Vijay Tatkar's Blog

HPCS/DARPA and Sun: An invitation to a podcast
You have probably seen the news by now that IBM and Cray won the third phase of the HPCS/DARPA contract. Of the three finalists, Sun was the unlucky player not to be invited to play here, despite presenting a strong portfolio. I was asked what my views were, of this news.
Let me start off by saying that I am not privy to any details of the Sun bid, except maybe a few reports from talking to colleagues who have participated in the bid. So I'll give you the somewhat more useful set of pointers that might give you a better clue, first, then I'll opine (in a strictly IMO fashion).

First of all, you might want to check out these blogs within Sun:
http://blogs.sun.com/innovation/

An in-depth podcast, in MP3 is available here. Jim Mitchell and David Douglas talk about life after HPCS. Jim Mitchell is a Sun Fellow and leader of this project; David Douglas is Sun's Associate Director of the HPCS program. [The MP3 program is a 13+-minute podcast]
A summary of their podcast:

• DARPA has given Sun a wonderful leverage to commercialization of technologies from single- to petascale machines
• Sun's involvement in HPC market is much stronger now than it was before this program and that Sun has used the program to advance both Hardware and Software aspects
• SunLabs has used this to achieve not some pure research advancements but to apply these technologies to give Sun products a leg-up
• DARPA gave an opportunity to look at computing at large/massive scale, particularly in software and productivity  and in particular, this will help drive growth opportunity in the industry, not just for compute-intensive sites
• Look at what Fortress brings to the table. See here for more info.  Fortress is a new programming language designed for high-performance computing (HPC) with high programmability.
• DARPA support or not, Sun is continuing the charge towards peta-scale computing.

Lastly, let me add that IMO, yes, it is unfortunate that Sun didnt get the grant. It would have helped this effort tremendously, especially in bringing a sharper focus to this endeavor. But the HPC effort within Sun is at a point of no return; there is great momentum around Sun's HPC programs (Sun has made inroads into Top and Sun has recently announced a slew of new hardware, software and services to enable customers to quickly build up an HPC environment
Sun does have a contrarian view in terms of where this scale of computing is headed. You might want to read Greg Papadopoulos's provocative blog on why "The World Only Needs Five Computers".  Greg P is Chief Technology Officer and Executive Vice President of Research and Development at Sun and is responsible for managing Sun's technology decisions and architecture. His team leads Sun Labs, the DARPA High Performance Computing Systems program, global engineering architecture and advanced development programs.


Posted by tatkar ( Nov 27 2006, 12:12:16 PM PST ) Permalink Comments [0]

Solaris 10 and Sun Studio 11 DVD Kits now available for FREE!
The Free Solaris/SunStudio Media kit is now posted at http://www.sun.com/solaris/freemedia/
The kit includes Solaris 10 6/06 (aka Update1 release) for both SPARC and x86 and contains the Sun Studio 11 Compilers and tools.
Dont wait to download, just sign up and have Sun send the DVD to you directly. Encourage your customers to do the same!

Posted by tatkar ( Nov 22 2006, 10:56:01 AM PST ) Permalink Comments [2]

Sun Studio powers new Opteron Workstation to record SPEC INTrate and SPEC FP Sun recently introduced the so-called AM2 variant (the next-generation of AMD processors) of Sun Ultra40 Workstation. For details, see here. With this introduction, Sun also announced new World Records with this machine. I am particularly happy with this particular one (words from the product page, directly).
The Sun Ultra 40 M2 workstation, with two Dual-Core AMD Opteron model 2220SE processors, has reached a new milestone on the SPECint_rate2006 suite of the SPEC CPU2006 benchmark, by utilizing the most advanced features of Sun Studio 11 software and Solaris 10 OS.
Leading the x86 segment and surpassing competing workstations, the next-generation Sun Ultra 40 M2 workstation produced a SPECint_rate2006 result of 48.8. For a while now Woodcrest had retaken the SPEC INT lead held previously by AMD's Opteron chips. The performance of Woodcrest on SPEC CPU2000 has been particularly spectacular. So it is particularly pleasing to see that with CPU2006 INTrate, Sun has been able to reclaim the World Record here for the dual-core AMD Opteron model 2220SE processors. The rate measure is particularly important as we move into the dual- and quad-core world for the x86/x64 architecture machines.
In addition, Sun Ultra 40 Workstation continues to claim World Record performance for SPEC CPU 2000 FP with best numbers (Peak) of 3545 and a 4-core FPrate of 121. This beats the Woodcrest based numbers, handily, by about 40+%
The following table shows these comparisons:
SPEC CPU2006 INTrate (ratios, higher is better)

System	Description	#Threads	INTrate
Sun Ultra 40 M2	AMD 2220SE (dual-core 2.8GHz), 2CPU	4	48.4
	SuperMicroWoodcrest, Intel 5160 (dual-core 3.0GHz), 2 CPU	4	45.2
Dell Precision 380	Intel 3.73 GHz, Pentium Exteme Edition 965	4	23.1

SPEC CPU2000 FP(ratios, higher is better)

System	Description	#Threads	FPrate
Sun Ultra 40 M2	AMD 2220SE (dual-core 2.8GHz), 2CPU	2	121
Dell Precision 690	(Xeon 5160, 4cores, 2 chips, RHEL 4AS U3)	2	81.3

Required Disclosure Statements:

SPEC, SPECCfp and SPECfp Rate are Registered Trademarks of Standard Performance Evaluation Sun's results were submitted for review. For SPEC comparisons, socket equates to chip.
Competitive results from www.spec.org as of Nov 17, 2006.

Sun Ultra 40 M2 (AMD Opteron model 2220SE, 4 cores, 2 chips): SPECINT_rate2006: 48.4
Dell Precision 690 (Xeon 5160, 4 cores, 2 chips, RHEL 4AS U3): SPECint_rate2006 - 45.2

Sun Ultra 40 M2 (AMD Opteron model 2220SE, 4 cores, 2 chips): SPECfp_rate2000: 121
Dell Precision 690 (Xeon 5160, 4 cores, 2 chips, RHEL 4AS U3): SPECfp_rate2000 - 81.3

Sun Ultra 40 M2 (AMDOpteron model 2220SE, 4 cores, 2 chips): SPECfp2000 – 3545
Dell Precision 690 (Xeon 5160, 4 cores, 2 chips, RHEL 4AS U3): SPECfp2000 - 2872




Posted by tatkar ( Nov 17 2006, 04:31:45 PM PST ) Permalink Comments [1]

SunTech Days in Korea
I was at SunTech Days in Korea for a SunStudio presentation ( you'll find a copy here right now and later on a the SunTech Days Korea site as well).
This was my first trip to that country and it was absolutely wonderful! I didnt really know what to expect, so I was pleasantly surprised to find

• State of the art airport facility at Incheon, Korean
• Easy checkout, baggage retrieval, airport to hotel shuttle facility, etc. The whole experience of getting to the hotel in the city (Seoul) which is about an hour or two (depending on time of day) was unbelievably easy.
• A very professional business center which hosted the event and the surrounding blocks around it.
• A very superb, underground shopping mall featuring over 800 stores.
• Everything was just so spic-and-span: the city streets in the business district, the underground mall, etc
• An unbelievably courteous population.

The attendance on day 1 was over 1200 (I didnt stay all throughout the second day to note how it was on that day) and we had three tracks: a Solaris development track and two Java development tracks. The Solaris track had generally around 200 or so attendees. Jim Hughes was the first speaker, doing a Solaris Keynote (you can find his presentation here). I was second, and I talked about SunStudio highlights, roadmap and strategy behind the features we are developing in compilers and tools. Both of us stressed on Whats unique in Solaris and SunStudio . In particular, I liked Jim's theme that Different is not necessarily Better, but Better is always Different and used this theme to highlight SunStudio's synergy with Solaris and systems design to add overall value to Sun. On a personal note, I was very pleased with my own presentation (a rarity, because I always come out thinking I should have said something more as well); additionally I got in two demos as well and the overall feedback seemed very positive.
The same day, we also covered the Pod for SunStudio and Solaris (for Solaris Development) as much as we could, between the three of us. There was a steady stream of curious attendees streaming by and I talked to perhaps between half- to a dozen engineers, who seemed fairly knowledgeable with our tools (a bit of a surprise to me) and with compiler technology in general. Not speaking the local language clearly got in the way of being able to communicate with others; we had Korean speakers at the POD that helped out.
On day 0, there were also three tracks: an OpenSolaris track, a Java ME track and a Netbeans track. From what I could make out, it was fairly well attended as well, tho I couldnt make out how big the attendance was, on that day.
All in all, I was very happy about the trip and it was a good insight into some developer thinking in Korea; the next time around, I will keep a day for sight-seeing. That was the one thing I missed doing this time around!

Posted by tatkar ( Nov 16 2006, 04:43:53 PM PST ) Permalink Comments [0]

Two more reasons to move to Solaris
This Microsoft deal with Novell announcement to support SuSE users to jointly develop interoperability solutions (and co-market them) and a patent covenant ie. a guarantee not to sue SuSE customers and developers is an interesting one.
This has generated some interesting buzz and reaction as in here.
This one comes on top of the Oracle deal to support RedHat customers with a more competitive support offering. This earlier announcement has generated a flurry of opinions some of them captured here and in this BusinessWeek report.
These two deals actually favor Solaris. Why? Consider this:

• It re-emphasizes one of Sun's core values that Operating Systems still matter in the choice of Entreprise Deployments
• Go ahead, create more forks: SuSE, Oracle and Redhat! Whereas more forks is good in some markets, it definitely is an adoption barrier when eg. Solaris offers years (7+ and counting) of binary compatibility. You've just validated Sun's other core proposition around binary compatibility
• The discussion refocuses on who owns IP and who supports it. Sun provides a uniquely popular opportunity with over 3000 applications committed to Solaris 10, 6M+ downloads and a bigger ISV pool than any of these forks individually. Sun is the only company in this lot that owns it own IP, will make it available for free, will protect its customers from litigation and play into the OpenSource community as well as any Linux vendor can.
• Sun has a co-operation/interoperation agreement with Microsoft already in place. This just makes the current announcement a "been there, done that" deal.

Frankly, I see this as opportunity for Sun to make greater inroads with Solaris 10, especially in the x86 market. Sun continues to play in the open source community in a very open, transparent manner, embracing and leading the movement with a license (CDDL) that is freer of restrictions than GPL.
As a Sun employee, I see this as increasing validation of Sun's approach to Solaris.
Posted by tatkar ( Nov 07 2006, 03:48:54 PM PST ) Permalink Comments [0]

I am off to present at SunTech Developer Days in Seoul, Korea
Next week I will be travelling to Korea to present at the SunTech Developer Days in Seoul. This is my first visit to Korea and I'm looking forward to presenting there. The challenge is not of presenting the material- which I've done about 25 times last year to various Sun customers etal- but it is in presenting it to this audience which does not fit the profile of any of my previous presentations. Also, there will be the challenge of having these talks translated simultaneously. More to report on that when I get back.
It should be a lot of fun! I heard the crowds there were much larger than similar Developer Days here in the US.
My technical presentation isnt quite ready for prime time yet... there are always last minute changes I want to make but the presentation should be some small variant of this one posted here.
Posted by tatkar ( Nov 03 2006, 06:04:55 PM PST ) Permalink Comments [0]