What: Talk by Robert W. JKembel on Unified Data Center Network
When: Wednesday, 21st March, 2012, 5 pm to 8 pm.
Where: Yashada, Rajbhavan Complex, Baner Road, Pune

Registration and fees: This event is *FREE* for all to attend. Please RSVP to amarjeet.sharma@qlogic.com.

I was invited to visit the Nvidia Tech Week this past weekend (February 25-26, 2012) at their facilities in Pune. This is a great concept – getting employees to invite friends and relatives to actually see what their company is all about is very good social outreach and a fantastic marketing initiative. If more tech companies in the area do similar events once or twice a year, it will help lift the shroud of technical opaqueness around them. I think hosting similar events in area colleges will also help students realize that even VLSI/Embedded Systems Design is cool.

I was given a personal tour by Sandeep Sathe, a Sr. Development manager at Nvidia and also met with Jaya Panvalkar, Sr. Director and head of Pune facilities. There was enough to see and do at this event and unfortunately I was a bit short on time. It would have taken a good two hours for a complete walk-through, so I decided to spend more time on the GPU/HPC section though the Tegra based mobile device section was also quite impressive. It’s been a while since I actually installed a new graphics card in a desktop (actually, it’s been a while since I used a desktop), but graphics cards have come a long way! Nvidia is using standard PCI Express form factor cards for the GPU modules with on-board fans and DVI connectors.

The following are key takeaways from the demo stations I visited

GeForce Surround 2-D
Here, Nvidia basically stretches the game graphics from a single monitor to three monitors. Great for gamers as it gives a fantastic feel for peripheral vision. The game actually doesn’t have to support this. The graphics card takes care of it. The setup here is that while the gamer sits in front of the main monitor, he also sees parts of the game in his peripheral vision in two other monitors that are placed at an angle to the main monitor. I played a car rally game and the way roadside trees, objects moved from the main monitor to the peripheral vision monitors was quite fascinating.

GeForce 3-D Vision Surround
This is similar to the above, but with 3D. You can completely immerse yourself in the game. This sort of gaming setup is now forcing monitor manufacturers to develop monitors with ultra small bezel widths. I suppose at some point in the next few years, we will be able to seamlessly merge graphics from different monitors into one continuous collage without gaps.

Powerwall Premium Mosaic
Powerwall is a eight monitor setup driven by the Quadro professional graphics engine. Two Quadro modules fit into one Quadroplex industrial PC to drive four monitors. Projectors can also be used in place of monitors to create a seamless view. The display was absolutely clear and highly detailed. The Powerwall is application transparent. Additional coolness factor – persistence data is saved so you don’t lose the image during video refresh and buffer swaps. This is most certainly a tool intended for professionals who need high quality visuals and computing in their regular work. Examples are automotive, oil and gas, stock trading.

PhysX Engine
PhysX is a graphics engine that infuses real time physics into games or applications. It is intended to make objects in games or simulations move as they would in real life. To me this was very disruptive, and highlight of the show. You can read more about PhysX here. It is very clear how PhysX would change gaming. The game demo I watched had several outstanding effects: dried leaves moving away from the character as he walks through a corridor, glass breaking into millions of shards as it would in real life. Also running was a PhysX simulation demo that would allow researchers to actually calculate how objects would move in case of a flood. What was stunning was that the objects moved differently every time as they would in real life. PhysX runs on Quadro and Tesla GPUs. It is interesting to note that Ra.One special effects were done using PhysX.

3D photos and movies
Next couple of demos demonstrated 3D TV and photo technology using Sony TVs and a set of desktops/laptops. Notably, the Sony 3D glasses were much more comfortable compared to others. Nvidia is working with manufacturers to create more comfortable glasses. There was also a Toshiba laptop that uses a tracking eye camera to display a 3D image to the viewer regardless of seating position without glasses. It was interesting. However, the whole 3D landscape need a lot of work from the industry before it can become mainstream.

Optimus
What was explained to me was that Optimus allows laptops to shut off GPUs when they are not needed. They can be woken up when high performance work is required. This would be automatic and seamless, similar to how power delivery is in on a Toyota Prius. This sort of a technology is not new to computing – a laptop typically puts a lot of components to sleep/hibernate when not being used, but the GPU is not included.

Quadro Visualizations
This allows 2D/3D visualizations for automotive, architectural and similarly complex systems for up to one thousand users at a time. You can easily change colors, textures, views so everyone can comment and give constructive feedback. I was not sure if the design can be changed on the fly as well. Nvidia is working with ISVs like Maya and Autodesk on this.

Tesla
Tesla GPUs use chips that are used for high performance computing and not rendering, which is different from what Nvidia typically does. The Tesla modules do not have any video ports! It has a heterogeneous GPU/CPU architecture that saves power. In fact, the SAGA-220 supercomputer, dubbed India’s fastest, at ISRO’s Vikram Sarabhai Space Center facility uses 2070 Tesla GPUs along with 400 Intel Xeon processors. In addition to supercomputing, Tesla is very useful in 3D robotic surgery, 3D ultrasound, molecular dynamics, oil and gas, weather forecasting and many more applications.

Tegra Mobile Processor
Next few demos showcased the Tegra mobile applications processor based on ARM Cortex A9 cores. The HD quality graphics and imaging were impressive. It is clear that smartphones and tablets of the day are clearly far more powerful compared to desktops of yesteryear and can support highly impressive video and audio in a very handy form factor.

In all, I had a great time. As I mentioned earlier, Nvidia along with other tech companies in Pune should hold more of these kinds of events to give technology exposure to the larger population in general. I think it is important for people to know that the stuff that makes Facebook run is the real key and that is where the coolness is.

Viewed in its rightful context, this is a major achievement. They are competing with Cadence, Calxeda, ST Micro and Xilinx. Please vote for them and create history.

Visit http://www.edn.com/info/2399-2012_ACE_Awards_Ultimate_Products.php to vote.

The $45 Akash

The irrepressible Mr. Sibal has done it again. He thinks that a high powered govt. committee can supervise the manufacture of a mil-spec Akash (Siblet – I have borrowed this name from @comicproject on Twitter) through BEL and ECIL for $45. Let me get this straight – a few current and retired bureaucrats who probably majored in English or Hindi Lit will now tell some folks  at a sarkari company how to design a tablet and get it done. Secondly, he wants the tab to be Indegenized.  This is getting beyond ridiculous to say the least.

Let’s count the reasons:

#1: While DataWind may not be the best manufacturer, why not eject them for someone better?
#2: Do we really thing that BEL and ECIL have the wherewithal to actually build a device that customers will like. Name the last consumer electronics box either of these companies built? The answer is none! Even Videocon would be better suited than either of these. I have closely observed the inner workings of BEL in an earlier life – not pretty! They might turn around and give the contract back to DataWind.
#3: Why MIL-SPEC? What the heck is IIT Jodhpur thinking?
#4: What’s with the $45 price – you can’t even buy a decent phone with it, let alone a tablet.
#5: Why the insistence on Indigenization? The current Akash imports about 70% of components and it is impossible to indigenize it any further unless India starts manufacturing chips, LCD screens and other related electronics overnight.

Bold vision is bold only if supported by a path to getting there. Otherwise, it just looks stupid and megalomaniacal.

This is exactly the reason why India needs its own Sony and disband the departments of science and technology and telecommunications, and others but that is another story!

Semiconductor Manufacturing Plant

I came across a blog written by Deepak Sekar, the Chief Scientist at Monolithic 3D and he makes several interesting points as to how Korea became the De-facto memory hub. The story of Korea in the 1960s and where India is now is uncannily similar. Hopefully, the Indian Government takes lessons from this and formulates a policy that works here. Let’s look at the points Deepak makes and see what could be applicable in the Indian context:

Agrarian Economy

Like Korea in the 1960s, India today is primarily an agrarian economy, in the sense that a majority (70%+) of its populations lives in villages performing farming or related occupations. As the government is tries to move the teeming masses to higher paying occupations, a focused policy to dominate in particular semiconductor manufacturing segments might just be the ticket. Like Korea did it with memories, India could focus on 3D chips or microprocessors or DSPs.

Chaebols and Access to Capital

Chaebols are classical Korean conglomerates often owned by a single family. All old school major Indian businesses have a similar structure – Tata, Birla, Reliance, Mahindra, Videocon to name a few. These business groups for the longest time have been using funds from successful groups companies to start new ventures. It is not inconceivable that they could be goaded or cajoled to start semiconductor manufacturing in a huge way. In fact, Reliance was to enter the Semiconductor manufacturing space, but the plans were shelved after the crash of 2008. If the Indian government as a matter of policy actually gives easy access to capital for forming semiconductor businesses, these Indian Chaebols could step up to the plate and take on the challenge. They certainly have done similar things in the past and this would be no different.

New Technologies

Starting a manufacturing business from scratch requires that you look at brand new cutting edge technologies without any of the baggage of the past. You don’t have to worry whether a legacy technology could be migrated to the new line. Newer technologies with a clean start can allow Indian semiconductor manufacturers to leapfrog external competition that must worry about process node transition and migration. If executed properly, India can take a huge lead in a particular semiconductor space.

Low Cost Manufacturing Location

India is still far cheaper compared to the US, the EU, Japan and even China. Additionally, it has a deep pool of resources who actually know how semiconductors are built. This is an inherent advantage that Korea did not have in the 1960s when they only had lower cost.

In fact, all the ingredients for a perfect storm, low costs, talent, land, capital availability and successful business houses are present. All it needs is policy implementation. Of course, this blog and many others can talk ad nauseam about the Indian semiconductor policy. The question is whether the governments, state of central, act and help bring visions into reality.

This is a PuneChips event, a forum for Pune people interested in semiconductors design/apps/EDA

What: A Talk on PCI Express Architecture and Applications for FPGAs by Kiran Puranik
When: July 30, 2011 from 10:30 am to 12:00 noon
Where: Venture Center, NCL Innovation Park, Pashan Road

PCI Express Architecture and Applications for FPGAs

Modern FPGA devices offer great advantages for designers of industrial imaging, networking, automation and control, data acquisition systems for test, industrial and medical applications. Apart from offering high performance programmable fabric, FPGAs offer a wide variety of IO standards  to interface with networks, motors, sensors, transducers, offer built in high density data storage and the ability to interface to high speed external memory devices. But, most importantly FPGAs offer Gigabit serial connectivity via standards based protocols such as PCI Express^TM. The ubiquitous nature of PCI Express technology enables development of FPGA based plug and play board and card products that interface with standard off-the-shelf embedded compute and communications platforms, running Windows^TM, Linux or other operating systems and custom device drivers. PCI Express 3.0 Architecture offers many reliability, availability and scalability features to address application needs, as well as advanced features such as relaxed transaction ordering, transaction processing hints, optimized buffer flush-fill, active power management to achieve the highest throughput performance possible within the platform’s power and thermal budgets.

About the speaker: Kiran Puranik

Kiran is a Principal Architect at Xilinx, Inc., responsible for serial connectivity protocol products such as PCI Express. He has spent the last 10 years at Xilinx engaged in architecture definition, design, development and verification of Intellectual Property blocks for several generations of FPGAs. Before Xilinx, Kiran held various engineering positions in the field of ASIC, ASSP design and ICCAD software development.

photo credit: Patrick Hoesly

Dr. Ramachandran also presented this at the 2011 VLSI conference. They have posted it online.

Click here to download the PDF.

This event is jointly brought to you by PuneChips and LSI Corporation.

What: Technical Talk by Dr. T.R. Ramachandran on Building and Autonomoous and Scalable Semiconductor VLSI Business
Where: Sargam Auditorium, 4th floor, LSI India Research & Development Private Limited,  T +91 20 4010 4700
When: Wednesday, July 13th 2011, 9:30am-11am. Please arrive by 9:00am for security registration and snack

RSVP: Reshma Arthani:  Reshma.Artani@lsi.com, Mobile: +91.992.320.3557

Abstract:

The presentation focuses on effective ways to build autonomous and scalable semiconductor VLSI businesses. The trends in the VLSI industry and inherent challenges of growth make autonomy & scale-building essential elements of long-term success. This is particularly relevant to emerging geographies like India where there is increased focus on enhancing end-to-end capabilities and overall management.

About the speaker:

T. R. Ramachandran is Senior Director for Product Operations in the Storage Peripherals Division at LSI. In this role, he reports to the Senior Vice President and General Manager of the division and is responsible for the operations infrastructure, business processes, IP and customer program management across the entire product lifecycle from planning through manufacturing ramp for LSI’s highest volume semiconductor business. Before assuming this role, TR held a number of positions in LSI where he brought to bear a unique blend of expertise in a range of areas from business, operations & program management, strategic/competitive analysis, large-scale M&A and business transformations, global product development and deployment, and supplier & manufacturing management. He lives in the United States in Northern California, and is keenly interested in various aspects of technology & broader public policy as well as problems of scale tied to private, public and/or non-governmental sectors.

TR received a Bachelor’s degree in Metallurgical Engineering from IIT-M (Indian Institute of Technology in Madras/Chennai) and is a recipient of the Dr. Dhandapani Prize from IIT-M and the Vidya Bharati Prize conferred by the Indian Institute of Metals. He received his Masters and Ph.D. degrees in Materials Science from the University of Southern California, Los Angeles. His Ph.D. was focused on structural and optical studies of semiconductor thin films & quantum dot nanostructures and innovative forays into nanotechnology using scanning probe microscopes.

photo credit: oskay

What: A seminar on Digital Design and Prototyping with Verilog by Mr. Basu
When: April 28, 2011 from 9:00 am to 6:00 pm
Where: Classroom E, 100 NCL Innovation Park, Dr. Homi Bhabha Road, Pune 411008

This event is not free. The fee is Rs. 1000 per person, which is hardly anything in our opinion. You can get a 25% discount if you bring your own laptop and a further 25% discount with a valid student ID. I want to encourage everyone interested to attend this seminar even though it is not free. Please call +91 20 2590 2984 to register. You can view the workshop details here.

About the Speaker
Mr Basu is a Design Engineer with experience in both Digital and Analog Design using a multitude of EDA and simulation tools. He has strong interests in Embedded systems design and multicore code design. He is a Hobby Robotics fan and entrepreneur in a related field. He has a B.Tech (H) ’03  and  M.Tech ’04  from Indian Institute Of Technology , Kharagpur.

He was the Component Design Engineer for Intel India’s first Multicore project where he also co-authored the Enhanced
Structural Tester Based Functional Test methodology for Intel Multicore processors. He was also the Mixed-Signal Design
Consultant for National Semiconductor’s Sponsored Project at IIT Kharagpur.

He is an accomplished researcher and speaker. A few of his research projects include:

• Behavioral Modelling for Mixed Signal Sytems using Verilog-AMS speeding up simulation times by 1000x
• Analysis of spice simualtion engine for simulation speedup
• Multi-core programming using Message Passing Interface and CUDA

Previously, he has done similar seminars on Behavioral Modelling at IIT Kharagpur

SystemVerilog Guru Cliff Cummings is back in town and he will be holding another seminar on April 19th at the MCCIA auditorium on Senapati Bapat Road from 4:00pm to 7:30pm. Most recently, Cliff was here in November 2009 and this seminar gives a great opportunity for engineers to  re-engage with him. This event is completely free, but registration is required. Please visit this link to register and view the agenda.

 This event is co-sponsored by Qlogic and Cadence who I must thank profusely on behalf of the PuneChips community. It is not very often that internationally renowned experts visit our city and hold free seminars, but QLogic and Cadence have made it possible. So, I encourage everyone who has any interest in SystemVerilog to attend and participate.