. Overall I had an excellent time at ASPLOS 2014.

ASPLOS experiences

I have been to many computer conferences, but it was the first time that I have attended the ASPLOS. It was one of the greatest experiences that I have ever had. There are many things that make the ASPLOS unique. The quality of papers is very high. They are very informative and cover different aspect of computer architecture. The atmosphere of the conference was also very friendly. I made good friends and could expand my network.

I got the chance to present a paper in one of the collocated workshops (WACAS). The feedbacks that the community provided to me was very useful and thoughtful. Not only many of my questions are answered, but also many new and interesting challenges and questions come to my mind.

I also want to suggest one thing that might help to motivate new students to be part of this amazing community. It would be great if we could have a session/event for undergrad students to be able to present their ideas. Usually it is hard for undergrad students to publish in top-tier conferences. This event would provide them the opportunity to be part of this great conference and observe the existing challenges in this area.

At the end, I encourage students to attend and publish papers in ASPLOS. It provides a great experience and shed the light for your future works. Hope to see you soon in the future ASPLOS!

ASPLOS experiences

I’m coming from the computer architecture community and this was the second time I attended ASPLOS. Now I’m sure that ASPLOS is the most exciting conference to attend and publish at, for the following reasons: (1) ASPLOS encourages interdisciplinary research, unlike most of the computer science conferences, (2) it’s an open minded conference that promotes creative ideas, (3) the amount of effort invested in the organization and the review process makes it a top conference.

What made this ASPLOS specifically good was:
1. The poster session - a nice way to clarify any misunderstandings
2. The session with lightening talks was a great idea which I hope will become practice
3. The debate on specialization was exciting and the outcome was fun
4. The ice-skating excursion was an awesome idea
5. SLC is a great place with exceptionally friendly people!
6. The following paper: “ASC: Automatically Scalable Computation”, which I found very creative. Hope to see more of papers like this one.

A couple of things that could have been better:
1. The first keynote would have been interesting had it been delivered 6-7 years ago.
2. It was probably not the best time of the year to be in SLC. A conference like ISCA would be more suitable for this place (it would be more green and less gray).
3. Parallel tracks: this is becoming a standard annoying feature of all conferences. The two sessions I was the most interested in attending (Approximate computing & Datacenters) were fully overlapping with each other.

All in all, this year’s ASPLOS was a fantastic experience. Many thanks to the organizers!

Sunday, March 16, 2014

ASPLOS 2014 -- Salt Lake City

ASPLOS 2014 was the second ASPLOS conference that I've attended, the first being last year's ASPLOS in Houston.  Compared to 2013, this ASPLOS was significantly better across all fronts:  organization, quality of papers and presentations, provided food, hotel room quality, and especially the organized outings. The keynote speakers were not as good as previous conferences, but I really enjoyed the debate panel session. Also, I can't complain much, as the K2 paper from our research group was selected as the best paper!

The best sessions, in my opinion, were the virtualization, heterogeneity, and parallelism. Even the other sessions had at least one paper that attracted me and was relevant to my interests. Despite the high quality of presentations, I found that the physical layout of the two session rooms was very poorly thought out. During almost every session, I was easily distracted by the adjacent room's speaker, whom I could clearly hear through the walls. While this seems like a trivial issue, I would highly recommend that future ASPLOS sessions not be located directly next to each other unless the rooms are more soundproof. The poster session was very enjoyable, although the space was a bit cramped. I thought the quality of each poster was good and that each author was well prepared to answer difficult questions about their works.

Unfortunately, since most of the sessions had several high-quality papers, it was very difficult choosing which session I should attend. I think that ASPLOS is already large enough, as evidenced by the double-scheduled sessions, and I was forced to miss a number of presentations that I wanted to see due to conflicts with other quality works. I would heartily encourage the steering committee to not colocate ASPLOS with any other conferences (like HPCA), because it's already so large that most attendees are unable to see all the presentations they want.

As a co-author of a work published in this year's ASPLOS, I have to commend the PC on their impressively objective review process which resulted in very valuable feedback, both before and after the rebuttal period. I really am honored to have our work be accepted into such a prestigious conference, and I very much appreciate the generous travel grant which made the trip more affordable for me.


How was my experience at ASPLOS 2014? In a word: great. This was my first time attending ASPLOS and I really enjoyed myself. I really liked the fact that this conference attracts people from a wide array of fields. I got a change to meet many interesting people I probably would not have met at a more targeted conference.  This also allowed to me learn about interesting research in areas I don't actively follow.

 I also got a chance to meet many people working in the same field as me. I was a little star struck meeting two of the professors who's work my thesis was based on. I have recently left academia for industry, but attending this conference had me really excited about research again. There was lots of interesting work in my area of interest and plenty of people as excited as me about current developments. 

There was also a vibrant social life at the conference. For me, the biggest highlight was visiting the Olympic Oval. Having grown up in Canada, I'm a little ashamed to say that it was my first time curling. It was really fun and I am seriously considering taking up the sport. I also got a chance to try long track speed skating. While I have done plenty of skating in the past, hockey skates did not leave me well prepared for speed skates. I had a few falls and my ankles were a little worse for wear, but I have no regrets. It was a once in a lifetime opportunity.  I also won't soon forget spending a night at the bar with a group of profs. 

I had lots of interesting discussions at ASPLOS. Some technical: What will be next big thing in computers? As processors get more heterogeneous, will programming them become easier or more challenging? What are the merits of approximate computing? Can you trust any hardware component or piece of software? Some less so: how do we get more women interested in Engineering, and how early do we have to get to them? Should students help review papers? What is the best kind of beer? How bad was this winter in your part of the country?

Thanks to all chairs for putting together a great conference.

Interesting Times at ASPLOS

This was my first experience at ASPLOS and I really enjoyed myself. I come from the programming languages community and having attended conferences like POPL, PLDI, and OOPSLA previously, it was interesting to attend a more architecture-focused event. ASPLOS does very much bill itself as an interdisciplinary conference, and it is eye-opening to see all of the different ways researchers are crossing boundaries between programming languages, operating systems, and computer architecture.

There were several interesting themes that I perceived as prevalent at this year's ASPLOS. One was neural computing, which came up in a number of presented papers, as well as notably the keynote about Qualcomm's Zeroth spiking neural network processor. We got some hints about what programming such devices might be like, and I spoke to some of my PL colleagues about interesting potential avenues for research in this area. We heard a little bit from Qualcomm about programming models, and it seems like program analysis of programs for neuromorphic systems would be an interesting challenge.

Another theme that I perceived was heterogeneous computing, which showed up in a whole host of papers on a myriad of topics, and the debate on the last day of the conference. Talks on accelerators for various purposes, like the Q100 for database processing and the DianNao machine learning accelerator, made me think about the potential role of verification in designing reliable coprocessing hardware. It would be interesting to explore the possibility of building provably correct accelerators, and also to explore augmenting such accelerators with verifiable security or privacy guarantees.

Finally, on a less serious note, I thought the conference was very well organized and enjoyed the chance to see a bit of beautiful Salt Lake City during a spot of downtime and also the planned excursion to the Utah Olympic Oval. As other posters have commented, the presence of Indian food at the banquet was an unexpected and quite welcome surprise! Thanks to the organizers and to everyone on the program committee for making this such a fulfilling conference!

ASPLOS 2014: Curling, FSMs and fun!

An applause for the all the organizers and people who worked really hard to make this happen, and for making this so great. The sheer number of people from a lot of disciplines, the sharing of ideas, and just getting a chance to know whats out there made this a memorable experience.

The curling experience was a lot of failure and a lot of fun! The dinner was awesome, I never expected good Indian food at a conference. I got a chance to interact with some really smart people, and got a lot of useful feedback about my research.

My favorite talk was "Data-Parallel Finite-State Machines" from Microsoft research. The idea of enumeration and convergence is really cool, and I can see a lot of applications and optimizations stemming forth from this.

This was my first ASPLOS, and my first conference as well. It would be wrong to say I wasn't nervous, giving my first talk and being in the company of giants. However, the whole conference turned out to be a really fun experience, and I look forward to ASPLOS 2015 (Istanbul!!)

Intelligent Robots, Uncertain< T >, K2, and the Olympic Oval @ ASPLOS 2014

I enjoyed attending the conference, meeting and talking to new and old connections in both academia and industry. I wanted to share some of the highlights for me and I hope you find them interesting.

Keynote: Neuromorphic Processing: A New Frontier in Scaling Computer Architecture 
Jeff Gehlhaar, Qualcomm
This is one of the most exciting and inspiring talks I've attended. I think Qualcomm's forward-looking research in neuromorphic processing is in line with quite a few research ideas built upon specialization and accelerators. I believe that these are viable solutions to help the architecture community combat important challenges such as the utilization, dark silicon, and enable hardware-software co-design to provide greater efficiency. I also am very intrigued by this talk to explore different application space that spark innovative ideas, such as combining machine learning with sensory devices using a expanded mobile platform. I've already told more than 10 non-computer-geeks about the cool robot video and cannot wait for Qualcomm to reveal more about what they are doing next.

Uncertain< T >: A First-Order Type for Uncertain Data James Bornholt (Australian National Universit); Todd Mytkowicz (Microsoft Research); Kathryn S. McKinley (University of Texas at Austin/Microsoft Research) 
I thought the lightning 90s introduction for this talk did a great job piquing the interest of the audience. I probably would have not wanted to attend the talk without the convincing GPS use cases the authors thoughtfully put together because at first glance, this material is too "mathy" for my taste. However, the authors made it compelling to use this new data abstraction to express and compute using uncertain<t> in applicable applications. 

K2: A Mobile Operating System for Heterogeneous Coherence Domains Felix Xiaozhu Lin (Rice University); Zhen Wang (Rice University); Lin Zhong (Rice University) 
This paper was selected as one of the best papers this year and it is well deserved. K2 is on the track to solve one of the most important problems that the operating systems community is facing, a heterogeneous coherence domain. The presenter was articulate, and was a great story teller, leading the audience from one problem to the next, showing insights at every step. I really enjoyed it even when I'm an architect, and not an OS expert by trade.

The Utah Olympic Oval
Kudos to the event organizers to be so creative and take us to skate on the Olympic-quality ice rink and have staff members who are at the ready to teach us non-Olympians ice skating and curling. I had so much fun learning how to throw the stone, how to sweep, and most importantly, how not to run and fall on the ice during curling! 

Overall, I throughly enjoyed the conference and I hope you did too!


ASPLOS 14 experience


Attending ASPLOS'14 was a delightful experience, and it is hard for me to say so, as a person who usually sees the half empty glass. However, since the papers were interesting, the organisation was faultless, and the special sessions ("wild and crazy ideas" and "debate") were exhilarating and intriguing, I must applaud.

Most of all, ASPLOS was a great opportunity to meet and talk to thought-provoking researchers and get productive feedback. Learning about their point-of-view and interests was of a great use. I was delighted to have a conversation with the acclaimed Carl Waldspurger, who shared some of his vast knowledge with me.

As for the conference topics, ASPLOS was also a great opportunity to get acclaimed with research ideas in various fields. I guess I would not have read many of the papers if I had not attended the conference. Once I attended, I find the relevance of most of the research work to my research. Although most of the papers are in different field than mine, I found in almost each work useful ideas, methods or benchmarks.

Personally, I am sure I could have enjoyed ASPLOS better. Printing the poster at the last moment, and making changes to the presentation instead of going to the excursion (curling) were not the greatest calls I have ever made.

Of course, this blog post would not be complete without a great "thank you" note for the conference PC and ACM for assisting me financially in attending the conference. So, thank you, and I appreciate that you made it possible for me to attend the conference.

Saturday, March 15, 2014

Memorable ASPLOS experience

This was my first time attending ASPLOS, and I was so excited to present my work at such high-standard conference. During the main conference, I was really impressed by the great organization and many other things. I would like to highlight some of my impressions in this post.

The broadness and real-world impacts of the presented work were significant. The papers accepted by ASPLOS may vary in topic and discipline, but their commonality was the real impact and inspiration to the CS-in-general research and industrial practice.

The review process was more effective and objective than many other conferences. I was so impressed by the double-blind multi-stage process that was able to produce rigorous and high-quality reviews. Designing and performing such complex process required significant efforts from reviewers. I would like to express my sincere appreciation to them.

The lightning session and the poster session provided chances to authors to greatly improve the visibility of their work. In addition to the regular presentation, I enjoyed having deep discussions on multiple pieces of work during the poster session.

Finally, the excursion was great! It was my first time playing curling, although I realized I was not good at it :P.

ASPLOS'14

This was my first time at ASPLOS and my first conference talk. I was excited to meet many senior and junior people from different communities and hear about their research. Some of them I have met before and this was a great opportunity to see them again and discuss what is new. My talk went well and I would like to thank everybody for their useful feedback and comments. I also enjoyed other talks and I think I learned a lot.

ASPLOS is really unique because it gathers people from different fields (i.e. OS, PL, Arch.) and with different backgrounds and research interests. This allows attendants to meet and have productive conversations with people working on related (often the same) problems but at different levels and using different approaches. Also, they get the opportunity to learn about certain problems and solutions across the stack they were not aware of before.

Finally, I really appreciated the debate that took place at the end. I think whoever came up with that deserves some kind of an award. :) Hope this will become tradition (if it is not already).

ASPLOS 2014

This year's ASPLOS was one of the most enjoyable conferences that I've attended.  Part of that was a bit coincidental: several ASPLOS attendees this year were either former students of our research group or people that I've met at other conferences and meetings.  As a result, ASPLOS this year felt a little bit like homecoming because I had the opportunity to meet with people that I haven't seen for awhile.

There were several other things, though, that made this ASPLOS especially enjoyable.  I enjoyed the talks (especially the 2A Security and 4A Virtualization sessions, although I might be a little bit biased on Session 2A :) ), and since I am on the job market, it was great to speak with people from both industry and the academia.  The idea of having lightening talks and poster sessions to handle the downsides of having a dual-track conference was clever, in my opinion.  I'm interested in hearing from people to see if it helped them decide which talks to attend.

All of this said, I think my favorite part of the entire conference was the ice skating.  It was an absolutely great time.  Who says micro-architecture and ice don't mix?

Friday, March 14, 2014

ASPLOS 2014 Experience

This is my second time attending ASPLOS, both time presenting a paper at GPGPU workshop. I was impressed by the great organization and high quality of main conference/workshop papers presented. Here are some highlights of the event that I want to share:

-GPGPU7
  It was the second time that I presented a paper on this workshop. The keynote was great that gave a very clear idea about future direction of heterogeneous architectures. I love the idea of HSA and truly thinks this is the future.
  People were really interested in my presentation and asked lots of interesting and challenging questions. They are very useful to advance the work. I've got opportunities to talk to other presenters who provided different perspectives.

-Main Conference
  My main focus on the ASPLOS are works in GPGPU areas. Although they are not many, they are very innovative. Other works may not directly related to my research, but they did provide me with some fresh angles to look at research problems.
  I love the lunch session on the other day when several tables were reserved for women. I've talked to several genius women researchers about their work, one of which was the main leader of a very closely related work of my research.
  The excursion was great! I tried the first time in my life walking (or rather stumbling) on the ice. That was fantastic experience. I've got some help from nice people by holding my hands and teaching me how to control my balance. I got too excited that nearly forgot to eat the delicious dinner!

My experience @ ASPLOS 2014

The ASPLOS 2014 took place in Salt Lake City, Utah from March 1-5, 2014. After 25 hours of flight from Singapore to Salt Lake City, I was expecting myself to be sluggish, tired and jet lagged. On the contrary, I felt otherwise. I was welcomed by an enthusiastic group of people at the registration. I was surprised to receive a fleece as a compliment. It was followed by sessions of papers which enlightened me with immense knowledge and forethought.

The keynotes were well organized. Among the keynotes, I was excited to hear about the information shared by Brad Calder from Microsoft. His in depth knowledge about cloud domain was very informative. The keynote talk on brain computing by Jeff Gehlhaar from Qualcomm  was really innovative and impressive. The lightning sessions were quite crisp and helpful in deciding the sessions to attend. A special mention goes to the debate session. Any debate session would be as interesting as its topic.

This is my first ASPLOS as an attendee / author / presenter. It was genuinely an humbling experience to meet fellow researchers from versatile fields.  The papers presented in ASPLOS were quite knowledgeable. The presenters were very interactive in sharing their knowledge. I was fortunate enough to be to introduce to works on heterogeneous mobile platform (K2), virtualization (KVM/ARM), Cider on android etc,. My presentation was well received and positively criticized. I thoroughly enjoyed the Q&A session after my talk.    

The excursion trip was fun. The oval is a pleasant place to visit and have fun. I would like to take this opportunity to thank all the people involved in organizing ASPLOS 2014. Especially, my sincere thanks goes Prof. Adve, Prof. Rajeev and Prof. Al Davis for their strenuous hard work to make this event a successful one.

My fond memories of ASPLOS'14

That was my first time to attend ASPLOS. Previously I also attended some other top conferences such as OSDI and ISCA, but I find ASPLOS is very special.

First, ASPLOS is a conference with multiple disciplines. It is really amazing to see researchers from Architecture, PL and OS communities to get together, and to communicate with each other. I believe this kind of communication would improve our understanding towards our own research areas, while providing a good opportunity for us to know about the whole software/hardware stack, and to understand the real problems from the reality. This is really important for the current research. For example, our lab is doing research in Non-Volatile Memory. To efficiently utilize these new memories, we need to rethink almost every aspect of a computer system, from the programming model, operating system, to the architecture. Thus, attending ASPLOS is really beneficial to me.

Although I didn't present any paper at ASPLOS, I did give a short talk about my crazy idea in the WACI session. Previously I have never heard of such sessions from other conferences. Actually I was surprised to know that as a top tier conference, ASPLOS encourages and emphasizes on those wild ideas, and it has been a tradition for a long time! From the WACI session, I learned some interesting ideas from other presenters, and I really enjoyed the process of talking on the stage. I really appreciate the chance given by ASPLOS WACI to make me there.

I also find the debating is interesting and impressive. Although the general-purpose team won, I am still a believer in specialized architecture, since I do believe that optimization on some special/common case will gain better performance/resource utilization.

Over all, I did have a great time at ASPLOS'14. The organization of this conference is terrific! The organizers are so considerate to provide us with an opportunity to play curling and skating. I am a good skater, but it was my first time to try curling (and yes, it was much more difficult to play than it seems on the TV). So I really appreciate the efforts those organizers put in this conference. It is you guys who made this great conference happen, so successfully!

Thank you so much for such a good conference and the opportunity for me to be part of it. I am really looking forward to the next one!

Thursday, March 13, 2014

Great ASPLOS experience

This ASPLOS has been well organized. I really like the lightning talks, the debate session and the excursion to curling in the Olympic oval of course. The lightning talks gave me a high-level idea on which talks I want to go. This is very helpful when there are parallel sessions. I would like to see lightning talks in all future ASPLOS.

The debate session is very interesting. It is really fun and inspiring to hear faculty members to debate on interesting research topics. The debate is about the argument: “Specialized architectures, languages, and system software should largely supplant general-purpose alternatives within the next decade”. Before the debate, 47% audience agree and 43% disagree. After the debate, only 35% agree while 50+% disagree. So I guess the reverse side successfully convince people to believe general-purpose hardware will not be supplanted. I actually voted the other way. I voted “disagree” before the debate and “agree” after the debate. I believe specialized architecture will be very useful, but it will not supplant the general-purpose alternatives. But I want to support the specialized architecture and make it pervasive if it is ever possible.

The papers and talks are amazing. I mainly work on the system side. Still, I can find several programming language and architecture works that interest me. And discussing with the authors and people is also a great way to learn their work and how they think.

There is some discussion on whether or not merging ASPLOS with other similar conferences. I disagree with the merge because the merged conference will be too big. It will be harder for us to find and talk to other people. Besides, such merge may harm the system part. I would like to see more system work in ASPLOS and hope ASPLOS keeps its own identity.

Overall, I have had a really good time at ASPLOS. Thanks a lot for such great organization!

ASPLOS'14 trip

I was excited to attend ASPLOS, since it is such a top conference in our field. By attending it, I got chances to talk with some outstanding junior researchers and prestigious senior scholars.  From the conversions, I got direct feedbacks on my research ideas, meanwhile, extended my knowledge by learning interesting ideas from others. It is always a happy process when exchanging ideas with excellent researchers.

ASPLOS is a multi-discipline conference. As a programming system researcher, sometimes I found it was hard to understand the details of some research topics. But I was not bothered too much by that. Instead, it was fun to see some unfamiliar topics from other areas, especially when I realized there were some ideas in common, though they targeted different problems. This also gave me a chance to view my own research problems from a different perspective.

Besides research, the happiest thing to me is getting some nice friends. I felt relaxed when I spent time with them. Hope the friendships can last even after we graduate and join either industry or academia.

My Great Experience to ASPLOS


It was my first time to present my paper at a conference! I really want to thank ASPLOS for recognizing my paper at the beginning and thank my advisor for giving me patient instructions and great supports.

ASPLOS is such a stage and a forum where people from different disciplines present their work and exchange their ideas. I like the fact that it is not just a computer architecture conference.  I learn from different people from different research areas.  It is awesome to know some great research such as k2, scale-out NUMA, etc.

Traveling to Salt Lake City is also exciting. Although as a paper presenter I did not have enough time for travelling around, taking part in the excursion to the Olympic Oval with a group of brilliant people is an experience I will never forget. I learnt how to skate on that day, talked to Prof. Kim from Columbia University, chatted with students from EPFL, met a group of Chinese students...And the food was great! I normally do not eat meat, but meat cooked with Indian spicy was too good to resist.

I gave a good talk on the second day. Thanks my advisor David again! He accompanied my and gave me great supports. After the talk, I talked about the Sharing Architecture with a lot of people. People came for questions, really challenging questions, and even collaborations. I was so excited to know that many people like the Sharing Architecture!

To sum up, it is an experience that I will cherish for a long long time.

ASPLOS experience

I am very thankful to the ASPLOS organizers for providing us with the student grants. I think it’s a generous grant and it definitely helps many students to attend the conference.

I was one of the submissions chairs for the conference. It is really satisfying to see something that you were part of the making actually happen. Most of my work was in the background - setup the submissions website, make sure everything is online all the time, the site is backed up frequently, the submission forms and the review software tweaked as desired, and so on. It was a very interesting experience for me. But the organization of the conference itself is a totally different aspect. I think it was really well done. I agree with the other posts in this regard. The location of the conference was also perfect. Salt Lake is a beautiful city and this being my first visit to Utah, made it a very memorable experience. I very much enjoyed my stay there.

The Keynote talks were really impressive, especially, the one on Neuromorphic Processing by Jeff Gehlhaar. It was interesting to see how Qualcomm is investing in “brain like” computing and I thoroughly enjoyed the talk. This is my second visit to ASPLOS and it always inspires me to meet and talk to so many people from different but related fields. My primary field is computer architecture and I don’t get a similar exposure when attending pure architecture conferences. At least for me, ASPLOS inspires me to think cross-discipline research.


Not sure if David Wood would be reading this. But here’s my alternative to the idea of combining ASPLOS and HPCA. If the community thinks that it is hard to go to HPCA and ASPLOS because they are so close to each other, may be we should think about moving the times of these conferences around a little? In Fall, we have only one deadline (talking from the perspective of an architecture student) - ISCA. Whereas, between May and August, we have three deadlines - Micro, ASPLOS, and HPCA. May be if we space things out a little, we won’t have the problem of combining the two conferences to begin with? There was some discussion at last year’s Micro about the same. Just a thought to ponder.

Wednesday, March 12, 2014

Adwait's report on ASPLOS-2014

This was my 2nd ASPLOS and I was impressed with the quality of the papers and delivery of the talks. I attended GPGPU and Approximate Computing Workshops in addition to the ASPLOS conference.

Workshop/Tutorials
The workshops were conducted very nicely, and I liked both GPGPU and Approximate Computing Workshops

- Both workshops had papers that would interest to broad audience. Even though I don't know much about all the research areas in-depth, I got the high-level ideas of almost all the workshop papers. I would like to thank the organizers of both workshops -- Great Work!

- As most of the works were in their nascent stages, it was nice to talk to authors (and also get feedback on my GPGPU workshop paper), and get their thoughts on potential future works. 

- In GPGPU Workshop, talks/papers from Martin Burtscher (K20 Power) and Nalia (instrumentation of GPGPU kernels) were most interesting to me. Martin's talk was hilarious :-)

- In approximate computing workshop, talk/paper from Benjamin Ransford (Approximate Semantics for Wirelessly Networked Applications) and Mehrzad Samadi (CPU-GPU Collaboration for Output Quality Monitoring) were most interesting. 

Main Conference
- The lightning session helped me significantly in identifying the papers that are of interest to me. 

- Poster Session was also nice -- as I could talk to authors directly and in-depth.

- WACI session was interesting -- Liked the idea of Dirty Caches for Useless (FaceBook) updates. Talk was hilarious. Although not all talks were as WACI as I would have expected. 

- The conference program was very strong. I really appreciate PC Chair and Committee's effort in carving out a strong program. I liked two papers the most : (1) "Disengaged Scheduling for Fair, Protected Access to Fast Computational Accelerators" from University of Rochester, (2) REF: Resource Elasticity Fairness with Sharing Incentives for Multiprocessors from Duke. 
In my opinion, both papers are very neat and can spur lot of future work.

- According to me, talks from Di Wang (Underprovisioning Backup Power Infrastructure for Datacenters) and Seyed Majid Zahed (REF: Resource Elasticity Fairness with Sharing Incentives for Multiprocessors) were most interesting and clear. Conclusion of Di's paper was: "We don't need Diesel generator for backup power". I think this is a very strong conclusion and can have long-term impact. 
Although in Seyed's paper, there was lot of maths, but his talk was very clear and simple, and I could understand almost all of the math he presented in the talk. Looking forward to reading his paper in detail!. 

- Debate: It was very informative, but there was lot of confusion about "YES" or "NO". I guess many people voted in wrong direction :-)

General
- Organization was great. Thanks to Rajeev, and all students of Utah :-). Presence of Indian food in conferences (especially held in US) is probably rare (first time for me). Thanks for it. 

-  The idea of giving fleece jackets was superb. This is the first conference souvenir, which is probably I am going to use for long time. Food during Sunday's reception was a good idea too.

-  ASPLOS was co-located "not-deliberately"with many funny non-technical events :-). It was both good and bad :-)





First ASPLOS Experience

As a first-time ASPLOS presenter / author, I really enjoyed both the conference program and the people. The program was not only strong, but I felt it perfectly exemplified the diverse yet interconnected fields of architecture, programming languages, and operating systems. From automatic, on-demand stats calculation of Uncertain<t> to database processing units; from operating system support for heterogeneous coherence domains to neural network processors and GPU fairness mechanisms, the papers and presentations were top-notch.

As a presenter, I also enjoyed both the feedback during the Q&A session, and the interaction with other authors and attendees during the poster session. People were receptive to my work, and genuinely excited to discuss technical facets of both my work and other author's work.

Although the dual-track made attending every paper presentation impossible, the lightning sessions helped narrow down which presentations I wanted to attend and which posters I wanted to visit.

Finally, the outing to Utah's olympic oval was not only a fantastic idea, but perfectly executed. The skate rental was seamless, the curling lessons were really fun, and the food was incredible! A big thank you to Sarita, Al, Rajeev, and everyone else involved in planning, organizing, and executing this conference!

A few take away points

A few take-away points of ASPLOS 2014, at least in the eyes of a first-time attendee:

1. Quality of publications is indeed top-notch
2. Topic diversity is wilder than you 'd realize by reading the program
3. The lightning-presentation round is very fun indeed, and at least helps set the presentation expectations
4. Dual tracking is just fine; perhaps I was lucky, but I was able to catch almost every talk I wanted to attend
5. There is a great level of energy in the conference; new ideas are born here, I guess new collaborations too
6. The WACI session was fun --- memorable stand-up delivery by Michael Wei
7. I seriously suspect people made a mistake and voted NO when they wanted to vote YES in the Wednesday morning debate of whether heterogeneity is the future. Seriously people. 2 best paper awards on heterogeneous systems, awesome works and presentations in that session (cough cough, I presented there, cough cough) why won't you get the hint?   :)
8. Curling is not as bad as it looks on TV. It is actually fun. It's kind of like bowling, but more fun, and it can actually give you a good workout
9. At the same time, it was great to see snow far in the mountains (nice landscape at SLC by the way) but not in the city. That is the way things should be.
10. Nothing here. Just needed to reach number 10 because I went past 8.

A few follow-up thoughts: 

If you 've never been to ASPLOS, it is totally worth attending --- especially as a student. Though all publications at this level are the product of long and strenuous effort, attending a top-tier conference also brings this work down to earth. As you attend the talks and sync with (other) authors and attendees, your mind is spinning with new or revitalized or clarified ideas. The authors are there, they 'll be happy to brag about their work, you just ask them and get what you need to go back to your drawing board. It is through this process that research moves forward and stays at this high quality level.

Second, organizationally, it is hard to realize the effort invested in making a conference successful until you witness it. The program committee chair, Sarita Adve, explained the reviewing process in detail on Tuesday, and it was crystal clear that a great effort had been put to select the best work. But, beyond that, experiencing a generally stress-free time throughout the conference was possible thanks to an army of people (including students from the University of Utah) that helped with organizational details. Thanks to all of them.

That's about all I can think of. Great work by everyone at ASPLOS14. Looking forward to a repeat :)