To install click the Add extension button. That's it.

The source code for the WIKI 2 extension is being checked by specialists of the Mozilla Foundation, Google, and Apple. You could also do it yourself at any point in time.

How to transfigure the Wikipedia

Would you like Wikipedia to always look as professional and up-to-date? We have created a browser extension. It will enhance any encyclopedic page you visit with the magic of the WIKI 2 technology.

Try it — you can delete it anytime.

Install in 5 seconds
4,5
Kelly Slayton
Congratulations on this excellent venture… what a great idea!
Alexander Grigorievskiy
I use WIKI 2 every day and almost forgot how the original Wikipedia looks like.
Live Statistics
English Articles
Improved in 24 Hours
Added in 24 Hours
What we do. Every page goes through several hundred of perfecting techniques; in live mode. Quite the same Wikipedia. Just better.
Great Wikipedia has got greater.
.
Leo
Newton
Brights
Milds

Minisupercomputer

From Wikipedia, the free encyclopedia

Not to be confused with Superminicomputer.

Minisupercomputers constituted a short-lived class of computers that emerged in the mid-1980s, characterized by the combination of vector processing and small-scale multiprocessing. As scientific computing using vector processors became more popular, the need for lower-cost systems that might be used at the departmental level instead of the corporate level created an opportunity for new computer vendors to enter the market. As a generalization, the price targets for these smaller computers were one-tenth of the larger supercomputers.

Several notable technical, economic, and political attributes characterize minisupercomputers. First, they were architecturally more diverse than prior mainframes and minicomputers in hardware and less diverse in software. Second, advances in VLSI made them less expensive (mini-price). These machines were market targeted to be cost-effective and quickly manufactured. Third, it is notable who did not manufacture minisupercomputers: within the USA, IBM and the traditional mainframe makers, outside the USA: the Japanese supercomputer vendors and Russia (despite attempts to manufacture minicomputers).

The appearance of even lower-priced scientific workstations (e.g., Dana Computer/Ardent Computer/Stellar Computer (the merger of these companies)) based on microprocessors with high performance floating point units (FPUs) during the 1990s (such as the MIPS R8000, IBM POWER2), and Weitek eroded the demand for this class of computer.

The industry magazine Datamation coined the term "Crayette" which in short order meant instruction set compatible to Cray Research, Inc.

YouTube Encyclopedic

  • 1/3
    Views:
    1 937 093
    1 520
    318
  • The RPiCluster
  • Four Computer Science Master's Graduates Receive Awards at MUM
  • Beowulf Boot Camp at CCT

Transcription

This is the Raspberry Pi cluster it's a Beowulf cluster built out of 33 Raspberry Pis I started working on this couple months ago in response to a need I had for my dissertation work in which I was developing a distributed simulator which needed a cluster environment to actually operate so this uh... seemed to be about the cheapest way you could build a 32 node or 33 node cluster so, some of the novel things about this build uh... are i think the the rack design and power management system. So if I take a look at one of these stacks... you can see i have used PCB stand-offs to build each of these columns and allow each of the Pis to stacked one on top of another ..and that allowed plenty of room for component clearance and for air to get through ...to keep things cool... ...and then on the top and bottom... I have drilled plexiglas in that same pattern to hold the whole structure together. Now for the power management... you'll notice...so, normally you would use USB, this little microUSB connector on the corner to power a Raspberry Pi and you can see, I'm not actually using that in the system. Instead, I built a custom PCB that distributes 5 volts to each of these stacks... as well, as you can see, providing a tri-color LED and there are these wire jumpers that go between each PCB to allow power to be moved between the top and bottom of the stack. So i just bring power into one place on the stack and the whole thing is powered and so my regulated 5 volts I'm actually getting from a PC... ...couple PC power supplies because they provide up to 30 amps at 5 volts, so you got plenty of power that's clean... plenty clean for running Raspberry Pis. So one other thing is... with these... ...the PCB port here is that when ever you power through the 5 volt pin on the GPIO header you are actually bypassing a fuse that's normally... uh... that is present between the USB connector and the 5 volt rail so i went ahead and added that back in to this PCB as well so that I still maintain short-circuit protection... on all the boards. So, if I pull around the back here, I've got some air cooling which is certainly overkill for the amount of heat generated by the Raspberry Pis, but I am overclocking all of these to 1 GHz so they do get a little bit warmer than they normally would but even so uh... they don't ever get to temperatures really require the fan cooling... ...but it looks cool... ...and does keep them a lot cooler than they normally would be as well. So, on the corner here, this is one of of them the PC power supplies and I'm just using the 5 volt rail off of that... ...and another one on the other side. Then we've got a 48 port switch that all these Ethernet cables are running into... and then just a couple power cables coming from each of these PC power supplies to power each stack. So, the program that I have running on here is just a parallel program using MPI... and effectively... ...this node on top, the master node is directly controlling the LEDs on all the rest of the nodes in the cluster. This was done to showcase the uh... whole thing operating as a cluster... ...to show how tight the synchronization is actually possible with communications using parallel programming uh... in any case where you have a single LED that's being lit uh... on a node... it is just doing a.regular MPI send/receive operation uh... however on cases where you have multiple LEDs on distinct nodes lighting synchronously... ...those are actually in response to broadcast messages so, you see you can get very tight timing between the different nodes. The are truely operating as one... ...as one program... one machine... So, lets let it sit here and run through some of the different patterns. We have a spiral... ...blink all... ...different colors... So software I have running on here it is Arch Linux... I chose that because it's uh... far more lightweight than Raspbian... ...a lot smaller process list once you have booted the system because you don't have a graphical environment installed by default. Really, aside from that, it's just a matter of having and an NFS share setup so that you have a common filesystem between all the nodes... iand some kind of MPI package installed ...there's the rainbow look... ..with strobing... So at this point, I have been able to move all of my simulation uh... over to this platform and it is performing really well. It has been a really fun project to work on... It is pretty cool to have a 32 node cluster that can sit on your desk...just about.

Notable minisupercomputer companies

This list is sorted alphabetically, and many entries here are to companies that no longer exist.

References

  1. ^ Getting Up to Speed. 2004-02-03. doi:10.17226/11148. ISBN 978-0-309-09502-0.

External links

This page was last edited on 9 May 2024, at 00:39
Basis of this page is in Wikipedia. Text is available under the CC BY-SA 3.0 Unported License. Non-text media are available under their specified licenses. Wikipedia® is a registered trademark of the Wikimedia Foundation, Inc. WIKI 2 is an independent company and has no affiliation with Wikimedia Foundation.
Contact WIKI 2
Introduction
Terms of Service
Privacy Policy