Friday 14 November 2014

Microwulf Is The World’s Cheapest Supercomputer, A Personal, Portable Beowulf Cluster

Microwulf Is The World’s Cheapest Supercomputer, A Personal, Portable Beowulf Cluster

If there ever comes a time when you find yourself in the need of a supercomputer, would you fancy making one for yourself or will you spend insane cash on purchasing one? We know we’d go for building it on our own. We know one person who’d agree with this approach; the same guy who has created this supercomputer known as Microwulf that has qualified for the cheapest supercomputer ever build.Microwulf – DIY Cheapest Supercomputer2
The system is a Beowulf cluster that is capable of running at 26.25 Gigaflops and costs $1,256. This may not qualify for a budget computer, but for budget supercomputer it sure is heck of a solid candidate. The figure of 26.25 Gigaflops is insane when you look at it from price/performance ratio.Microwulf
Sun’s Spart Enterprise M9000 Supercomputer costs $511,385 and is capable of working at 1.03 Teraflops, breaking this in dollars per Gigaflop tells us that M9000 costs $496 per Gigaflop whereas Microwulf costs $48 per Gigaflop.
You can find out how to make your own mini-supercomputer by checking Cluster Monkey out!

Microwulf: Design


Microwulf is designed to be a cost-efficient, high performance, portable, "personal" Beowulf cluster. The basic idea is to pack a lot of processing power into a small volume using multicore CPUs.
Microwulf design To do so, we use motherboards with a smaller form-factor (like Little Fe) than the usual ATX size, and we space them using threaded rods (like this cluster) and scrap plexiglass, to minimize "packaging" costs. By building a "double decker sandwich" of four microATX motherboards, each with a dual core CPU and 2 GB RAM (1 GB/core), we can build a 4-node, 8-core, 8GB multiprocessor small enough to fit on one's desktop, powerful enough to do useful work, and inexpensive enough that anyone can afford one.
Since our microATX motherboards have an on-board Gigabit Ethernet adaptor, that is the least expensive way for the processors to communicate. To keep the two cores from competing for this adaptor, we add a second Gigabit Ethernet adaptor in each motherboard's PCI-Express slot. We then rely on Open MPI (see below) to spread the communication load across these two adaptors. Then we connect all the adpators via an inexpensive 8-port Gigabit Ethernet switch. This provides a Gigabit Ethernet link's worth of bandwidth for each core.
The bottom motherboard acts as the "master" node, which is configured to boot from Microwulf's single hard disk (and/or DVD-ROM drive). The other three motherboards are configured as "server" nodes, and boot from the network using PXE.
The following schematic diagram shows the interconnections between Microwulf's components:
Microwulf schematic
At present, Microwulf is running Ubuntu Linux.

Microwulf Pictures

Tim Brom and Microwulf
Tim Brom and Microwulf
Microwulf "west" view
Microwulf "west" view
Microwulf "southwest" view
Microwulf "southwest" view
Microwulf "south" view
Microwulf "south" view
Microwulf "southeast" view
Microwulf "southeast" view
Microwulf "north" view


Microwulf: Cost Efficiency


When you have measured a supercomputer's performance using HPL, and know its price, you can measure its cost efficiency by computing its price/performance ratio. By computing the number of dollars you are paying for each floating point operation (flop), you can compare one supercomputer's cost-efficiency against others.
With a price of just $2470 and performance of 26.25 Gflops, Microwulf's price/performance ratio (PPR) is $94.10/Gflop, or less than $0.10/Mflop! This makes Microwulf the first general-purpose Beowulf cluster to break the $100/Gflop (or $0.10/Mflop) threshold for measured double-precision floating point performance.
For comparison purposes:
  • In 1976, the Cray-1 cost more than 8 million dollars and had a peak (theoretical maximum) performance of 250 Mflops, making its PPR more than $32,000/Mflop. Since peak performance exceeds measured performance, its PPR using measured performance (estimated at 160 Mflops) would be much higher.
  • In 1985, the Cray-2 cost more than 17 million dollars and had a peak performance of 3.9 Gflops, making its PPR more than $4,350/Mflop ($4,358,974/Gflop).
  • In 1997, IBM's Deep Blue defeated world chess champion Gary Kasparov. Its price has been estimated at 5 million dollars, and it produced 11.38 Gflops of measured performance, making its PPR more than $439,367/Gflop.
  • In 2003, the U. of Kentucky's Beowulf cluster KASY0 cost $39,454 to build, and produced 187.3 Gflops on the double-precision version of HPL, giving it a PPR of about $210/Gflop.
  • Also in 2003, the University of Illinois at Urbana-Champaign's National Center for Supercomputing Applications built the PS 2 Cluster for about $50,000. No measured performance numbers are available; which isn't surprising, since the PS-2 has no hardware support for double precision floating point operations. This cluster's theoretical peak performance is about 500 Gflops (single-precision); however, one study showed that the PS-2's double-precision performance took over 17 times as long as its single-precision performance. Even using the inflated single-precision peak performance value, its PPR is more than $100/Gflop; it's measured double-precision performance is probably more than 17 times that.
  • In 2004, Virginia Tech built System X, which cost 5.7 million dollars, and produced 12.25 Tflops of measured performance, giving it a PPR of about $465/Gflop.
  • In 2007, Sun's Sparc Enterprice M9000 with a base price of $511,385, produced 1.03 Tflops of measured performance, making its PPR more than $496/Gflop. (The base price is for the 32 cpu model, the benchmark was run using a 64 cpu model, which is presumably more expensive.)
At $94.10/Gflop, Microwulf is by far the most cost-efficient platform available today for high performance double-precision computation. While it may not provide Tflop performance, it provides more than twice the general-computation performance of Deep Blue. Microwulf thus offers significant computational power at a highly affordable price.

Saturday 1 November 2014

Hendo Hoverboards - World's first REAL hoverboard

Hendo Hoverboards - World's first REAL hoverboard

Hendo is introducing the world's first REAL hoverboard and hover developer kit. We are putting hover technology in YOUR hands.
So where does the HENDO hoverboard stand today? Well, about 1 inch off the ground. As you can see from the video above, the prototype is real and it works! But to see it hover in person, and better yet, to defy gravity by riding it, is something you need to experience as well.
With the support of the Kickstarter community, we all can. We need your help to put the finishing touches on the Hendo Hoverboard, to help us produce them, and to create places to ride them. 

Ok, so you can't shred a halfpipe with one of these, but you can still show your support for Hendo with one of our Betaboards instead. The Betaboard is a smaller version of our hoverboard that floats on a magnetic base. It uses a a different type of levitation than our technology, based on electromagnetic stabilization.  It cannot be moved around too quickly and needs to stay plugged in, but it's perfect for any desk and makes a great gift!
(In theory, the Betaboard should float forever, as long as it's plugged in and not disturbed too much. It holds about 8 ounces/200g.)
Our engineering team has been amazing, rapidly iterating on design after design. In fact, this our 18th prototype, and we continue to make advances week after week.
 
The magic behind the hoverboard lies in its four disc-shaped hover engines. These create a special magnetic field which literally pushes against itself, generating the lift which levitates our board off the ground.
While our hoverboard is primarily intended to be self-propelled, the actions which stabilize it can also be used to drive it forward by altering the projected force on the surface below.
Currently, this surface needs to be a non-ferromagnetic conductor.  Right now we use commonly available metals in simple sheets, but we are working on new compounds and new configurations to maximize our technology and minimize costs. 
The hoverboard is simultaneously fascinating and exhilarating. The enabling technologies existed, but no one had yet been able to align them to bring a hoverboard forth. Hendo has done so, and our hoverboards are working in almost every way we could have imagined. But perfecting it will take a little more time and resources. 
Until the hoverboard is within everyone's reach, we are offering the same technology in a small, accessible form factor - the Whitebox™ Developer Kit.
SPECS: (Approximate)
  • Dimensions: 10" x 10" x 5" (25cm x 25cm x 12cm)
  • Hover Height: 1/4" - 1/2" (1cm - 1.5cm)
  • Battery Life: 10-15 min
  • Charge Time: 1-2 hrs
  • Weight: 10-12 lbs (4.5 kg)
  • Payload: ~ 5 lbs (3 kg)
We kept the Whitebox as simple and affordable as we could to get our technology out into as many hands as possible.
It is designed to be explored, taken apart, and analyzed, encouraging you to dare to wonder.
https://www.youtube.com/watch?v=jCvvQLnqpR4&feature=player_embedded&list=UU_qM32jvFjo08W9UfGhbeWg
The Whitebox+™ puts hover control and propulsion capabilities in your hands. It's operated through your mobile device (iPhone/Android). This control functionality contains advanced technology and hardware, so it's not designed so much with tinkering in mind. Powered by a set of rechargeable LiPo batteries, we integrate the charging system directly into the box. Right now we can give you 12-15 minutes of hover with a charge time of about 2 hours.
The Whitebox+™ has the ability to move forward and backward, left and right, and rotate around its axis. The app can even pre-program autonomous movements and "hover flight plans". And what's really cool - it can do a static start (that is, pop-up from a standstill when first turned on).
Feel like adding some extra visual oomph to the Whitebox? Check out these three new skins! More to follow, and if you have an idea for one let us know. 
While one day we expect to have hoverboards that can effortlessly float over any medium (even water!), our current technology requires special types of surfaces. 
Therefore, we need a hoverpark to go with our boards, and we have been busy designing a park befitting the awesomeness of our technology.
Hendo is driven by our supporters, backers, and collaborators that want to see a hoverpark come to life. So we're giving everyone an opportunity to be a part of the hoverpark by claiming a brick-sized piece of its surface. 
When you reserve a brick we'll engrave your name on it and place it on our hoverpark surface.  And on opening day you will see your name permanently etched into history, in the world's first hoverpark.
 "What we back today, echoes in eternity" - Maximus Kickstartus Magnetus
Levitation using magnets seems simple - just put one magnet over another, same poles facing, and the top one will float. Voila, right?  Sadly, as we all find out (usually as heartbroken little kids) this never works. Due to something called Earnshaw's Theorem, a stable static equilibrium between two magnets is impossible. There have been a number of ways around this, but none have proven feasible enough for everyday applications. Until now.
Lenz’s law explains how eddy currents are created when magnets are moved relative to a conductive material.  These eddy currents in turn create an opposing magnetic field in the conductor.  Our core technology, which we call Magnetic Field Architecture (MFA™), focuses this field more efficiently. 
You can go ahead and google both of these scientific principles, but to sum it up in regards to levitation: Lenz = Easy, Earnshaw = Hard.
The Hendo Hoverboard is a first-step product, a precursor to the broader implementation of the world-changing technology of MFA.  It enables a new generation of lift and motion technology that will change the way we view transportation. Additional applications for MFA technology are virtually limitless - from business, to industry, to healthcare, and beyond.
Unlike magnetic levitation systems employed today, our hover systems are comparably inexpensive and completely sustainable. Hovering modes of transportation are now possible and practical. Lifting a wide range of loads - whether it's a person riding a hoverboard (what we were all expecting) or a building riding out an earthquake (what we never imagined could be possible) - is all within reach.