Address Space

Question 1

Address Space

Answer 1

The number of wires on the address bus defines the maximum amount of RAM a CPU can theoretically address. This is called the address space. With a 32-bit address bus, for example, the maximum amount of memory the CPU can address is 232, or 4,294,967,296 bytes (4 GB). Modern CPUs have address buses larger than 32-bit, so they can support substantially more than 4 GB of RAM. (See “64-Bit Processors” later in this chapter for details.)

Question 2

Address Space

Answer 2

The number of wires on the address bus defines the maximum amount of RAM a CPU can theoretically address. This is called the address space. With a 32-bit address bus, for example, the maximum amount of memory the CPU can address is 232, or 4,294,967,296 bytes (4 GB). Modern CPUs have address buses larger than 32-bit, so they can support substantially more than 4 GB of RAM. (See “64-Bit Processors” later in this chapter for details.)

Question 3

Clock Speed and Multipliers

Answer 3

The CPU clock speed, or frequency, tells you how many calculation cycles a CPU can (in theory) execute per second. One calculation cycle per second is equal to 1 hertz (Hz). Of course, nobody measures clock cycles in hertz, but rather in billions of calculation cycles per second, or gigahertz (GHz).

Two things determine clock speed: the maximum speed of the CPU itself and the maximum speed that the motherboard can handle. The CPU speed is determined by the manufacturer and is set at the factory. The system crystal—a quartz crystal circuit that oscillates at a fixed frequency when fed current—sets the motherboard’s clock speed.

Early CPUs ran at the speed of the bus, but engineers realized the CPU was the only thing working much of the time; if they could speed up just the CPU, they could speed up the whole process. Modern CPUs run at a multiple of the motherboard clock speed. The system bus on my 3.4 GHz Core i7 system runs at 100 MHz, so the clock multiplier goes up to ×35 at full load to support the maximum speed of the CPU at 3.5 GHz.

In the old days, the system bus speed and the multiplier on the CPU had to be manually configured via jumpers on the motherboard. Today’s CPUs actually report to the motherboard through a function called CPUID (CPU identifier), and the system bus speed and multiplier are often set automatically. Many motherboards offer manual configuration, but this is primarily for tweaking systems.

Question 4

CPU Packages

Answer 4

All current CPUs come in a square package called a pin grid array (PGA), but variations exist within and among manufacturers. Collectively, Intel and AMD have used close to 100 variations of the PGA package over the years for hundreds of different CPU models. They’ve used names such as staggered-PGA, micro-PGA, ball grid array, or BGA (which uses tiny balls instead of pins), and land grid array, or LGA (which uses flat pads instead of pins).

Question 5

zero insertion force (ZIF) sockets

Answer 5

To make CPU insertion and removal easier, these sockets—officially called zero insertion force (ZIF) sockets—use a small arm on the side of the socket or a cage that fits over the socket to hold the CPU in place. ZIF sockets are easily identified by their squarish shape.

Question 6

T Socket

Answer 6

Most modern Intel sockets have official names such as “Socket T,” but they are usually referred to by names such as “Socket LGA 2066” to reflect the number of pins in the CPU socket. AMD’s latest sockets go by names such as “Socket AM4” and are generally not referred to by the number of pins on the CPU or in the socket.

Question 7

Hyper-Threading

Answer 7

Hyper-Threading Technology is Intel’s implementation of simultaneous multithreading, which executes multiple threads at the same time on a single processor core to increase performance. To the OS, a multithreaded processor looks like more than one CPU, but it’s limited by having just one set of CPU resources.

Question 8

64-Bit Processors

Answer 8

The address and general-purpose registers in a 64-bit CPU are typically 64 bits wide and can access tons more memory than a 32-bit CPU. Whereas 32-bit CPUs can only allocate 4 GB of RAM to a process, 64-bit CPUs already support hundreds of terabytes of RAM, even though they’re using just a fraction of what they can theoretically support. In practical terms, this greatly enhances the performance of programs such as video editors, which work with huge files. Most new CPUs are 64-bit, and 64-bit CPUs are even increasingly common in smartphones.

Question 9

Multicore CPUs

Answer 9

CPU clock speeds hit a practical limit of roughly 4 GHz around the years 2002–2003, motivating the CPU makers to find new ways to get more processing power for CPUs. Although Intel and AMD had different opinions about 64-bit CPUs, both decided at virtually the same time to combine multiple CPUs onto the same physical processor die. CPUs with more than one processor core are known as multicore CPUs.

The first batch of multicore processors combined two CPU cores onto a single chip, creating a dual-core architecture, but today’s CPUs often have four or more.

Question 10

Mobile CPUs

Answer 10

The inside of a laptop PC is a cramped, hot environment, where no self-respecting CPU should ever need to operate. Since the mid-1980s, CPU manufacturers have endeavored to make specialized versions of their processors to function in the rugged world of laptops. These are called mobile processors.

Intel and AMD create mobile versions of their CPUs that use less power than an equivalent desktop model. This enables the laptop’s battery to last longer and makes it easier to keep the CPU and laptop running cool. Modern CPUs also use throttling to reduce processing cycles and lower clock speeds when temperatures climb or the system is idle.

Question 11

Throttling

Answer 11

Modern CPUs also use throttling to reduce processing cycles and lower clock speeds when temperatures climb or the system is idle.

Question 12

Virtualization Support

Answer 12

Intel and AMD have built-in support for running more than one operating system at a time, a process called virtualization. The key issue from a CPU standpoint is that virtualization used to work entirely through software. Programmers had to write a ton of code to enable a CPU that was designed to run one OS at a time to run more than one OS at the same time.

Think about the issues involved. How does the memory get allocated, for example? Or how does the CPU know which OS to update when you type something or click an icon? With hardware-based virtualization support, CPUs took a lot of the burden off the programmers and made virtualization a whole lot easier.

Question 13

Integrated Graphics Processing Unit

Answer 13

The video-processing portion of the computer—made up of the parts that put a changing image on the monitor—traditionally has a discrete microprocessor that differs in both function and architecture from the CPUs designed for general-purpose computing. The generic term for the video processor is a graphics processing unit (GPU). Because graphics processors can handle certain tasks much more efficiently than the standard CPU, integrating a GPU into the CPU enhances the overall performance of the computer while at the same time reducing energy use, size, and cost. With the proliferation of mobile devices and portable computers today, all of these benefits have obvious merit.

Both major CPU manufacturers have released CPUs with integrated GPUs. The architecture differs in some ways, such as how they use the cache on the chip. The Intel HD Graphics integrated into many Core i3/i5/i7 processors, for example, has the CPU cores and the GPU core sharing the “last level cache,” which is either L2 or L3, depending on the processor. With the AMD accelerated processing unit (APU), such as the AMD Fusion, the GPU has access to all levels of cache on the CPU.

Question 14

AMD accelerated processing unit (APU)

Answer 14

the GPU has access to all levels of cache on the CPU

Question 15

Overclocking

Answer 15

Most modern systems use the CPUID to set the clock speed and clock multiplier automatically for any CPU you install, so all but true enthusiasts are done with installation at this point. Boot up the PC, and watch the screen. Assuming you’ve done everything correctly, your system will start without any undue alarms or errors. If, however, something has gone wrong, you need to turn to troubleshooting.

Some motherboards enable you to adjust CPU settings manually by changing a UEFI BIOS setting, or by using other software. Many enthusiasts deliberately change these settings to enhance performance; this is called overclocking.

You can make a couple of adjustments to overclock successfully. First, increase the bus speed for the system. Next, you often have to increase the voltage going into the CPU just a little to provide stability. Always do only one thing at a time, document your change, and then reboot.

Question 16

Overridng

Answer 16

Overriding the defaults can completely lock up your system to the point where even removing and reinstalling the CPU doesn’t bring the motherboard back to life. (There’s also a slight risk of toasting the processor, although all modern processors have circuitry that quickly shuts them down before they overheat.)

Question 17

CMOS clear

Answer 17

Most motherboards have a jumper setting called CMOS clear that resets the UEFI/BIOS to default settings. Before you try overclocking on a modern system, find the CMOS clear jumper and make sure you know how to use it.

Question 18

Installation Problems

Answer 18

Look (and listen) for the following signs of installation problems:

It’s normal for the CPU fan speed to fluctuate, but if the fan doesn’t spin at all for more than five seconds or so, turn the system off and make sure that the fan is plugged in correctly.
If the fan is spinning but the PC doesn’t boot, make certain that any jumpers governing speed or voltage are set correctly.
If, after a few minutes, you hear a loud alarm, this is an indication that the system is overheating. Turn it off immediately, and check your fan installation to make sure you’ve got good CPU/heat sink contact and that the fan is functioning properly.
If the system does not boot and the fan does not spin, check to ensure that the CPU is seated properly and that no other components or cables have been unseated or disturbed.

Question 19

Heat Issues

Answer 19

Messy cables
The warning signs will include: spontaneous reboots, especially on warmer days or with the heat blasting in winter time; occasional alarms from the tinny PC speaker; lockups; and excessively loud fans. The fix is easy. Check the fans, and replace when necessary.

Question 20

Note

Answer 20

Note: Installing good fans in a case can make a world of difference to the user experience. No one wants to use a new or upgraded PC that sounds like a helicopter! I always install the quietest fans I can get, generally as big as I can get into a case. If you have a case with options for 80 mm or 120 mm fans, for example, go for the latter. Bigger fans move more air at lower speeds. Check the Supplementary Material links for some excellent product links.

Question 21

RAM

Answer 21

Manufacturers have produced RAM in a wide variety of physical form factors (sometimes called packages) and technologies over the years. So first we’ll look at several types of RAM, including those covered on the CompTIA A+ certification exams.

Question 22

DIMM

Answer 22

Dual inline memory modules (DIMMs) dominate the current desktop market, including a popular variant used in laptops, small outline DIMMs (SO-DIMMs). Full-sized DIMMs have 184, 240, or 288 pins; are 64-bits wide; and have a capacity of up to 256 GB at this writing (although that’s not a common size; most DIMMs are somewhere between 8 and 32 GB today).

Question 23

SO-DIMM

Answer 23

The SO-DIMM variant comes in 72-pin, 144-pin, 200-pin, and 204-pin configurations. Generally, you can find SO-DIMMs with capacities comparable to those of their full-sized counterparts.

Question 24

RAM Packaging

Answer 24

RAM manufacturers produce an amazing variety of RAM packages. You’ll find sticks that offer the same capacity, for example, but vary in number or location of memory chips. One stick, for example, might have four chips on the front and four on the back—a double-sided stick—whereas another might have eight chips on the front only—a single-sided stick. A third one might be single-sided with only four chips, and a fourth, double-sided with 16 chips!

The general rule on memory packaging is never mix packages if you can help it. It leads to overall system instability in enough cases that a prudent tech sticks with uniform RAM sticks.

Question 25

Memory Channels

Answer 25

Current motherboards use multichannel memory configurations to increase memory performance. With only a 64-bit data bus, DIMMs deliver 64 bits of data at a time. That's pretty logical! This is a single memory channel. Memory controllers can be designed to use two channels simultaneously, a dual-channel configuration. As you'd expect, triple-channel and quad-channel configurations use three and four channels, respectively. In dual-channel mode, the memory controller's 128-bit-wide data path uses two DIMMs per bank. In triple-channel mode, the 192-bits data path requires three DIMMs per bank. On multichannel systems, the DIMM slots for a given bank are typically the same color. Use identical RAM sticks to populate a bank for multichannel memory modes, and always check the motherboard book for details. Motherboards that support a multichannel mode can also run in the lower-channel modes, but you'll get the best performance using the highest mode the motherboard supports.

Question 26

RAM Technologies

Answer 26

RAM comes in two forms: static RAM (SRAM) and dynamic RAM (DRAM). SRAM is faster than DRAM because it never needs to be refreshed; however, it is also much more expensive. SRAM is, therefore, reserved for special purposes, such as L1, L2, and L3 system cache memory. DRAM is used for system memory and needs to be refreshed every few nanoseconds (ns). During this time, it is inaccessible to the CPU. This interval is called a wait state. Faster RAM has a shorter wait state. Next, I'll describe the various types of DRAM that you'll need to know for the 1001 exam.

Question 27

DDR SDRAM

Answer 27

The original type of SDRAM ran at the same speed as the system bus; it was synchronous with the bus. The "S" in SDRAM stands for synchronous. It is also known as single data rate SDRAM. The next technology leap past that was double data rate SDRAM, usually referred to simply as DDR.

Question 28

DDR2

Answer 28

DDR2 succeeded the DDR SDRAM standard. Improvements in DDR2's electrical characteristics enable it to run faster than DDR while using less power. The big speed increase from DDR2 comes by clock doubling the input/output circuits on the chips instead of speeding up the memory itself. Speeding up the input/output and adding special buffers (sort of like a cache) makes DDR2 run much faster than regular DDR. DDR2 comes in two packages: 240-pin DIMM and 200-pin SO-DIMM. DDR2 modules and systems are not compatible with DDR memory. You'll find motherboards running both single-channel and dual-channel DDR2. The table below shows common DDR2 speeds.

Question 29

DDR3

Answer 29

DDR3 boasts higher speeds, more efficient architecture, and around 30 percent lower power consumption than DDR2. DDR3 uses these packages: 240-pin DIMM (slotted differently than 240-pin DDR2) and 204-pin SO-DIMM; neither will fit into a DDR2 socket. DDR3 doubles the size of the prefetch buffers, giving its bandwidth a huge boost when reading contiguous data. The table below lists common DDR3 speeds. Many DDR3 systems support dual- and triple-channel memory configurations.

Question 30

DDR4

Answer 30

DDR4 arrived on the scene in late 2014 with much fanfare—and slow adoption. DDR4 offers higher density and lower voltages than DDR3, and it can handle faster data transfer rates. In theory, manufacturers could create DDR4 DIMMs up to 512 GB. As of this writing, DIMMs running DDR4 top out at 16 GB, like DDR3, but run at only 1.2 V. (There's also a performance version that runs at 1.35 V, and a low-voltage version that runs at 1.05 V.) DDR4 uses a 288-pin DIMM, so they are not backwardly compatible with DDR3 slots. DDR4 SO-DIMMs have 260 pins that are not compatible with DDR3 204-pin SO-DIMM slots. Some motherboard manufacturers have released boards that offer support for both DDR3 and DDR4 by providing both slot types. With DDR4, most techs have switched from bit rate to megatransfers per second (MT/s), a way to describe the bandwidth as the number of data transfer operations happening at any given second. For DDR4, the number is pretty huge. The table below shows common DDR4 speeds and labels.

Question 31

Parity

Answer 31

These types of RAM use special circuitry to detect and, in some cases, correct errors in data. Parity RAM, the older technology, used a dedicated parity chip mounted on the RAM stick that added an extra bit—the parity bit—to each byte of data. Parity checking was useful in the early days of desktop computers when DRAM had a relatively high failure rate.

Question 32

ECC RAM

Answer 32

ECC RAM contains circuitry that not only detects errors, but corrects them on the fly without causing interruption to system processes. ECC RAM is common on performance-enhanced workstations and server systems. Note that the computer's motherboard and basic input/output system (BIOS) must be designed to support either parity or ECC RAM. Most motherboards that are designed for parity or ECC RAM can be configured in BIOS to disable error checking, thus enabling you to install RAM that is not parity or ECC RAM.

Question 33

Common Symptoms of RAM Errors

Answer 33

The majority of genuine memory errors happen after you've made a change to your PC's hardware and restart the system, at which time it goes through the power-on self test (POST) routine. Common startup errors include the following: System won't boot, accompanied by a loud, continuously repeating beep System boots, but the display screen is blank System boots, misreports RAM size System boots, reports memory or address error Continuous reboots after installing new RAM Once you've determined that the error you're getting is a genuine RAM error, you can determine the cause and take corrective measures.

Question 34

Common Causes of RAM Errors

Answer 34

Faulty RAM is less common, but it does happen. Sometimes it comes out of the factory that way, but it's more likely that the RAM was damaged somehow during transport or storage. Don't forget that our old enemy, ESD, can cause good RAM to go bad in no time flat. To test for bad RAM, replace it with known good RAM and see if the error stops or recurs. You can also use the included Windows Memory Diagnostics tool to check your system for hardware memory errors.

Question 35

Common Causes of RAM Errors

Answer 35

The majority of genuine memory errors happen after you've made a change to your PC's hardware and restart the system, at which time it goes through the power-on self test (POST) routine. Common startup errors include the following: System won't boot, accompanied by a loud, continuously repeating beep System boots, but the display screen is blank System boots, misreports RAM size System boots, reports memory or address error Continuous reboots after installing new RAM Once you've determined that the error you're getting is a genuine RAM error, you can determine the cause and take corrective measures.