Follow Slashdot stories on Twitter


Forgot your password?

Comment Re:56 gigabit InfiniBand (Score 1) 55

I know about this...I worked on the SilverStorm's Fabric Manager while I was there. I remember going into the VT System-X room and seeing piles of bad cables from earlier setup. If I remember correctly, the very first network had more switch ASICs than hosts...both were around 2000 or so. I think the first switches used 8-port ASICs internally. We made massive improvements to our fabric scan time and reaction time to moving cables, nodes going down, etc. This was a good thing because the non-silverstorm IB switches that were there were at the start were having a failures all the time. I believe System-X eventually moved to SilverStorm IB switches (those 288s and 24 port switches). That 288 was fun to work on. Moving to 24-port ASIC based switches really cuts down on fabric scan and setup time.

Comment 56 gigabit InfiniBand (Score 1) 55

They claim they will use 56 gigabit InfiniBand. Has anyone tested Mellanox's FDR adapters and switches? From what I understand, that is 14 gigabit over 4x cabling. I remember all the problems just getting 10 gigabit to work over 4x 2.5 gigabit copper. I imagine this must use fiber to get any distance from the server to the switch.

Their asic seems to support only 36 ports. Building a 2000 node network with 36 port switches will take a lot of interconnected switches. I wonder what topology they are going to use. Is anyone building bigger switches based on many interconnected 36-port asics?

Comment what about engineering? (Score 1) 694

According to this article: engineering jobs are doing quite well. This seems to suggest it might even be sexy!

It seems the first article in the main post talks about STEM but doesn't provide any real evidence but may be talking about an article noting the brain drain of the finance world and that the percentage of MIT grads going into finance increased. It was not the majority. This isn't a bad thing. I am a software engineer with a BS in computer science and an MS in computer engineering. I worked on financial software, device drivers, internet advertising software, remote control car embedded software, and wireless meshing software. STEM can take you into all sorts of industries. I don't know why this is a bad thing and I don't think that was what the President was saying.

The author says, "First, American culture has always realized this 'stuff' is important." I find that hard to believe. When I was in school, there were almost no computer science majors. My graduation had 3 CS majors walk with me with over 10,000 students enrolled at the university. Other STEM majors did a bit better but the reality is that our society looks down on STEM folks. It is considered weird and odd. Hollywood uses this stereotype often because the public believes it. We are geeks, nerds, dorks, etc. STEM really does need better PR since keeping the world running doesn't seem to matter much anymore.

The second article attacks pure science jobs. Most folks that major in STEM probably do not go on to pure research jobs. Why is this a bad thing? Many folks become engineers and use the research to create other things. They compare school teachers to scientists even though many are both. Many "pure" researches are college professors. "The women I know who are university professors, by and large, are unmarried and childless. By the time they get tenure, they are on the verge of infertility. " My wife is a university professor and we have two kids thank you. I know several other professional female mathematicians who married and had kids. It seems they these folks get married at the same rate as most other folks and have kids at about the same rate too. Also, the author ignores that the PhD can go into public school teaching and will be paid extra for having the degree. They can leave college and go into the private sector. Some do both. My wife programmed for a while and then came back to teaching because she liked teaching a lot more.

"men tend to lack perspective and are unable to step back and ask the question 'is this peer group worth impressing?' " That is an incredibly sexist statement. Does the author have any proof that women do not do the same? As far as I can tell the sentence should be changed to "people sometimes lack perspective..."

"When Albert goes to graduate school to get his PhD, his choice will have the same logical foundation as John Hinckley's attempt to impress Jodie Foster by shooting Ronald Reagan. " The author compares getting a PhD to being a stalker. This is why STEM needs better PR! Americans see most PhDs as potential stalkers who are just trying to impress the other potential stalkers. Maybe someone gets a PhD because they are actually really interested in the field and really want to study it for its own sake. Maybe someone cares about something other than money. Maybe there is more to life satisfaction than just making more money. Do you want a job that pays 20% more that you will hate or a job that you will love? Perhaps the PhD candidate is actually smarter enough to do what he/she loves.

"What about women? Don't they want to impress their peers? Yes, but they are more discriminating about choosing those peers." Really? There are many pieces of media about who girls get caught up trying to impress the wrong peers. Are they all wrong? I have female friends and relatives that have fallen into this trap just as much as my male friends and relatives.

"If you are extremely introverted, you might prefer to work as a computer programmer. " I work as a computer programmer and I have developed many great friendships through this field. I met my wife through a fellow programmer. I have friends that have been there for well over a decade through my career as a programmer. Contrary to popular belief, computer programmers do have to talk and work with other people, even non-programmers. I have work with folks in documentation, testing, marketing, sales, accounting, finance, etc. I find that introverted folks are only good at a very limited subset of programming. Most complex problems require teamwork.

"For whatever reason we've decided that science in America should be done by low-paid immigrants." What? 1. What is wrong with immigrants? 2. Who says that immigrants are the majority of scientists? 3. Who even cares? I have worked with folks from many corners of the earth and been richer for it. It has not hurt my career and in fact has helped it as I have seen things from very different points of view. If someone is going to try to outwork me for less money, than I will just try to make myself more valuable to justify earning more than them. Life is competitive. Get over it when someone from another country out works you. Prove that you deserve your salary but being more productive than the other person.

"Imagine if one of those kind souls that Summers was speaking to had taken Condoleezza Rice aside and told her not to waste time with political science because physics was so much more challenging. Just think how far she might have gone... " She seems smart. Perhaps she would have discovered something that would have changed the world for the better. Perhaps should would have changed our view of the universe itself. But perhaps she would not have made as much money. Perhaps that isn't always the most important thing.

Comment GPU to network (Score 1) 175

I wonder when a GPU will be able to directly access a network of some sort. Right now, you would need glue code on the CPU to link multiple GPUs in different systems together. I imagine that some HPC applications would run quite well with 100 GPUs spread over 25-50 machines with a QDR InfiniBand link between them.

Comment what is the interconnect (Score 1) 1

I wonder what the interconnect will be? An earlier Riken system uses lots and lots of 4X SDR InfiniBand (10gbps) . They used lots of 32 port IB switches from InfiniCon Systems (I worked there at the time). Now IB is up to 12X SDR (30 dbps) 4X QDR (40 gbps). I wonder what IB company will get the deal. InfiniCon become SilverStorm and is now owned by QLogic. There are also Mellanox, Voltaire, and Sun.


Submission + - What Britney Spears Can Reveal About Alzheimer's

Hugh Pickens writes: "Time Magazine reports that researchers at the Cleveland Clinic report that they may have found a way to identify those most at risk of developing Alzheimer's long before symptoms develop — simply by asking them whether they recognize celebrities such as Britney Spears and Johnny Carson. It turns out that when people who are at highest risk of Alzheimer's try to recognize a famous name, their brains activate in very different ways from those of people who aren't at risk and scientists can actually see this difference using functional magnetic resonance imaging, or fMRI. A team led by Stephen Rao, a brain-imaging specialist, found that those who were at the highest risk of developing Alzheimer's showed high levels of activity in the hippocampus, posterior cingulate and regions of the frontal cortex, all areas involved in memory. "This pushes the envelope further in attempting to detect dysfunction in the brain at a stage earlier than any detectable clinical measurement of cognitive decline," says Dr. Ralph Nixon. The idea is not necessarily to diagnose Alzheimer's earlier but to help identify those most vulnerable to cognitive decline so they can participate in clinical trials of new drugs designed to postpone or reduce symptoms. "If we can delay the onset of Alzheimer's by five years," says Rao, "by some estimates we can cut the incidence of Alzheimer's in half. If we can delay the disease by 10 years, we could almost eliminate it because people would die from other conditions first.""

Submission + - Nvidia predicts 570x GPU performance boost (

Gianna Borgnine writes: "Nvidia is predicting that GPU performance is going to increase a whopping 570-fold in the next six years. According to TG Daily, Nvidia CEO Jen-Hsun Huang made the prediction at this year's Hot Chips symposium. Huang claimed that while the performance of GPU silicon is heading for a monumental increase in the next six years — making it 570 times faster than the products available today — CPU technology will find itself lagging behind, increasing to a mere 3 times current performance levels. "Huang also discussed a number of 'real-world' GPU applications, including energy exploration, interactive ray tracing and CGI simulations.""

Submission + - Fujitsu SuperComputer With 10-Petaflop Performance 1

An anonymous reader writes: Fujitsu is currently developing a new supercomputer which will be able to achieve 10-petaflop performance — ten times faster than today's most powerful supercomputer. The new system configuration with a scalar processing architecture makes use of the world's fastest CPU (Fujitsu SPARC64 VIIIfx with 8 cores, 128 gigaflops), and will be unveiled by 2012.

Submission + - Holiday shipping statistics

Meostro writes: It may seem that your house has become a shipping hub for the past few days, with holiday goodies being delivered daily or even several times a day. However, what you see is infinitesimal versus the actual volume of stuff that goes through the major shipping and logistics networks daily. UPS, FedEx and DHL all have holiday statistics available, and they're impressive.

For example, the FedEx website will see double their average number of tracking requests, at nearly 6 million per day during the month of December. UPS will deliver about 243 packages every second (for a full 24 hours) on their peak day, December 20th.

Submission + - FCC won't release cell carrier reliability data

imuffin writes: MSNBC is reporting that the FCC has been collecting data on the reliability of different cell phone carriers in the US. This data could be invaluable to consumers trying to choose a company to sign a lengthy contract with, but the FCC won't release the data to consumers, citing national security risks.

From the article:
A federal Freedom of Information Act request for the data, filed in August by, has been rejected by the agency. The stated reasons: Release of the information could help terrorists plan attacks against the United States, and it would harm the companies involved.

Submission + - HP's Bundle Trouble

narramissic writes: A French consumer group has filed 3 lawsuits against HP, saying the company's practice of selling consumer PCs with Windows pre-installed violates a French law that 'prohibits linking the functionality of a product to another product' — not to mention that consumers wind up paying for an unwanted OS. For its part, HP contends that it is not in violation of the law because the OS is integral to the PC. 'The PC without an OS is not a product because it doesn't work,' said Alain Spitzmuller, legal affairs director for HP France. 'We believe the market is for products that work.'

How Skype Punches Holes in Firewalls 215

An anonymous reader writes "Ever wondered, how P2P software like Skype directly exchanges data — despite the fact, that both machines are sitting behind a firewall that only permits outgoing traffic? Read about the hole punching techniques, that make a firewall admin's nightmares come true."

Slashdot Top Deals

1 Mole = 25 Cagey Bees