Thursday, February 14, 2019

Supercomputing to Modernize the Electric Grid

We open the show by talking some weather. It’s so cold at Henry’s house (in Minnesota) that he’s becoming a human superconductor and quantum computing experimenters are showing up at his house to test their systems under uber cold conditions. Shahin adds an inane joke about cold and levitation that Dan threatens to cut out of the final edit of the show.

Our first topic is how Lawrence Livermore National Lab is working to simulate and then help modernize the electric grid. We talk about how the ‘new grid’ will need to be two-way, both delivering and accepting electricity. The new grid will also have to communicate with smart homes and other buildings in order to predict demand and adjust real time pricing.

When the discussion turned to solar power, Henry related the problems of low payouts from utilities to consumers who have installed solar panels. Dan pointed out the current shortfalls in solar power, bringing up an example of the world’s largest solar plants still not generating enough juice to power the NYC subway system. Henry called Dan a dirty liar and an embarrassment to his family. Dan provided the following links to justify his take:
  1. On an annual basis, the NYC subway system uses 1.8 billion kilowatt hour of electricity. This is according to NYCsubway.org. This is 1,800 megawatts of electricity
  2. According to an article published by Origin Energy on 10/24/18, the largest single location solar field is located in India and generates 648 MW of electricity. This is obviously less than the 1,800 megawatts necessary to power the NYC subway. Dan is vindicated.
Next up, we discuss some of the applications that are being run on the Summit supercomputer, the world’s largest system. Some of the applications include exploring the origin of the universe and whole-cell simulation, along with a host of other stuff. Our discussion strays into the recent announcement that scientists in Israel have supposedly cured cancer. This claim has since been debunked, or at least partially debunked…leaving it barely bunked at all.  As the conversation strays even further, Shahin suggests putting a giant mirror behind the sun in order to give us more solar energy. One hell of a good idea.

Catch of the Week

Shahin’s Catch of the Week starts as a mix of buzzwords combined together but clarifies itself (a bit) through explanation. What he’s talking about is a paper titled “Semi-device-independent quantum money with coherent states” that discusses using quantum computing to create unforgeable quantum banknotes and credit cards – definitely a good thing.

Dan’s Catch of the Week is the dust up between Apple and Facebook and how the two goliaths have become embroiled in a slap fight. In the ensuing discussion, Dan coins the phrase “if you’re not paying for an app, it’s a virus.” The gang also points out Facebook’s naiveté (whether it’s real or put on) when it comes to user privacy issues. Dan, warming to the topic of tech giants controlling our lives, brings up the example of Microsoft’s new “Newsguard” browser feature that passes judgement on whether news sites are credible or nor credible. Newsguard is a browser extension that users can activate on new versions of Microsoft’s Edge browser. Here are a few representative discussions about Newsguard and possible implications:  Gizmodo, Publishing Insider, and Breitbart.

Shahin believes that Newsguard is an AI fueled tool, which, upon further research, turns out to be incorrect. Newsguard uses ‘trained journalists’ to review and rate thousands of news and information websites. After a little more desultory conversation, the podcast ends on this disquieting note.

Download the MP3 * Subscribe on iTunes * RSS Feed

Sign up for our insideHPC Newsletter

Monday, February 11, 2019

What's an AI Supercomputer? What's up with software SMP?

We start our discussion by contemplating the fact that Shahin doesn’t have a middle name (he says he never needed one) and touching on why Henry has picked up the nick name ‘Gator’ Newman.

What's an AI supercomputer?

Our first topic is whether a supercomputer can or cannot be a “AI Supercomputer.” This is based on France (along with HPE) unveiling a new AI system which will double the capacity of French supercomputing. So what are the differences between a traditional super and a AI super. According to Dan, it mostly comes down to how many GPUs the system is configured with, while Shahin and Henry think it has something to do with the datasets. Send us a note or a tweet if you have an opinion on this.

Software SMP hits 10k

The guys also discuss ScaleMP and how their announcement of record results, with close to 10,000 customers as of the close of 2018. This led to talk about SMP vs. MPP from a performance standpoint. Henry asserted that a clustered approach will always be superior to a big SMP approach, all things being equal. Dan doesn’t agree and Shahin confesses his love of ‘fat node’ clustering. Dan agrees with Shahin, but wonders why no one is doing it.
We also note that Mellanox got a nice design win with the Finns, as they’ll be installing 200 Gb/s HDR InfiniBand interconnect in a new Finnish supercomputer to be deployed in 2019 and 2020. The interconnect will be used in a Dragonfly topology.

Catch of the Week

  1. Shahin’s catch of the week is a mathematical puzzle titled “The most unexpected answer to a counting puzzle.” Here’s a link to the video.
  2. Dan likes a good comeback story and in light of that, his catch of the week is AMD nabbing a design win at Nikhef.
  3. Henry HAS NO CATCH OF THE WEEK. This makes him the “RF-HPC Villain of the Week” :)
Download the MP3 * Subscribe on iTunes * RSS Feed
Sign up for our insideHPC Newsletter

Friday, February 1, 2019

China Exascale Again (Tianhe-3 is coming), GDPR shows its teeth

After a short talk about the weather in Henry’s basement (it had just reached 60 F by the time we recorded the show), we got right down to business with an important announcement:  our pal Rich Brueckner is leaving the show. He just has too much on his plate and something had to give.

While we’re worried about the impact Rich’s departure might have on our listenership, we did take note of and welcome listeners 13, 14, and 15, who made themselves known to Henry on one of his recent business trips. Yay us.

Our first topic is China rolling out a successor to Tianhe-1, dubbed Tianhe-3. According to news articles, Tianhe-3 will be 200 times faster than Tianhe-1, with 100x more storage. What we don’t know is if these comparisons are relative to Tianhe-1 or Tianhe 1A. The later machine weighs in at 2.256 PFlop/s which means that Tianhe-3 might be as fast as 450 PFlop/s when complete. We also made a reference to a past episode, which we know you remember vividly, where we discussed China's three-pronged strategy for exascale.

As we’re moving into our popular “Catch of the Week” segment, Shahin hijacks the conversation by questioning if anyone knows the real-world utilization rates of non-commodity configurations in public clouds. This leads to this bold estimate from Dan “I’ll bet that there isn’t a public cloud out there that has a higher than 60-65% utilization rate.” We have a spirited discussion about this pseudo-metric and how infrastructures are sized to handle peaks. We also brought up a story that malware can bring down public clouds, although someone would have to own your system before doing it.

Catch of the Week:

  1. Henry hipped us to a website that shows whether your email address or password have been powned: https://haveibeenpwned.com/
  2. Shahin brought up Google’s recent 50 million euro fine for GDPR violations:
  3. Dan discussed the case of a Dutch surgeon who won a landmark case to get her medical disciplinary records removed from Google searches.

Download the MP3 * Subscribe on iTunes * RSS Feed

Sign up for our insideHPC Newsletter