TexBlog

Moth on parallel debugging

stevetei — Fri, 13 Nov 2009 12:48:07 GMT

Daniel Moth just posted a great entry to his blog that provides one-stop-shopping for all things VS 2010 parallel debugger, including videos, samples, slides, etc. It’s great to see how much content we’ve built up over the past few months. I recommend you pay The Moth a visit to check it all out!

TechEd panel discussion video available

stevetei — Wed, 11 Nov 2009 17:52:47 GMT

Earlier this week, I sat down with Don Syme, Keith Yedlin, and our moderator, Tiberiu Covaci, for a panel discussion on the impact of parallelism on software developers. Check out the video here.

Meet me at TechEd Europe and SuperComputing09

stevetei — Tue, 03 Nov 2009 00:36:58 GMT

I’m about to embark on a 2-week road trip, speaking at TechEd Europe in Berlin next week and SuperComputing09 in Portland the week following. If you’d like to get together to talk parallel programming, please let me know. I’m particularly interesting in learning more from folks that are considering adoption of the new parallel computing functionality in Visual Studio 2010, developers targeting Windows HPC Server, and also those dabbling in GPGPU development.

In addition to speaking on our upcoming Visual Studio 2010 technologies, I’ll be spending a lot of time at the Ask-The-Experts pavilion at both conferences. If you’re in the neighborhood, please drop in and say, “hi.”

Visual Studio 2010 beta 2 and new MSDN launch today!

stevetei — Mon, 19 Oct 2009 18:31:03 GMT

Visit the landing page for VS2010 and .NET 4.0 to learn more and to download the bits. One of the most exciting aspects of this release is that the quality and completeness have met high enough standards that we have declared it “Go Live” quality, which means you can begin developing production applications with VS 2010 beta 2.

I’d love to hear your feedback on building parallel applications with this release, especially around the completeness of experience between the programming models (PPL, TPL, PLINQ), debugger, and profiler.

We’ve also launched a new MSDN online experience, which included updates to over 11k DevCenter pages across 37 worldwide locales! You might particularly appreciate the new MSDN Library, which includes new “Lightweight” and “Script Free” that enable you to dial your experience based on bandwidth and web browser type.

Enjoy!

Parallel Computing in Visual Studio 2010 beta 1

stevetei — Mon, 22 Jun 2009 23:04:08 GMT

Beta 1 has been available for a few weeks now, and I’ve been seeing a lot of encouraging customer pick-up and feedback. If you haven’t had the opportunity to try it yet, I encourage you to head over to MSDN and download a copy. Once you grab a copy, the VS Beta 1 or Parallel Computing forums are great places to ask questions. You can also use the Connect site to submit bugs or other feedback directly to the product team.

If you’re looking for a few tips & tricks, I’ve gathered pointers to a few resources that might be of interest to you. Let me start with some team members:

Daniel Moth is the Program Manager responsible for the parallel debugging features, and his blog is always informative and sometimes even entertaining. :)
Hazim Shafi is the architect of the parallel performance analysis features and truly the horse’s mouth when it comes to that entire feature set.
The native concurrency team blog is chock full of pithy code snippets and discussion of the finer points of PPL and asynchronous agents.
Not to be outdone, the managed code concurrency team blog has been posting regularly on the details of TPL and PLINQ.
Feng Yuan is a principal developer on the parallel debugging features, and I’m glad to see he’s recently dusted off his blog!
Not strictly VS2010-related, but to get the scoop on one of the hottest incubation projects on the team, check out the Axum team’s blog.

Stretching out beyond our humble team here in Microsoft building 112, here are a few reviews and reactions from the community that might be of interest:

Playing around with VS2010 and the Parallel Extensions
First look at debugging .NET 4.0 dumps in Visual Studio 2010
Work Stealing Queues in .Net 4 and in Previous Versions
…and a better resource list than I could compile myself :)

Have you checked out VS 2010’s parallel computing features yet? If so, I’d love to hear what you think.

Survival of the Nicest… or is it the Meanest?

stevetei — Tue, 20 Jan 2009 04:10:19 GMT

I’m a regular Slashdot reader. Yes, yes, people from Microsoft do read Slashdot; it keeps us grounded to know what is on the minds of the ABM crowd. Slashdot is often an insightful leading indicator of information technology thought and opinion. However, every once in a while I come across a Slashdot article that I find remarkable either for it’s hater-ness or it’s naïveté. This posting, asking whether Nice Engineers Finish Last In Tough Times, is an example of the latter. Whether it’s sloppiness or sensationalism that allows articles like this to slip through the editorial sieve to the home page, I don’t know. But this particular article got my goat for a number of reasons. The first couple of reasons are pretty obvious:

The author obviously has pretty strong opinions about the two managers. Is this really a case of Jim vs. Dwight battling to keep their jobs on The Office, or is there a more subtle reality being masked by the author’s opinions of the people involved?
I’m not a professional statistician, but I feel pretty secure in asserting that a sample size of one doesn’t meet the bar for statistical relevancy. Does this guy’s experience with one instance of office politics really portend some sort of “nice guys finish last” trend in worldwide IT?

But here’s the thing that bugs me most about this article:

The article lacked any relevant data that a second-level manager would actually use to determine who kept their job. For example, which of these two people is more effective at their job, has the most relevant experience, or judged more capable of managing the combined organization would all be important pieces of information for such a decision. We’ve all worked with really nice people that just weren’t very good at their jobs as well as people we didn’t like that we begrudgingly admitted were effective at their jobs.

Of course, we would all prefer to work with people that were both great at what they do as well as total sweethearts, but reality is rarely as simple as the “new trend: bad guys kick sand in face of good guys” caricature painted by this article.

Technical Summit Keynote Video Goes Live

stevetei — Mon, 08 Dec 2008 21:03:33 GMT

The video of the keynote presentation for Technical Summit 2008 in Berlin has gone live. Unless you're fluent in German, you'll want to skip to about 14:55 into the video for my English bit. :)

Parallel Computing Road Trip Debrief

stevetei — Fri, 05 Dec 2008 22:02:00 GMT

I returned last week from a two-week road trip in Europe that involved stops in Barcelona for TechEd Developers EMEA 2008 and Berlin for Technical Summit 2008.

At TechEd I gave two technical breakout sessions, one entitled "Bringing Out the Best in Multi-core Systems," where I highlighted performance engineering and analysis on multi-core hardware (featuring our Visual Studio 2010 technology, naturally) and the other called, "Parallel Programming for C++ Developers in the Next Version of VS," where I focused on Parallel Pattern Library (PPL) and the other upcoming parallel goodies for C++ developers in Visual Studio 2010. There is nothing quite like presenting pre-beta developer tools to keep you at the top of your presentation-giving game. In my case, it's possible that I may or may not have had to tap dance a little when my machine blue screened in the middle of a demo...

What can I say? If it worked right we'd already be shipping it... and luckily there is still a good deal of runway left before we ship it. :)

My colleagues Daniel Moth and Joe Duffy also delivered presentations on parallel computing technologies. In Daniel's case, he did an overview of our parallel technologies for .NET developers, while Joe delivered a deep dive on the Concurrency Runtime. The video of Daniel's talk is available to the public on the TechEd website, but you can only get to Joe's and my talk if you are a registered attendee.

If you're jonesing for video, I did managed to do an interview with the good folks at the Intel Software Network, which you can watch here:

I also did a podcast with those wild and crazy NxtGenUG guys out of the UK (I start about 5:59 into the podcast). To make it clear how I rate with the NxGenUG folks, you'll note that the claim-to-fame by which I am introduced is as "Daniel Moth's boss." :)

After a long weekend off with my wife, Helen, in Madrid, I shuffled off to Berlin for TechSummit. At TechSummit I delivered the keynote presentation as well as a slightly abridged repeat of my "Bringing Out the Best in Multi-core Systems" session from TechEd.

Luckily, no blue screens were seen during the course of these presentations.

In addition to my presentations, I also made time for a briefing with the German IT press and also video interviews with Channel8 and Channel9. Channel8 is geared toward the student community, so in this interview I kind of talk about what I do for a living and riff a bit on interning at Microsoft. In that Channel9 interview I drill into the capabilities of the profiling technology that we are building for Visual Studio 2010. A portion of this video is a shot of my laptop screen, so you'll be able to follow it best if you view the video in full screen mode.

If you can read German or are handy with a web translator, you can check out the blog buzz on Technical Summit here.

Finally, I should also mention that the Parallel Computing Platform team delivered a number of great session as PDC this year. If you haven't had a chance to see the videos yet, you can find them here (filter on "Parallelism").

Parallel Computing: see us at a conference near you

stevetei — Thu, 30 Oct 2008 01:26:24 GMT

I've been getting periodic reports back from the team this week on how things are going at PDC. Customers seem to be very jazzed about parallel computing, our sessions have been enthusiastically received, and my team's debugger and profiler work seems especially popular. The team also tells me that those customers interested in parallel computing are particularly intelligent and attractive people. Okay, I made up that last bit; even the homely and dull-witted are interesting in parallel computing. Okay, I should really stop now. Anyhow, if you're like me and didn't have the opportunity to attend PDC in person this year, there are still a few more opportunities this month to meet the team and see first hand what we're doing for Visual Studio 2010:

Nov 10-14: TechEd EMEA in Barcelona, Spain
Nov 15-21: SC08 in Austin, TX
Nov 19-21: Technical Summit in Berlin, Germany

I'll be speaking at the two European events, so feel free to get in touch if you'd like to chat in person.

Parallel Tools on Channel9

stevetei — Fri, 10 Oct 2008 22:11:58 GMT

Sean Nordberg, Daniel Moth, and I sat down with Charles earlier this week to talk about tooling for building parallel apps and show of some of the things we're working on for Visual Studio. If you'd like to learn more, we'll be presenting at both PDC and TechEd Europe over the next month, and we'd love to see you there! At PDC, in addition to regular sessions, we are also hosting a pre-conference tutorial and a post-conference symposium, so -- as the advertisement goes -- if you're only going to one conference this year... make it PDC. :)

Action-packed video: Steve and Steve on parallelism

stevetei — Wed, 17 Sep 2008 20:31:20 GMT

InformIT just posed a new video (part 1, part 2, part 3) in their OnMicrosoft series in which Steve Toub and I chat with Ted Neward about parallel computing. I just watched some of it, and I was pleased to observe that Steve and I don't look like total dorks. Aside from them repeatedly misspelling my name during the video, I hope you'll find it interesting and informative if not dramatic and gripping. :)

Hear Steve on .NET Rocks!

stevetei — Thu, 11 Sep 2008 20:42:15 GMT

I had an opportunity to chat with Carl and Richard of .NET Rocks! fame recently about the future of parallel computing on the client and some of the near-term technologies we're building on the Parallel Computing Platform team. Listen in and let me know what you think.

Supercomputers in Dresden and Tear Gas in Paris

stevetei — Thu, 11 Sep 2008 20:38:14 GMT

A little preamble: this blog entry may win some sort of infamy award for sitting in my Live Writer drafts folder since late June. Dang, time sure does fly. Anyhow, the irony of sitting on a blog entry for for three months doesn't escape me, but what the hell... I'm posting it anyway.

I was on the road with my colleague, Keith Yedlin, last month for the International Supercomputing Conference in Dresden followed by Parallel Computing Initiative presentations in Paris. Keith manages the sibling product unit to mine, developing the platform and programming model components for Parallel Computing, such as the Parallel Extensions to the .NET Framework, while I manage the Visual Studio tooling that complements new and existing parallel programming models.

At ISC08 we continued to commune with the Supercomputer community and with members of the Microsoft field and product teams focused on Windows HPC Server. There is an interesting continuum in parallel computing... my team occupies mostly the middle portion of that continuum with multi/many-core computing experiences on the client, the high end of the continuum is occupied by HPC/cluster computing, and at the low-power end you'll find handheld devices with multi-core CPUs and systems-on-a-chip. ISC gave us an opportunity to explore the overlapping chunk of the client and cluster ven diagrams and how we can create experiences for developers and users that scale easily between the two. One of the more interesting things I took away from my conversations with customers and industry folks was that the need for sane and lovable programming models exists both on a multi-core client and on individual nodes of a cluster (which are now also multi-core and must deal with shared-memory parallelism on the node).

From Dresden, Keith and I stopped briefly in Paris to deliver presentations on the Parallel Computing Initiative and Parallel Computing Platform technologies to internal-to-Microsoft as well as external audiences. Upon arriving to Paris, and not being sleepy after being on the go all day, we elected to go for a walk and maybe grab a drink. We found ourselves in the park area behind the Eiffel Tower, and, man, it was PACKED! "Cool," we said to ourselves. "Must be a concert going on or something." As we wade through the wall-to-wall young people, we see little fireworks shoot into the air and then observe crowds of folks running away from the arcing fireworks. Except these are pretty meager fireworks; they don't really go bang and they seem to get unnecessarily smokey when they land. As the crowd of fireworks-runners go past, we observe what they're running from: police in riot gear marching forward. And as our eyes and noses begin to sting we realize that what we thought were little fireworks being launched by happy-go-lucky folks in the park are actually tear gas canisters being launched by the stern looking cops in decked out in riot gear. We had stumbled into the middle of a quaint little gathering of students celebrating completion of end-of-year exams just as the local authorities decided it was time for them to go home.

So, yeah... never been tear gassed before, so that was cool.

Anyhow, we managed to find our way back to the hotel without being swept up in the dragnet. Presentations the next day went really well, with lots of excitement about the parallel computing experience we're creating for developers from both the local Microsofties and the external organizations and press that stopped by. Our presentation even earned us some ink in the French press (multi-coeur, how cool is that?) -- ink with a hint of sarcasm in this case, but I totally get that the proof what we're doing will be in the pudding, and it's crucial that we deliver on the promise for making the construction of parallel software dramatically simpler.

The indefatigable EricMitt has posted the videos from our presentation in Paris (you'll be treated with presentations in English if you can make your way through a little bit of French UI).

Parallel Developer Tools

stevetei — Fri, 21 Mar 2008 23:21:00 GMT

Having been on the job for a few months now, I feel like the team and I are finally beginning to settle in. It's been a period of significant flux for the Parallel Developer Tools team that included a new leader and (largely) new team trying to adjust to one another, building a plan for the next release of Visual Studio, recruiting and hiring top talent for the team, physically moving to a new building on a different part of campus (and separate from most of the rest of VS), building and strengthening partnerships with teams internal and external Microsoft to drive the Parallel Computing Initiative forward, and, and, and...

I came across this InfoWorld article today regarding developer skills and parallel computing. I was amused the fact that the issues raised by Dan Reed in the article echoed a conversation I had with Dan just yesterday when I met him for the first time in his office (which happens to be just across the street from mine). It's a fact that developers largely really are not (yet) trained for parallel computing. What's more, those that are trained typically have their experience base in high performance computing (think massive amounts of data being crunched on clusters), which is a slightly different problem than that we're trying to deal with on the PCP team. Our focus in more squarely on leveraging the increasing parallel-capable hardware of a single node, typically a client PC running Windows, and dealing with the complexities of shared memory, making existing software more parallel-scalable, and how we'll design next-generation, inherently parallel software, and -- of course -- how to make parallelism second nature to current and future developers.

By the way, I should mention that I'm looking for the industry's best developers, testers, and program managers to assist us in this effort. Here are some of the current job openings on my team:

Software Development Engineer

Lead Software Development Engineer

Software Development Engineer

Group Program Manager

Software Development Engineer in Test

I'm on PCP (and you can be too)

stevetei — Thu, 20 Dec 2007 03:01:00 GMT

No, not that PCP. The PCP to which I refer is the Parallel Computing Platform, a fairly new organization within Developer Division that is working to unlock new experiences for PC users by making it much, much easier for developers to build applications that leverage the opportunities for parallelism afforded by new multi-core CPUs. Read more on the coming shift to parallel computing and what it means for developers in this Microsoft white paper and Herb Sutter's The Free Lunch is Over article from Dr. Dobb's Journal.

Yes, this means I've moved on from the Visual C++ team. In my new role I'll be serving as Product Unit Manager for the Parallel Developer Tools team within the PCP organization. My team is focused on the tooling developers need to build highly parallel software, such as debuggers, profilers, analysis, and designers.

So how can you be on PCP? You can start by visiting our Parallel Computing MSDN dev center. If you're really ambitious, you can start playing with our Parallel Extensions to .NET 3.5 CTP. Of course, there's more where that came from, and in the coming months you can expect to see things like parallel development libraries and tools for both native and managed code developers.