RFHPC69: A Hard Look at the Two 2017 Coral Supercomputers

In this episode, the Radio Free HPC team meets at SC14 in New Orleans to discuss the recent news that Nvidia & IBM will build two Coral 150+ Petaflop Supercomputers in 2017 for Lawrence Livermore and Oak Ridge National Laboratories. The two machines will feature IBM POWER9 processors coupled with Nvidia’s future Volta GPU technology. NVLink will be a critical piece of the architecture as well, along with a system interconnect powered by Mellanox.

Henry thinks that IBM will have trouble delivering, and he reminds us what happened when they stumbled and pulled out of the Blue Waters deal at NCSA. Will this be the same story? Rich thinks they will come through, though probably 18 months late. Dan is being the pragmatic one this week. Go figure.

Download the MP3 * Subscribe on iTunes * RSS Feed

RFHPC65: StartupHPC Meetup at SC14

In this episode, the Radio Free HPC team previews the StartupHPC Meetup at SC14.

Does your Startup have ties to High Performance Computing? SC14 will be in New Orleans this year, and we are holding our first meetup on Nov 17th in New Orleans. Please come, meet like minded people, listen to industry notables, and kick off StartupHPC as a support community.

Download the MP3 * Subscribe on iTunes * RSS Feed

RFHPC64: AWS Profitability – Ready for the Cloud Apocalypse?

In this episode, the Radio Free HPC team discuss the possibility of a future where the Big 3 (Amazon, Google, and Microsoft) figure out that Cloud is not profitable and pull the plug. If that Cloud Apocalypse sounds far fetched, check out these recent news nuggets:

  • An article in The Register that describes how Amazon lost money despite $20 Billion in revenue.
  • Business Week says that Amazon’s cloud unit, AWS, is one of the fastest-growing software businesses of all time, as Bloomberg Businessweek‘s Ashlee Vance recently wrote. It’s also in a hotly contested market. Google, Amazon, and Microsoft (MSFT) have been slashing prices to lure customers from one another.
  • The Wall Street Journal says that “The developments point to the possibility that profit margins in Amazon’s cloud unit—believed to be much higher than its online retailing business—face a long-term squeeze that add to investor concerns about the company’s profitability.”

If you’re more into numbers, the team has put together a spreadsheet of AWS revenue that does not paint a rosy picture.

In fact, Henry has got thatAmerican Pie song by Don McLean in his head. What would the new lyrics be? I’m thinking they’d go something like this:

The Day the Cloud Died

A long, long time ago
I can still remember how cloud profit used to make me smile
But I knew if I had my chance
That I could make shareholders dance
And maybe they’d be happy for a while

But second quarter made me shiver
With forward-looking statements I’d deliver
Bad news on the doorstep
Google price-dropped ten percent

I can’t remember if I cried
When analyst expectations lied
But something touched me deep inside
The day the cloud died

So bye-bye, the Cloud had gone dry
Drove my margins to the brink and now my board’s asking why
And them good old boys spit out the Kool-Aid and cried
Singin’ “This’ll be the day that Cloud died
This’ll be the day that Cloud died.”

Download the MP3 * Subscribe on iTunes * RSS Feed

RFHPC63: Paypal Leverages DSPs for Systems Intelligence

In this episode, the Radio Free HPC team discusses a new Paypal project that is leveraging TI Keystone DSP processors for systems intelligence.

Paypal has developed a novel approach to systems intelligence. By analyzing their chaotic real-time server data, they can produce organized, intelligent results using HP’s Moonshot server powered by TI DSP processors.

The challenge confronted here is a data set that must be consumed as soon as it is created in order to yield maximum value. In PayPal’s example, they are dealing with a data set that must be analyzed as soon as it is generated and if this can be done, it could help solve problems in areas such as fraud, risk assessment, forecasting, and business analytic. Just imagine the value of on the fly, real time analysis to businesses when you can identify a pattern and act on it instantaneously. The saying “time is money” is as true as ever and acting quickly on trends exhibited by big data analysis is an untapped resource with tremendous potential. PayPal’s novel approach is to convert events represented in a plain text format into a numeric format which can be analyzed in real-time using mathematical techniques with hardware specifically designed to operate on such numeric data. The first instantiation of this approach uses ProLiant m800 cartridges powered by TI’s 66AK2Hx processor that integrates eight c66x DSP cores and four ARM Cortex-A15 cores using TI’s KeyStone II architecture running in HP’s Moonshot server platform. The 66AK2Hx SoC is ideally suited for this type of application as it possesses some unique advantages to aid in real time processing.

In this video from the 2014 HPC User Forum in Seattle, Arno Kolster and Ryan Quick from PayPal present: Update on HPC at PayPal.

Download the MP3 * Subscribe on iTunes * RSS Feed

RFHPC61: Xennet Free-Market Alternative to AWS

In this episode, the Radio Free HPC team looks at the Xennet initiative, a “public, distributed, and decentralized Supercomputer.” As the brainchild of Israeli computer scientist Ohad Asor, Xennet is essentially a free-market alternative to AWS that sounds a lot like the marriage of BitCoin and SETI@Home.

Read the Full Story * Download the MP3 * Subscribe on iTunes * RSS Feed

RFHPC60: New TPCx-HS Benchmark for Big Data

In this episode, the Radio Free HPC team discusses the new TPCx-HS benchmark for Big Data. Designed to asses a broad range of system topologies and implementation methodologies, TPCx-HS is the industry’s first objective specification enabling measurement of both hardware and software including Hadoop Runtime, Hadoop Filesystem API compatible systems and MapReduce layers.

TPCx-HS is a major achievement in two distinct arenas,” said Raghunath Nambiar, chairman of the TPCx-HS committee, and a distinguished engineer at Cisco. ”TPCx-HS is the first vendor-neutral benchmark focused on big data systems – which have become a critical part of the enterprise IT ecosystem. Secondly, TPCx-HS is the first Express-class benchmark issued by the TPC. Express-class benchmarks are being developed in response to overwhelming demand for a turnkey alternative to enterprise-class benchmarks, which have distinct advantages but are also substantially more time-intensive and costly to run.”

The TPCx-HS benchmark specification is now available online. For more details, you can check out our slidecast interview with Raghunath Nambiar, chairman of the TPCx-HS committee.

Read the Full Story * Download the MP3 * View the slidesSubscribe on iTunes * RSS Feed