This is a split board - You can return to the Split List for other boards.

Eaglerulez's guides to building a computer Version 2.

#1EaglerulezPosted 8/20/2007 12:01:36 PMmessage detail
Table of Contents.

1. Introduction
2. Why Build a computer?
3. How Computers Work
- The Processor
-The Video Card
-RAM
-Motherboard
- The 4 Big companies
- Power Supplies
- Hard drives
- Optical Drives
- Cases
4. Buying Advice.
5. List of Recommended Manufacturers
6. Assembling your computer
7. Configuring your computer
8. Overclocking
9. Cooling




1. Introduction This is Version 2 of my building a computer guide. Not much has changed from version 1, though I have spruced it up a bit, and have corrected some minor flaws. This guide was designed towards first time builders and thus most of this content is directed solely towards first time builders. If you are not building a computer though, you can still learn a lot from this guide.

Furthermore feel free to offer any comments, critiques, insights, or questions, as I feel it will help me improve this guide. The only thing I ask is that people not ask me questions pertaining directly towards their individual builds in this topic, as I will have another help topic to answer specific questions.

Finally, I cannot stress enough, that this is a CRASH COURSE to building computers, I will tell you right now, you must do loads of your own research to truly get an understanding of how computers work and what you need to know when building a computer. Heck anyone with technical knowledge will laugh at this guide, because my explanations on the components of the computer are very primitive, and simple. This guide is merely the starting step when it comes to understanding computers.


THIS IS A MASSIVE GUIDE OVER 27 PAGES LONG, DO NOT POST UNTIL I SAY SO
---
www.actsofgord.com . Long live Green Earth. Escalators can't break, they can only temporarily become stairs. I Love GREEN TEA
#2Eaglerulez(Topic Creator)Posted 8/20/2007 12:02:32 PMmessage detail
2. Why build a computer?

Building a computer is so simple that I can't even find an adequate analogy. Find yourself a decent building guide (some on the 'net are dated, but that is no matter...you just want the general idea). Once you understand what parts you need and order, the particular components will have their own specific instructions, which may or may not vary from the guide you used.

The beauties of building your own computer are numerous.

- Yes, you do save money, no matter what other people try to say. Every commercial computer manufacturer cuts corners. They use cheaper versions of well-known parts, like video and sound cards. They are almost like their retail brothers, but something might be clocked differently or they might use RAM that failed certain tests. And when it comes to invisible parts, don't get me started. Even if you buy the highest quality stuff, you will spend about the same amount and know that you don't have junk inside your computer.

- Every part that goes into the computer is something you actually wanted. You aren't limited to the cookie-cutter options that companies give you.

- You get warranties on all your parts, so you can simply return them for a replacement, rather than screwing around with a company's customer service, which often make you jump through hoops.

- You will learn so much about computers by reading up and then putting the computer together than you will never, ever again be at the mercy of tech support staffs or friends/family members who know more about computers. You will be the guy who people are always begging for help.

If someone is just looking for an entry-level or mainstream machine, then they should by all means go the commercial route. But if you want a gaming machine or a computer at the center of your entertainment center, building is by far the superior route.

(By Codenameplasmasnake)

---
www.actsofgord.com . Long live Green Earth. Escalators can't break, they can only temporarily become stairs. I Love GREEN TEA
#3Eaglerulez(Topic Creator)Posted 8/20/2007 12:04:11 PMmessage detail
3. How Computerís work

There are 4 main components that will determine the overall performance of your computer. These components are the CPU (also known as the processor), RAM, GPU (also known as video card), and the motherboard. Oddly enough there are 4 main companies in computing that have essentially developed the way computers work. So now letís get down to business.

The CPU

First we will talk about the CPU. The CPU is the brain of the computer. Just like your brain subconsciously keeps your heart pumping, organs working, etc, the CPU figures out a way to accomplish every task it is asked to do. Simply put everything that you do on a computer is accomplished with the help of the CPU. The CPU is essentially a processor that can do math really, really quickly. What happens is the program you are using sends the CPU instructions in the form of mathematical equations. The CPU solves the equations, allowing the program to do what it has to. There are 4 main things that determine a CPUís speed.

Frequency

The simplest to understand is the frequency of the processor. As youíve most likely learned in science class the frequency of something is measured in hertz. At first computers ran at only a few megahertz, then a couple hundred megahertz, but now, most are running at atleast 1.8+Gigahertz. This frequency determines how fast the CPU can process information, do mathematical problems, ect. So obviously the more Ghz, means the faster the CPU. But this isn't always the case

Caches

The second way to determine a CPUís power is looking at the Caches. The caches are essentially small caches of memory built inside the CPU. When the CPU gets a particularly complicated equation it can use these memory caches to help solve the equation. There are three levels of caches (although the first two are the ones that are usually used in processors) The first and fastest cache is the Level 1 Cache. The L1 cache usually has a smaller amount of memory then the later two levels, but this doesnít mater because it is the fastest cache . Some modern day processors have as low as 64KB of L1 cache, while others can have 128Kb of L1 cache, the number varies among processors. The second level of cache known as the L2 cache this cache has a much larger amount of memory but is slower. Most L2 caches can range from 512KB to 4MB of cache. The final level is the L3 cache, the L3 cache is by far the slowest of the caches, but has the most memory. In most cases you will never find an L3 cache listed in a processorís tech specs because it is not found in most processors, but it does exist.

The levels of caches are usually larger then the level before it. For instance if the equation is too large for the L1 cache to handle, the processor will send it down to the larger L2 cache which may send it to the L3 cache if your processor has one. All in all a processor with larger cacheís but slower frequency can sometimes perform toe to toe with a processor of higher frequencies and lower caches. However keep in mind cache does not make too much of a difference in performance. When comparing two processors with the same architecture and clockspeed, the one with more cache will perform better, however the difference in performance is nothing major.

For instance an Athlon 64 X2 4200 with 128MB L1 Cache, 512MB L2 cache, at 2.2Ghz performs equally to an Athlon 64 X2 4000 with 128MB L1 cache, 1MB L2 Cache. And 2 Ghz.



---
www.actsofgord.com . Long live Green Earth. Escalators can't break, they can only temporarily become stairs. I Love GREEN TEA
#4Eaglerulez(Topic Creator)Posted 8/20/2007 12:09:03 PMmessage detail
Architectures

The most important factor in determining performance is the processor's Architecture.

You see each processor on the market is named, for instance the name Pentium 4. What most people donít know is that Pentium 4 is actually the branding for the type of architecture applied to the processor.

The architecture affects the way the CPU handles information, and will affect the frequencies, heat dissipation, power consumption, and overall everything that has to do with the processor. Now hereís the part where most people get confused. Architecture is actually the best way to tell a CPUís performance. Let me give you an example.

An AMD Athlon 64 3200 that clocks in at only 2.0Ghz, 128Kb L1 cache, and 512KB L2 cache, can out perform a Pentium 4 630 that has a whopping 3.0Ghz, 64+52Kb L1 Cache, and a whopping 2MB L2 cache. From a purely technical standpoint the P4 significantly spanks the Athlon. Even though the Athlon has more L1 Cache, the P4 has 4 times the L2 cache, and a whole 1000Mhz faster frequency, so why does the Athlon out perform it? Because the architecture is simply more efficient in the way it handles information.

To give another example say we have an Athlon X2 at 2.0Ghz that can process 2 pieces of information per Mhz. 2X 2000Mhz= 4000 pieces of information per second

Now say we have a Core 2 Duo running at 1.86 Ghz, and can process 4 items of information per Mhz, since it has a more efficient architecture.4X 1860 Mhz= 7440 pieces of information per second. Almost double that of the X2.

While the Core 2 has a slower clock speed it can process more items of information per Mhz then the Athlon and thus is a better performer.


The only problem with using the architecture to tell how fast a processor is, is the fact that there are so many names, and architectures out there that itís easy to get confused, plus what some people think as new and cutting edge, and thus fast, becomes obsolete 6 months later. So Iíve prepared this little factoid on the processor architectures that you should be looking at, keep in mind there are tons more on the market.

Guide to Architectures

To protect myself from elitists I would like to point out some things when it comes to architectures. Each processor is comprised of two things, a microarchiecture, and a core.

The Microarchtiecture, or as we call it in this guide, the architecture is the way the processor handles information, and executes instructions.

The Core is the physical makeup of the processor, which includes all the transistors, the physical memory of the caches, etc. To explain cores, say we are looking at a Pentium 4. The Pentium 4 will say something like Pentium 4 Prescott. Prescott would be the Core in this case. What this means, is that, that particular Pentium is using the Presscott core. A Pentium 4 using the Prescott core would be slightly different from other Pentium 4's using other cores (say a Pentium 4, that reads Pentium4 Norhwood) The differences between cores could be slight, for instance, one core may be designed with extra cache, or another would be designed for lower heat output and power consumption, however the differences between some cores could also be dramatic, with new cores introducing faster front side busses, more processing cores, etc even though they still use the same micro architecture.


---
www.actsofgord.com . Long live Green Earth. Escalators can't break, they can only temporarily become stairs. I Love GREEN TEA
#5Eaglerulez(Topic Creator)Posted 8/20/2007 12:11:51 PMmessage detail
It is not really important to know the differences between cores, as they are minor and will usually make a minor impact on performance assuming, they all have the same architecture. Also any major differences between the cores will usually have a different branding to distinguish the differences. For instance some Core 2's have a 1066FSB, and they are branded as E6XX processors. While some core 2's have a 1333FSB and they are branded as E6X50 processors. However it is important to know that there is a difference between architecture and core.

AMD

The Athlon 64 The Athlon 64 uses the AMD64 architecture and is the single core equivalent to intelís Pentium 4. The Athlonís tend to have slower clock speeds than the Pentium 4's, as well as smaller caches, but their architecture makes them cooler, more power efficient, and equally, if not stronger performers then the Pentium 4.

The Athlon 64 X2.: The Athlon 64 X2, is the dual core counter part to the Athlon 64, and the equivalent to an Intel Pentium D. This baby has all the features of a standard Athlon, however instead of one processing core, it has two, making it dual core. What this means is that within the die there are two L1 Caches one for each processing core, two L2 Caches, and 2 cores both of which run at the same speed. These babies outperform their Pentium D equivalents, like their single core brethren. Dual core processors are essentially two processors doing the work instead of one.

Intel

The Pentium 4. The Pentium 4 uses the Netburst architecture which was made to achieve high clock speeds (frequencies) at all cost, this meant sacrificing, efficiency, which made them draw more power, and give off more heat. The reason for this was because the general consumer assumed that higher Ghz meant the faster processor so when looking at an Athlon at 2Ghz, or a P4 at 3.0Ghz they would obviously chose the P4. Intel's general goal for the Pentium 4 was to get clock speeds of about 5-7Ghz. However the P4ís hit a road block since in order to achieve speeds faster then 4.0Ghz too much power would be drawn, thus a ton of heat would be produced, and thus anything faster then 4Ghz would be unsuitable for a standard home PC. So for a good while the P4 lost its influence to the cheaper, faster, and generally better Athlon 64.



The Pentium D. The dual core equivalent of the Pentium 4, can be spanked by the Athlon X2.


---
www.actsofgord.com . Long live Green Earth. Escalators can't break, they can only temporarily become stairs. I Love GREEN TEA
#6Eaglerulez(Topic Creator)Posted 8/20/2007 12:12:08 PMmessage detail
The Core 2 Duo: The Core 2 Duo utilizes the Conroe architecture which is intelís newest architecture that was released very recently. Instead of sacrificing everything for speed the Core 2ís are extremely efficient. In fact one of the cheapest Core 2's the E4300 which only clocks in at 1.80 Ghz can outperform an Athlon 64 X2 4200 which is clocked at 2.2 Ghz, a way faster frequency. So why can this baby out perform an Athlon? Well for one it has a shared L2 cache. In most other dual core processors, each core was given only a fixed amount of cache. This meant if a core needed more memory it would have to use slower alternatives by borrowing the memory from other sources. If an equation didnít require all of the L2 cache memory, the whole L2 cache memory would still be used, although it had no effect on the performance, meaning more heat would be produced, and more energy would be drawn. However with a shared L2 cache, each core can take however much it needs. So if it needs more for a certain equation, it is readily available. While if it doesnít need a lot of memory, power can still be saved by not dedicating the whole cache. The two cheapest Core 2ís have a shared 2MB L2 cache, while the later ones have a shared 4MB cache. In older processors this would mean each core would get 1 or 2MB, but now each core can use however much it needs. The Core 2ís are the best buy you can get today. The cheapest one is only about 100 dollars; it is a fast dual core processor, and is extremely energy, and information efficient. It is said to have roughly a 30% performance increase against itsí priced equivalents. The other value of the Core 2 Duo's is that they can overclock very well, allowing a budget 1.8Ghz Core 2 Duo to reach speeds of 3.3Ghz.

See Overclocking



Process types

Finally, the last and most minor form of determining a CPUís speed is its process type. What this means is how close the transistors in the CPU are together. Generally the closer together the transistors the less energy they have to use to communicate, meaning less heat, and overall better efficiency. The way this is expressed is in nanometers or NM. Meaning the transistors are only a certain number of nanometers apart. Most Athlon processors run on a 90NM process, while the new core 2ís run on a 65NM process (also contributing to why they are so efficient). As time goes on the processes will get smaller and smaller, meaning more cores can be fit into a single CPU wafer. For instance the new quad core processors that are out run on a mere 45Nm process which is why they can fit 4 processing cores on a single waffer, since everything is so compact together. The advantage of smaller process types is usually found when overclocking, since cores with smaller process types draw less energy and produce less heat, and are therefore ideal for reaching higher clock speeds.
---
www.actsofgord.com . Long live Green Earth. Escalators can't break, they can only temporarily become stairs. I Love GREEN TEA
#7Eaglerulez(Topic Creator)Posted 8/20/2007 12:13:03 PMmessage detail
Video cards

Video cards are what allow you to display images on your screen. What happens is a graphically intensive program, say a video game will send the graphical equations to the CPU, these equations are very complex, and with the CPU running a hundred other different things the computer/ program will slow down since the CPU is trying to deal with these graphical equations. So now with all of our extremely pretty video games, we must have a graphics card to solve all the complex graphic equations, taking a heavy load off of the CPU. With a graphics card, the CPU simply sends the equations to the graphics card for it to handle. Like the CPU there are 4 main things in determining a Graphic cardís speed.

The core: The core of a video card is very similar to that of a regular CPU. It takes equations, and solves them. However the answers to these equations are pixels, which are displayed on the screen. The cores of most modern video cards run around 400-600+ Mhz: This time though the faster the core, the generally faster the card, you donít have to worry about the architecture or any of that, just simply the frequency.

The pixel pipeline: After the core is finished with an equation, it sends the result of the equation, a pixel, down the pixel pipeline. Here textures, lighting effects, shadows, colors, ect. Are added to the pixel, and then displayed on screen. In most cases the more pixel pipelines, the better since the pixel pipelines usually determine the graphical power of a card. For instance a card with a 400Mhz core and 12 pixel pipelines is better then a card with a 500mhz core and only 8 pixel pipelines. Most graphics cards that are decent have around 12-24 pixel pipelines. One thing that you should keep in mind is that graphic card creator ATI uses less pixel pipelines, but more shaders per pipeline. If you where to look at an ATI cardís spec sheet it may say 16 pixel pipelines (48 shaders). What this means is that while the card has less pixel pipelines, it can pretty much do more with the pixels that are coming down the pipelines. However a card with 24 pipelines can perform closely to that of an ATI with 16 pipelines and 48 shaders. In general the more pixel pipelines the better image quality you will receive. Newer cards use Stream Processors, which are very similar to pixel pipelines. The same rule follows for these cards as well, as the more Stream Processors the better.

---
www.actsofgord.com . Long live Green Earth. Escalators can't break, they can only temporarily become stairs. I Love GREEN TEA
#8jamil2nicePosted 8/20/2007 12:13:42 PMmessage detail
[This message was deleted at the request of a moderator or administrator]
#9Eaglerulez(Topic Creator)Posted 8/20/2007 12:16:23 PMmessage detail
Memory Each independent video card has its own memory to hold the images that are displayed on the screen. These are the images that have been created by the pixel pipelines. Most cards can suffice with 256MB, but most higher end cards have so much data in the form of pixels (since higher end cards can create much more advanced effects) that higher end cards tend to have 512MB of memory or more. In all honesty a video cardís amount memory means very little as far as performance. You should only get a 512MB card if thatís the only way you can find a card that has many pipelines, and a fast core. Honestly though the improvements of 512MB or 256MB are minimal, and it is impossible to judge a cards performance by how much memory it has.

Memory speed: Video cards use high speed memory that communicates with the core at very high speeds, usually around 1-1.5Ghz. This serves as a way to tell performance, by how fast the memory can receive images, so generally the faster the memory speed, the smoother the amount of frames per second you will see on the screen. However there are several types of memory that video cards use. Some use GDDR3, which is clocked at only 700 or so Mhz, but is a performance equivalent of 1500Mhz in DDR2 RAM. Others are equipped with DDR2 ram which is clocked at around 1+ Ghz.

Things you should know:

Integrated Video cards: In order to save money, lower end computers (like low end dells, and such) use integrated video cards. These cards have weak cores, little pixel pipelines (like 2 pipelines) but worst of all it borrows its memory from your systemís RAM, which generally kills performance. This usually confuses people because they will think they have 512MB of RAM, but when they install a program that requires 512MB the program doesnít install because the integrated card reserves a small portion of RAM for itself thus leaving the program to only recognize 500MB of ram. Some known integrated culprits are Intel Extreme Graphics, Intel extreme graphics 2, Intel Graphics Media Accelerator 900, and 950, ATI Radeon 200 Nvidia Geforce 6150. These are the worst performing cards on the market and should be completely avoided (although you will most likely only find them in prebuilt computers)

---
www.actsofgord.com . Long live Green Earth. Escalators can't break, they can only temporarily become stairs. I Love GREEN TEA
#10Eaglerulez(Topic Creator)Posted 8/20/2007 12:18:26 PMmessage detail
Multi Card Configurations: About 10 years ago multi card configurations where established by the Voodo 2 card . However after the company that made the card (3DFX) collapsed, the configuration faded, until 2004, when Nvidia (who bought 3DFX's intellectual property) reintroduced it ( Thanks to HappyKhicken for informing me of this) What essentially happens is if you have two PCI-E X16 slots (the slot that all modern day video cards use) You can put two of the same graphics cards in each slot. Then you connect the cards via a bridge that is provided, or a dongle, and the cards essentially work as a dual core processor. The data is sent to both cards to figure out, and process data. Once they are done, they recombine the data to form an image. However Nvidia graphics cards have to use SLI capable motherboards, while ATI cards have to use Crossfire compatible motherboards. SLI, and crossfire are each companyís names for multi card configurations. Recently both companies announced the addition of a third card. This card will be used to process physics, which will make the game more realistic, and take a load off of the CPU, and the two other GPUís. ATI has stated that older ATI cards that are weaker then the current two cards doing graphics may be used to process physics.

So thatís my graphics card explanation. One thing you should know is that unlike CPUís you only need to look at the technical specs. So if you see a card that has a core that is 50Mhz faster, then a card with the same memory speed, and pixel pipelines, then chances are that that card is truly faster then the one with the weaker core. Thereís a lot more to it, but this is the base of it all.

---
www.actsofgord.com . Long live Green Earth. Escalators can't break, they can only temporarily become stairs. I Love GREEN TEA