How to Pick the Perfect Personal Computer


As someone in the IT industry, I find it highly annoying when vendors and retailers don’t provide me with the information I need to make educated purchasing decisions. Many PC buying guides exist, but they either aren’t flexible enough to accommodate your unique needs or are too focused on a narrow subset of available technologies and thus quickly become obsolete. They tend to be too general, covering only the most essential aspects of the specifications, leaving you confused when a salesperson or consultant offers alternative solutions or explanations. This tutorial is written for the intermediate PC user. If you are a master, you should be well-versed in all this.

To demonstrate how simple it is to be misled, one of the most popular PC manufacturers recently advertised a model of their ‘xyz-wizbang’ computer that boasted a whopping 12 GB of memory, an Extreme Intel Quad core processor, and four separate graphics cards. Wow, that’s a big deal, right? I was skeptical of the low price. If you follow the link for more information, then click the options button, then click the technical specifications button, then read them very carefully, you’ll learn that it comes with only 3GB of memory but can support up to 12GB, which it comes with a regular Intel processor but you can upgrade to the Extreme version, and that it supports four graphics cards but only includes one. Without knowing the specifics, the price probably seemed very appealing.

Give a hungry man a fish, and you feed him for a day; teach him to fish and feed him for life; this is one of my favorite sayings. That pretty much sums up the whole point of this manual. With just a little more knowledge, you can precisely define your needs, critically analyze the claims made by various vendors and retailers, and walk away with a high-quality personal computer capable of handling your tasks. Learning this method of purchasing has the added benefit of being timeless; the principles I outline here have generally been applicable since the 1980s. The key to comprehension is making sense of the jargon, which I will do as much as possible by explaining in layman’s terms. I’m occasionally asked, “What is the difference between 4GB RAM and 300GB of hard disk, and which do I need?” indicates that more education is necessary. ‘. You really should read this right now if you fit this description… Understanding the PC’s parts and how they work is essential before making any decisions.

The Central Processing Unit (CPU) is the “brain” of your computer; it processes your instructions at a rate of millions per second. Like my wife, modern processors will have multiple cores and go by names like “Dual core” (2 cores), “Quad-core” (4 seats), and “Octa core” (8 cores), making them capable of performing multiple tasks at once. To illustrate, suppose I ask my computer to give me a million lottery numbers, which takes eight seconds to complete (it would finish in the blink of an eye). With a dual-core processor, I could have half a million numbers from one core and the other half from the other core in just four seconds. Using the same reasoning, doing the same thing on a Quad-core would take just two seconds.

The process of dividing up work in this way is known as multithreading. If an enormous task can be partitioned into smaller tasks running in parallel, then more cores are preferable. There is a catch, though. To determine whether a Dual, Quad, or Octa-core processor is right for you, first determine whether or not your tasks can benefit from being divided up in this way and then check with the software vendors whose products you use. The CPU’s clock speed, measured in gigahertz, is another performance factor (cycles per second). The typical rate of a modern CPU is between 1.8 and 3.3 GHz. A multicore CPU’s processing units will all operate at the same frequency. Therefore, if a company advertises a computer as “12GHz,” they are likely multiplying the clock speed by the number of cores present (4 cores by 3GHz). Maybe they want everyone to think their computers are insanely faster than everyone else’s. A higher clock speed means quicker execution times, regardless of the number of cores present. Therefore, higher GHz should be available if the benefit of more hearts is unavailable.

The Memory (or RAM) – The CPU stores busy work while the computer is on. The computer’s memory is the fastest storage location, so that’s where it’ll do your job if given a choice. When RAM is exhausted, the computer must resort to storing data on the hard drive (a process known as paging to virtual memory), which causes a significant slowdown. Therefore, invest in as much memory as you can comfortably afford. The speed of the thing is secondary. Memory speed is measured in a combination of MHz, Type, and Latency, with a recommended minimum of 2GB (DDR2) or 3GB (DDR3) for general light use and 4GB (for DDR2) to 6GB (for DDR3) or more for demanding games or applications. Keep in mind that bandwidth is not the same as speed. Think of the memory bus as a highway. Each car will reach its destination quickly on a one-lane road going 70 miles per hour, but there will be a cap on the number of vehicles on the road. However, a four-lane highway, even at a slower speed of 55mph, will get more cars to their destination in the same period, even though each vehicle will take longer to reach its destination due to the higher bandwidth available. However, they are related because even a single traffic lane can equal a multilane highway’s bandwidth. Depending on the

specifics of the task at hand, your computer may need more bandwidth, processing power, or some combination of the two. Bandwidth (in this case) is also affected by the type of memory bus technology employed. RAM with DDR, DDR2, DDR3, etc. DDR stands for “Dual Data Rate,” and the number following denotes the total number of parallel communication channels used. The greater the media it uses, the greater the available bandwidth. So, compared to DDR, DDR2 has double the capacity, and compared to DDR3, DDR3 has an additional 150 percent. Since the latency penalty increases proportionally with each DDR level, the rule of thumb is to double the memory speed in MHz between each DDR level. DDR2 memory at 800MHz is required to match the speed of DDR 400MHz memory, and it offers much higher bandwidth thanks to its use of two independent channels.

Remember that the latest technology isn’t always better and that DDR3 1333MHz isn’t necessarily superior to DDR2 1100MHz for the reasons given. As of this writing, a new computer should have DDR3 memory running from 1333MHz to 1600MHz. Alternatively, DDR2 at 800MHz or higher. Latency is challenging to explain, but if you’re doing intensive work, you should use memory with low latency. While bandwidth is more crucial for tasks like video encoding and file transfers, processing speed is more important for gaming and general work.
The Hard Disk (HDD) or Storage is a slower but more permanent storage than memory. All your files will still be accessible after you turn your computer off. The store size is the only criterion that matters to most consumers. The price of mechanical hard disks has dropped, so you can buy one with far more storage space than you’re ever likely to use without breaking the bank. If you regularly work with large amounts of media (photography, databases, or videos), it’s recommended that your new PC has at least 300 GB of storage space. The speed at which data can be transferred between the hard disk and the computer is determined by four factors:

1) the disk’s rotational speed (typically 7200 rpm, but up to 15000rpm), 2) the amount of data stored per square inch of the disk’s surface (the platter), i.e., areal density, 3) the disk’s interface to the computer, which should now be SATA-II and can transfer data at up to 300MB/s, and 4) the drive head’ (average seek time usually around 8ms). Typically, the latter is the least crucial of the two. The solid-state disk (SSD) is a new technology that is rapidly developing. It functions similarly to a hard drive but has no moving parts, instead relying on a type of permanent memory (flash memory, like that found in USB sticks) to store the computer’s information. It’s a specialized field and a complex topic in and of itself. Cheap solid-state drives (SSDs) only outperform an excellent hard drive in specific situations, so they are not yet cost-effective for most users.

The Graphics Card (GPU, Graphics Processing Unit) — Skip this section if you don’t do any work requiring 3D graphics, such as gaming, video editing, CAD, or design. This includes photographers because the CPU, not the graphics card, is still the primary bottleneck for most photographic work. The need for real-time 3D graphics processing has motivated significant technological investment in graphics cards. Because of this demand, the GPU in high-end cards may now have more processing power than the PC’s central processor. The two leading players in this space are ATI and nVidia. Both are top-notch and provide comparable value in terms of quality and cost. They’re continually trading places regarding who’s the fastest in the world. A second or third-tier card from the top of the range should suffice for most people, and they’ll be a lot cheaper than the top card unless you need the entire card. The graphics processing unit (GPU) is the central processing unit of a graphics card. This GPU’s performance is reported regarding cores (streams) and MHz, just like your primary processor. Faster computers typically have more seats and a higher clock speed.

You cannot accurately compare core for core, MHz for MHz, between ATI and nVidia GPUs due to architectural differences in their design. You need to stick to the same provider to make a comparison like that. Almost all modern video cards have a pair of Digital Visual Interface (DVI) monitor outputs, allowing you to connect two displays simultaneously. Screen resolution is another crucial distinction between cards; at the very least, you should hope for support for 1600×1200 pixels. However, higher resolutions will necessitate more onboard memory (to serve as a frame buffer) and a more potent card to render the larger display quickly. The ATI Crossfire and Nvidia SLI inter-GPU communications standards and interfaces allow for the installation of multiple graphics cards. Two graphics processing units (GPUs) can also be installed on a single graphics card, with SLI or Crossfire software to facilitate communication between them. Remember that this approach to improving graphics performance does not scale up linearly. Each additional card (or GPU) will likely only provide an additional 40-60% performance gain over a single card (or GPU).

Memory is the key to determining whether you should use Windows 32 or 64-bit (or the OS). The number of ones limits a computer’s ability to handle large numbers and zeroes it uses to express them in the binary system of 1s and 0s. The sequential numbers that identify each memory address in a computer are analogous to street addresses in the postal system. While 32 bits allow for a maximum of 4GB of memory, modern computers require much more storage space. Now that 64-bit addresses are the norm; we can refer to 17.2 billion gigabytes (or 16 exabytes) of memory space. It will be quite some time before we have to make that adjustment again. Though most modern motherboards support up to six sticks of RAM, the practical limit is 24GB because RAM densities currently top out at 4GB per stick.

There are three types of optical drives: a DVD rewriter (DVD-RW), a DVD/Blu-ray reader (BD-R / DVD-R), and a Blu-ray rewriter (BD-RW). Blu-ray has two significant advantages over DVD: 1) it uses Blue laser light, which records at a much higher density than a red DVD laser, allowing for a greater capacity of 50GB over a DVD’s standard 8.5GB, and 2) Blu-ray movies can be played on a Blu-ray Reader or Rewriter. If you’re not interested in either of those things, then Blu-ray is unnecessary, and you can save yourself some cash. Blu-ray players are used mainly by movie buffs and video editors.

Connectors (called “interfaces” for short) that allow devices to communicate with your computer include Firewire (IEEE1394), USB 2.0 (USB 2.0), eSATA, and high-definition multichannel audio. So take this as the bare minimum.

As for the housing, any standard computer case will do. Choose a standard size and design instead of a branded case if you anticipate upgrading frequently. Many popular brands’ cases are purposefully made to be disposable because they lack upgrade space, come in non-standard sizes, or are otherwise difficult to work on and disassemble. Ensure it is quiet and has adequate cooling (ideally front, rear, and top fans). Instead of pressed steel, if you can afford it, go with a toolless aluminum case. They’re more aesthetically pleasing and straightforward to update. Verify that the chassis includes sockets for at least some of the PC interfaces, e.g., While I will touch on some of the more advanced factors that affect performance, such as USB2, audio, Firewire (IEEE1394), and eSATA, I can only scratch the surface here. If you want to know more, check out my other articles where I discuss these topics in greater detail (some justify having a whole article dedicated to them).
The Front Side Bus (FSB), HyperTransport (AMD), or Quick Path Interconnect (QPI) is a bus that connects the processor to the memory. Over the past decade, it has gone from the tens of MHz to 1600 MHz, officially reached by the last processor supporting it (the Core 2). (though with overclocking faster speeds where possible). Currently, QPI speeds can reach up to 6400GT/s.
When purchased from a reputable and knowledgeable vendor, overclocked computers are the fastest to buy while providing excellent value.
Sounds hot, but SpeedStep (or EIST) isn’t what it seems. When processor utilization is low, this technology reduces power consumption.

Through a feature called “HyperThreading,” a single CPU can trick Windows into thinking it has access to multiple cores. This means that a HyperThreaded quad-core processor will appear to have eight physical bodies. The extra virtual centers are just using the processor’s idle resources, so the benefit is not as great as it may seem initially. Therefore, it is highly likely that there are no unoccupied processor bits, and it will be of no use whatsoever if you are performing intensive work. In practical terms, you can expect a 10-15% performance boost. Under heavy loads, it can even slow down the processor by 5 percent or so.

Turbo is making a comeback in newer Intel Core i7 processors and operates in a manner opposed to SpeedStep. This means that the PC will automatically increase the processor’s clock speed by 200-300MHz in situations where the CPU is under heavy load. Its supposed benefits are greatly exaggerated.
The Redundant Array of Independent Disks (RAID) is a disk controller technology that can increase transfer speeds and provide redundancy in a drive failure. Available in several flavors, including RAID0 for speed, RAID1 for reliability, and RAID5 and RAID10 for both qualities.
You can now confidently call out vendors or retailers for providing incorrect or insufficient specifications or engaging in questionable sales practices. Get what you pay for by being specific about your needs and expectations. Make sure it’s up-to-date and has just the right level of detail for your purposes. Most importantly, relax and enjoy yourself.
Alan leads the technology team at UK-based computer manufacturer Cryo Performance Computers ( He is responsible for developing new methods for designing PCs for games and other resource-intensive contexts. Cryo PC provides high-end PCs, such as extreme gaming PCs and custom-built PCs for professionals.

Read also: Multilevel Monitoring Equipment – Light Beer Still Active and Powerful?