Hard Drive

Hard drives consist of a series of round metal plates called platters, also called cylinders. They are coated with an electromagnetic material which can support magnetic states that are capable of being electrically altered. This means some type of electrical signal can alter the magnetic polarization of various areas of the plates. The state of these polarized areas can also be sensed. Each platter can hold large amounts of data. There are several platters mounted on a hard drive. Between each platter is a head which is used to sense and modify the states of the platter. There are two heads on each platter. Each platter has data stored on it in a specific pattern for read and write access. The data is organized into tracks which are rings around the platter. The distance the head moves into the platter will determine which track is read. A sector is a section of data in the cylinder. Different hard drives have different numbers of sectors, tracks, and platters. The total storage space on the hard drive is traditionally equal to: Sector size times sectors/track times tracks/cylinder times the number of cylinders. With more modern drives, however, to increase storage space, some drives have more sectors on the outer tracks than the inner tracks. This is because there is more physical room for data on the outer tracks. Therefore this method of calculating hard drive capacity may not be effective in the future.

Controller Interface Types

A hard drive is a mass storage device where your operation system is installed along with many data files. There are two types of hard drives with regard to the controller:

1. IDE - Integrated Drive Electronics. A controller based interface. If your primary concern is low price with reasonable performance IDE is a good choice. It is still the most popular controller interface because of price.
2. SCSI - SCSI uses a separate bus hooked to the system bus using a host adapter. It is a more expensive system than IDE, but is better built and has a great deal of flexibility. If you are considering running a server or high performance system, this is the best way to go. There are several types of SCSI interface, the primary characteristic being the width of the data transfer (how many data bits are carried over the cable at a time). The important item is to be sure you get compatible controllers with your SCSI device such as your hard drive or CD-ROM drive.

Most hard drives have three characteristics of main importance for performance.

1. Size - The size of the hard drive is expressed in terms of Gigabytes which is roughly 1000 Megabytes. It is difficult to buy a drive less than 4 Gb today. Typical size are 8 through 20 Gb.
2. Speed - The data output of a hard drive is primary limited by the amount of time it takes for the electromagnetic head to reach the data at specific locations on the drive. The primary factor of limitation is hard drive rotation speed. Common speeds today are 5400 RPM (revolutions per minute), 7200 RPM, and 10000 RPM. Considering price and performance, we currently recommend 7200 RPM hard drives.
3. Reliability - The other performance factor that is worth considering is reliability. This is expressed as mean time between failure (MTBF) The higher the number, the better. Look for this specification on the manufacturer's specification sheets for each product.

Terms

• ATA - AT Attachment. This term refers to the type of IDE drive. Others are Microchannel (MCA IDE) and XT IDE. The ATA interface was used in the early 80386 based computers.
• ATA-2 - Refers to Enhanced IDE or EIDE.

Source:http://www.comptechdoc.org/hardware/pc/begin/hwharddrive.html

Microprocessor

The microprocessor is the center of your computer. It processes instructions and communicates with outside devices, controlling most of the operation of the computer. The microprocessor usually has a large heat sink attached to it. Some microprocessors come in a package with a heat sink and a fan included as a part of the package. Other microprocessors require you to install the heat sink and fan separately. This is not a difficult problem, but can be a bit daunting when the buyer wants to make sure they get the correct parts to fit their microprocessor. Also the buyer needs to make sure they will get the motherboard that their microprocessor will work with. This section will explain some of the differences in microprocessors and ways to be sure your parts match.

Microprocessors and Mounting

The mounting method refers to the type of connection the microprocessor makes with the motherboard. The following table lists the various mounting packages and some of the well known microprocessors that are mounted for that package.

• Socket 7 - AMD K5, K6, Intel Pentium 75-200Mhz, IBM
• Socket 370 - Some Intel Celerons
• Slot 1 - Intel Pentium II, Pentium III, Some Celeron 266-533
• Slot II - Intel Xeon
• Slot A - AMD Athlon

The Socket 7 processors are becoming less popular. We recommend socket 370, through slot A microprocessors at the current time. The prices on Socket 370 microprocessors are currently very low considering the performance of the systems. I recently bought a Celeron 500Mhz microprocessor with 66Mhz sidebus for under $120 with a motherboard for $84. When buying a microprocessor, make sure you get the type of socket you think since some processors are made for different sockets such as the Celeron. Be sure of one of the following.

1. The socket type is stated at the vendors website.
2. There is a microprocessor part number stated at the vendors website that can be traced to the manufacturers website which specifies the mounting package you want.

It would be no fun to get a Slot 1 motherboard and a socket 370 Microprocessor.

Microprocessor heat sinks and fans.

Being sure you get the correct heat sink and fan for your microprocessor can be a bit daunting. Who wants to get a $300 microprocessor, and risk it with an incorrect mounting of a heatsink or fan? Who wants to find out that they have purchased the wrong heatsink for their processor and spend days or weeks trying to sort it out? My solution is to purchase the microprocessor with the heatsink in the same package. Usually you get a better warranty and return policy this way and you don't need to worry about whether the two are compatible. I do not believe you can save enough money buying the heatsink and fan from anyone other then the vendor selling the microprocessor because of the time it takes for the additional research required and the potential trouble. The best solution to this problem is simply to buy a slot1, slot II or slot A microprocessor with the package that includes the fan and heatsink. These would be one of the Pentium II, Pentium III, Athlon, or Xeon packages. All that is required in this case is to slide the microprocessor carefully into its slot. With the exception of processors such as the Athlon which have a larger heat sink, requiring an extra plastic clip mechanism to help stabilize the heatsink, it is easier to install one of these processors than it is to install the computer's RAM memory or a hard drive.

Source:http://www.comptechdoc.org/hardware/pc/begin/hwmicroprocessor.html

Motherboard The Memory Slots Yes, the motherboard is the mother of all boards on your computer. The motherboard may have a form factor of AT or ATX. We recommend you use ATX motherboards with ATX cases since this is the newer alternative and most modern microprocessors run on ATX motherboards. The motherboard holds the microprocessor, the memory, and several card slots. The memory may be SIMM sockets or DIMM sockets. The current standard is DIMM socketed memory. This is usually 168 pin 3 volt unbuffered synchronous DRAM memory. PC100 or PC133 memory is the current memory of choice. Most boards have 3 or 4 memory slots, which may, depending on the size of DIMM used, allow up or beyond 1 Gb total system memory. Most boards commonly allow 384 to 512 Mb of system RAM. The expansion bus The card slots are used to put additional cards such as video cards, sound cards, internal modems, or network cards into. Some motherboards today include video and sound without the addition of a extra card. These cards slots today are mostly PCI type card slots. When talking about cards that are plugged into a PC you are talking about the expansion bus. The expansion bus is a means of a microprocessor extending its communication ability further into the outside world. It is a data exchange means between add on cards and the microprocessor and the motherboard. There have been several types of expansion buses. ISA - Industry Standard Architecture. Used when the original 8088 8bit microprocessor based personal computers were produced. EISA - Extended ISA used when the 80286 through 80486 series microprocessors were being produced. This bus is still used but is being phased out and is almost gone today. MCI - Microchannel architecture by IBM and used mainly on IBM brand computers. PCI - Peripheral Component Interconnect. The popular expansion bus of choice. It is significantly faster than EISA. AGP - Accelerated Graphics Port. This bus is developed for fast video cards. It is currently up to 4X mode speed. The current popular expansion bus is the PCI (Peripheral Component Interconnect) bus for all cards except the graphics cards. For graphics cards, the bus of choice is AGP. Most motherboards today have one AGP slot and several PCI slots. Your expansion cards will plug into these card slots. Be sure you get cards that match the available type of slots on your motherboard. My microprocessor runs at 500Mhz and my memory runs at 100Mhz. Why? As PC technology grew, eventually the access speed of the memory could no longer keep pace with the increased speed of the microprocessors. At this point, an I/O cache was placed on the microprocessor to be a buffer between the external memory on the motherboard and the internal processor registers. The memory was set to run at a different "side bus" speed which is some fraction of the microprocessor speed. Therefore when the speed of the microprocessor is set, it is set to some multiple of the side bus speed. In the case of a 500Mhz processor and 100Mhz PC100 capable memory, that multiple is 5. Sometimes this multiple and the sidebus frequency is set using jumpers on the motherboard, or it may be set with auto detection and the BIOS. You will need to consult your motherboard manual to determine how to set these parameters. Other motherboard items Other items on your motherboard that you should be aware of are the small pin connectors that are used to connect the following controls and indicatory to your motherboard. Power supply switch. Reset switch The power on indicator. Hard drive activity indicator. In the case speaker connector. You will need to consult your motherboard manual to see which connectors are used for which item and how to hook them up. There should be a bundle of cables near the front of the case (inside) which have labels on the connectors for these items. Chipset and BIOS One issue that will affect the operation of the motherboard is the chipset it uses and its BIOS it uses. The chipset is used to control the interface between the microprocessor and most of the devices and memory on the computer. The chipset used can have a significant affect on the performance of your system as can the overall design of the motherboard. The way to determine the best chipsets and motherboards is to read reviews and articles at various technical websites. Your system's BIOS is a computer program that allows your system to begin running and provides a small library fo function that your system will use to interface to various devices such as your hard drive. Some BIOS programs can limit the location on your hard drive where you can install bootable operating systems. The BIOS resides in a chip on the motherboard called a ROM chip. Usually part of this ROM can be reset or re programmed with updates. ROM that can be electrically re-written this way is called "flash" ROM. Source:http://www.comptechdoc.org/hardware/pc/begin/hwmotherboard.html

Computer Components

Computers come in all types and sizes. There are primarily two main sizes of computers. They are:

• Portable
• Desktop

The portable computer comes in various sizes and are referred to as laptops, notebooks, and hand-held computers. These generally denote different sizes, the laptop being the largest, and the hand-held is the smallest size. This document will mainly talk about the desktop computer although portable computer issues are also discussed in various areas.

Computer Components:

Computers are made of the following basic components:

1. Case with hardware inside:
   1. Power Supply - The power supply comes with the case, but this component is mentioned separately since there are various types of power supplies. The one you should get depends on the requirements of your system. This will be discussed in more detail later
      
   2. Motherboard - This is where the core components of your computer reside which are listed below. Also the support cards for video, sound, networking and more are mounted into this board.
      1. Microprocessor - This is the brain of your computer. It performs commands and instructions and controls the operation of the computer.
      2. Memory - The RAM in your system is mounted on the motherboard. This is memory that must be powered on to retain its contents.
      3. Drive controllers - The drive controllers control the interface of your system to your hard drives. The controllers let your hard drives work by controlling their operation. On most systems, they are included on the motherboard, however you may add additional controllers for faster or other types of drives.
   3. Hard disk drive(s) - This is where your files are permanently stored on your computer. Also, normally, your operating system is installed here.
      
   4. CD-ROM drive(s) - This is normally a read only drive where files are permanently stored. There are now read/write CD-ROM drives that use special software to allow users to read from and write to these drives.
      
   5. Floppy drive(s) - A floppy is a small disk storage device that today typically has about 1.4 Megabytes of memory capacity.
      
   6. Other possible file storage devices include DVD devices, Tape backup devices, and some others.
2. Monitor - This device which operates like a TV set lets the user see how the computer is responding to their commands.
   
3. Keyboard - This is where the user enters text commands into the computer.
   
4. Mouse - A point and click interface for entering commands which works well in graphical environments.

 Source:  http://www.comptechdoc.org/hardware/pc/begin/hwcomputer.html

So You Think You Can Solve a Cosmology Puzzle? Scientists Challenge Other Scientists With a Series of Galaxy Puzzles

So You Think You Can Solve a Cosmology Puzzle? Scientists Challenge Other Scientists With a Series of Galaxy Puzzles

ScienceDaily (Dec. 8, 2010) — Cosmologists have come up with a new way to solve their problems. They are inviting scientists, including those from totally unrelated fields,
to participate in a grand competition. The idea is to spur outside interest in one of cosmology's trickiest problems -- measuring the invisible dark matter and dark energy that permeate our universe.

The results will help in the development of new space missions, designed to answer fundamental questions about the history and fate of our universe.

"We're hoping to get more computer scientists interested in our work," said cosmologist Jason Rhodes of NASA's Jet Propulsion Laboratory in Pasadena, Calif., who is helping to organize the challenge, which begins on Dec. 3, 2010. "Some of the mathematical problems in our field are the same as those in machine-learning applications -- for example facial-recognition software."

JPL and several European Universities, including The University of Edinburgh and University College London in the United Kingdom, are helping to support the event, which is funded by a European Union group called Pattern Analysis, Statistical Modelling and Computation Learning. The principal investigator is Thomas Kitching of the University of Edinburgh.

This year, the competition, which has operated since 2008, is called GREAT 2010, after GRavitational lEnsing Accuracy Testing. The challenge is to solve a series of puzzles involving distorted images of galaxies. Occasionally in nature, a galaxy is situated behind a clump of matter that is causing the light from the galaxy to bend. The result is a magnified and skewed image of the galaxy. In the most extreme cases, the warping results in multiple images and even a perfect ring, called an Einstein Ring after Albert Einstein, who predicted the effect. But most of the time, the results are more subtle and a galaxy image is distorted just a tiny bit -- not even enough to be perceived by eye. This is called weak gravitational lensing, or just weak lensing for short.

Weak lensing is a powerful tool for unlocking the fabric of our universe. Only four percent of our universe consists of the stuff that makes up people, stars and anything with atoms. Twenty-four percent is dark matter -- a mysterious substance that we can't see but which tugs on the regular matter we can see. Most of our universe, 72 percent, consists of dark energy, which is even more baffling than dark matter. Dark energy is gravity's nemesis -- where gravity pulls, dark energy pushes. By studying lensed, or distorted, galaxies, scientists can create better maps of dark matter -- and by studying how dark matter changes over time, they can better understand dark energy.

Weak lensing is a promising method for tackling these questions. The 2010 U.S. National Research Council Decadal Survey on astronomy and astrophysics has ranked mission proposals using this method as high priorities.

The GREAT 2010 challenge is designed to improve weak-lensing know-how. Participants will start with fuzzy pictures of galaxies that have been distorted ever so slightly by invisible dark matter parked in front of them. The effect is so small that you can't see it with your eyes. The problem is even trickier because the telescopes are also distorting the galaxy images to an even greater degree than the dark matter. It takes complex techniques -- mathematical models and image-analysis algorithms -- to tease apart these various influences and ultimately discover how dark matter is warping a galaxy's shape.

"This is an image-analysis challenge. You don't need to be an astronomer or cosmologist to help measure the weak-lensing effect," said Kitching. "This challenge is meant to encourage a multidisciplinary approach to the problem."

Participants will have nine months to solve a series of thousands of puzzles. The winners will be announced at a closing ceremony and workshop held at JPL. Prize-winners can expect some kind of cool gadget -- as well as the satisfaction of having brought the world one step closer to understanding what makes our universe tick.

To participate in the venture, in-depth technical information is available online at: http://www.greatchallenges.info .

Tennessee's Kraken Named World's Third Fastest Computer, ORNL's Jaguar Is No. 1

East Tennessee is now home to two of the world's three fastest computers, according to new rankings released recently.


The Top500 list of the world's fastest supercomputers places University of Tennessee supercomputer Kraken in third place, where it also holds the title of world's fastest academic supercomputer, while Oak Ridge National Laboratory's Jaguar computer took first place overall.

Kraken, the result of a $65 million grant to UT from the National Science Foundation, recently became only the fourth computer in history to perform more than 1,000 trillion calculations per second, known as a petaflop.

"Winning a $65 million NSF award put the UT among the supercomputing elite, and now we have reached the pinnacle in having the world's fastest academic supercomputer," said Jan Simek, interim UT president. "This is a phenomenal achievement and is among growing distinctions that enable us to continue attracting the best faculty and the best students we have ever had, and to make our university the best it has ever been."

The twice-yearly Top500 is published by Jack Dongarra, a UT Knoxville distinguished professor of computer science and the director of the Innovative Computing Laboratory along with colleagues at Lawrence Berkeley National Laboratory and the University of Mannheim.

Since its creation in 2007, Kraken has been used for nearly 300 scientific projects addressing vital questions in areas from climate and weather modeling to applications in genetics and medicine.

"The beauty of Kraken is not just its computing power, but its problem-solving power," said UT Knoxville Chancellor Jimmy G. Cheek. "Scientists from universities around the country, including many here at UT Knoxville, have put Kraken to use to attack humanity's most pressing problems. It is an invaluable resource to this university to be home to such a powerful asset."

With the combined computing power of UT and ORNL, East Tennessee is now firmly ensconced as a center for supercomputing activities, a fact which is continuing to draw even more scientific resources to the area, leading not only to technological

"Kraken is the result of a powerful and expanding partnership between the University of Tennessee and the Oak Ridge National Laboratory to advance the frontier of scientific discovery and innovation from climate change to energy technologies and from basic to applied sciences," said Thomas Zacharia, deputy director for science and technology at ORNL and a UT Knoxville professor.

"When NSF made the award to UT Knoxville to develop Kraken just over a year ago I said that 'like the gargantuan sea monster Kraken, which inspired the naming of this supercomputer, the possibilities in scientific and engineering advances it enables are enormous, limited only by the confines of human imagination and vision beyond the frontiers of science,'" said NSF Director Arden L. Bement. "Today, Kraken is working to realize that vision. Consistent with ORNL's leadership in building what many would consider to be one of the most diverse and valuable computation centers in the world, Kraken will address some of the most complex problems of our era."

In October, the NSF awarded UT Knoxville an additional $10 million grant to create a new computer called Nautilus designed to help analyze and process the complex data created by massive computers like Kraken and Jaguar. Nautilus and its accompying research center will be part of UT's National Institute for Computational Sciences (NICS), which also manages Kraken.

Both Kraken and Jaguar are Cray XT5 supercomputers. Kraken alone has 100,000 processors that work simultaneously to produce the high speeds at which the computer is capable to address major scientific questions.

Jaguar clocked in at a sustained speed of 1.759 petaflops, while Kraken registered 831 teraflops.

In the most recent version of the list, Kraken ranked sixth while Jaguar was in second place. In addition to Kraken and Jaguar, East Tennessee also is home to two more machines ranked in the world's top 30, with ORNL's original Jaguar Cray XT4 system at 16th and UT's Athena in 30th.


Source:   Click

Sleep Talking' PCs Save Energy And Money

ScienceDaily (May 1, 2009) — Personal computers may soon save large amounts of energy by "sleep talking." Computer scientists at UC San Diego and Microsoft Research have created a plug-and-play hardware prototype for personal computers that induces a new energy saving state known as "sleep talking."

---

Normally PCs can be in either awake mode—where they consume power even if they are not being used, or in a low power sleep mode—where they save substantial power but are essentially inactive and unresponsive to network traffic. The new sleep talking state provides much of the energy savings of sleep mode and some of the network-and-Internet-connected convenience of awake mode.
UC San Diego computer science Ph.D. student Yuvraj Agarwal presented this work on April 23, 2009 at the USENIX Symposium on Networked Systems Design and Implementation (NSDI 2009). Computer scientists at UC San Diego and Microsoft Research in Redmond, Washington and Cambridge, UK collaborated on this project and the NSDI 2009 paper, "Somniloquy: Augmenting Network Interfaces to Reduce PC Energy Usage."
"Large numbers of people keep their PCs in awake mode even though the PCs are relatively idle for long blocks of time because they want to stay connected to an internal network or the Internet or both," said Agarwal. "I realized that most of the tasks that people keep their computers on for—like ensuring remote access and availability for virus scans and backup, maintaining presence on instant messaging (IM) networks, being available for incoming voice-over-IP (VoIP) calls, and file sharing and downloading—can be achieved at much lower power-use levels than regular awake mode," said Agarwal.
Following this realization, the team built a small USB-connected hardware and software plug-in system that allows a PC to remain in sleep mode while continuing to maintain network presence and run well-defined application functions. It supports instant messaging applications, VoIP, large background web downloads, peer-to-peer file sharing networks such as BitTorrent, and remote access. The computer scientists say their system is easily extensible to support other applications.
The computer scientists named their system Somniloquy, which means "the act or habit of talking in one's sleep." In fact, the system allows a PC to appear to "say" to other hosts on the network, "I'm awake and I can perform non-power-intensive tasks"—even though the PC is in sleep mode. If more computational muscle or resources present on the PC such as stored files are required, Somniloquy wakes up the PC.
The goal of Somniloquy is to encourage people to put their PCs in sleep mode more often, for example when they are not being used for computationally demanding tasks. "Reducing energy consumed by wall-powered devices, especially computing equipment, offers a huge opportunity to save money and reduce greenhouse gasses," said Agarwal.
"Somniloquy uses a very small low-power computer. It has a low-power processor, some memory, a lightweight operating system, and a small amount of flash to store data. Everything is scaled down and extremely energy efficient," said Agarwal, a self described "computer systems" researcher who uses hardware insights to build better energy-efficient computer systems.
Somniloquy's low-power secondary processor functions at the PC's network interface. It runs an embedded operating system and impersonates the sleeping PC to other hosts on the network. Somniloquy will wake up the PC over the USB bus if necessary. For example, during a movie download, when the flash memory fills up, Somniloquy will wake up the PC and transfer the data. When the transfer is complete, it will go back to sleep mode and Somniloquy will again impersonate the computer on the network.
The current prototypes work for desktops and laptops, over wired and wireless networks, and are incrementally deployable on systems with an existing network interface. It does not require any changes to the operating system on the PC, to routers or other network infrastructure, or to remote application servers.
The researchers evaluated Somniloquy in various settings and say that it consumes 11 to 24 times less power than a PC in idle state, which could translate to energy savings of 60 to 80 percent depending on their use model.
In the future, Somniloquy could be incorporated into the network interface card of new PCs, which would eliminate the need for the prototype's external USB plug-in hardware.
Yuvraj Agarwal worked on this project when he was an intern at Microsoft Research, Cambridge in the summer of 2007, and continued afterwards at UC San Diego. Steve Hodges, Ranveer Chandra, James Scott, and Victor Bahl from Microsoft Research are Somniloquy co-inventors and authors on the NSDI 2009 paper. Agarwal's Ph.D. advisor and paper-coauthor is computer science professor Rajesh Gupta from UC San Diego's Jacobs School of Engineering.
"Somniloquy: Augmenting Network Interfaces to Reduce PC Energy Usage," by Yuvraj Agarwal from UC San Diego, Steve Hodges, Ranveer Chandra, James Scott, and Paramvir (Victor) Bahl from Microsoft Research, and Rajesh Gupta from UC San Diego, was presented on April 23, 2009 at USENIX Symposium on Networked Systems Design and Implementation.

Computing: Heat Helps in Low Power Data Storage Scheme

Computing: Heat Helps in Low Power Data Storage Scheme
Random thermal fluctuations in magnetic memory can be harnessed to reduce the energy required to store information, according to an experiment reported in the current issue of Physical Review Letters. The development could lead to computer memory that operates at significantly lower power than conventional devices. Markus Münzenberg of Universität Göttingen and Jagadeesh Moodera of MIT describe the potential route to greener magnetic memory in a Viewpoint in the latest issue of APS Physics.

Heat is usually a problem when it comes to storing digital data. At the microscopic level, the molecules and atoms of anything at a temperature above absolute zero are in constant motion. Because magnetic memory relies on controlling and measuring the orientation of tiny magnetic particles, the jostling that comes about as components warm up can potentially scramble data. Thermal issues are a major concern as researchers build increasingly dense and fast magnetic memory. But heat isn't entirely bad, according to a collaboration of Italian and American physicists that has shown that random thermal motions can be helpful for writing magnetic data. Essentially, they found that applying an electrical current that should be too modest to record data can still be effective for writing information because thermal motion gives an added boost to help orient magnetic particles.
The researchers confirmed the effect by measuring magnetic fluctuations as the particles that make up memory were being aligned. Thermal motions are random, which in turn causes random variations in the amount of time it takes for magnetic particles to line up. The fact that alignment times ranged from one to a hundred billionths of a second made it clear that random, temperature-dependent motion must be at work in helping to flip the particles.
The experimental confirmation of the thermal effects on magnetic memory points the way to new, thermally-assisted data writing schemes. The advances could reduce the power required to store information, potentially helping to ensure that future PCs are increasingly green machines.

Story Source:

The above story is reprinted (with editorial adaptations by ScienceDaily staff) from materials provided by American Physical Society, via EurekAlert!, a service of AAAS.