Processors & CPUs: Architecture, Sockets, and Specifications

The CPU is the engine — it executes instructions, performs calculations, and orchestrates everything your computer does. You already know sockets matter (from the motherboard lesson), but understanding why a socket exists, how CPUs differ in architecture and instruction sets, and what "cores" and "clock speed" really mean will give you the full picture of CPU compatibility, performance, and thermal constraints.

CPU Sockets

A CPU sits in a socket on the motherboard, and the socket is the physical interface that determines which processors can be installed. Different sockets have different pin layouts, electrical standards, and power delivery schemes — so a CPU built for one socket won't fit or function in another.

The two dominant sockets in consumer and business desktops today are:

Intel LGA 1700 — the socket for Intel's 12th gen (Alder Lake), 13th gen (Raptor Lake), and 14th gen (Raptor Lake Refresh) processors. LGA stands for "Land Grid Array" — the CPU has flat contacts, and the socket has pins that the CPU lands on. A 1700-socket board pairs with a 1700-socket CPU; you cannot mix them with AMD. Note that newer Intel architectures, like the Core Ultra 200 series (Arrow Lake), have migrated to the LGA 1851 socket, so LGA 1700 does not support the latest generation of Intel desktop processors.

AMD AM5 — the socket for AMD's Ryzen 7000 series and later. AM stands for "AMD Socket M"; the CPU has pins, and the motherboard socket has contact pads. Like Intel, AM5 is AM5-only — it doesn't accept Intel chips. AMD has committed to long-term AM5 support, with backward and forward compatibility within the socket's lifespan.

There are older sockets (Intel LGA 1200, AMD AM4) that are still in circulation, but when you're specifying a system, verify that the motherboard socket and CPU socket align exactly. They are not backward- or forward-compatible across socket generations — a board designed for LGA 1700 will not accept LGA 1851 processors, and vice versa.

Instruction Sets: x86/x64 vs ARM

Not all CPUs use the same instruction set — the fundamental language they speak when executing code.

x86/x64 — the instruction set used by Intel and AMD desktop and server processors. "x86" refers to 32-bit instructions; "x64" (also called "x86-64") refers to 64-bit instructions. Historically, x64 dominated desktops and servers, but the landscape is shifting. For traditional desktop and server builds, x64 remains the standard, as 64-bit processors can address far more memory and execute more powerful instructions. A CPU advertised as "x64" or "64-bit" is an x64 processor.

ARM — a completely different instruction set, used traditionally in mobile phones, tablets, and embedded systems. However, ARM has recently entered high-performance consumer and cloud markets: Apple's M-series chips (ARM-based) power modern MacBooks, Qualcomm's Snapdragon X processors bring ARM to Windows laptops, and AWS Graviton processors are gaining adoption in cloud data centers. ARM processors are inherently power-efficient, which is why they excel in mobile and are becoming attractive for energy-conscious server workloads.

While ARM and x64 use different instruction sets, the gap in software compatibility is narrowing. Modern operating systems use advanced translation layers to run x86/x64 applications on ARM hardware:

• Apple Rosetta 2 — allows macOS to run Intel-compiled software on Apple Silicon (ARM)
• Microsoft Prism — enables Windows on ARM devices to run x64 applications

These translation layers are efficient enough for most tasks, though performance can vary depending on the workload. So while ARM and x64 are fundamentally different at the hardware level, the user experience is increasingly seamless.

For the A+ exam and for building traditional desktop systems, you'll primarily encounter x64. However, you should recognize that ARM is fundamentally different from x86/x64 and understand that it's increasingly relevant in consumer and enterprise computing.

Cores and Threads

A CPU has one or more cores — independent processing units that can execute instructions in parallel. Modern CPUs have multiple cores: a budget processor might have 4 cores, a mainstream system 8, and high-end systems 12, 16, or more.

A thread is a sequence of instructions that the CPU executes. On a single-core CPU, one core runs one thread at a time. But many CPUs support simultaneous multithreading (SMT), which allows a single core to execute two threads at roughly the same time, switching between them extremely quickly.

Simultaneous Multithreading (SMT) and Modern Architectures:

Traditionally, the rule was simple: each core with SMT = 2 threads. However, modern CPUs are increasingly using hybrid architectures that mix different core types:

• Performance cores (P-cores) — high-clock, high-power cores optimized for single-threaded and lightly-threaded workloads.
• Efficiency cores (E-cores) — lower-clock, lower-power cores optimized for background tasks and parallelizable workloads.

This changes the thread-counting math. For example, a modern Intel Core Ultra CPU might have 8 P-cores and 16 E-cores = 24 threads total. The P-cores each have SMT (contributing 2 threads per P-core), while the E-cores do not have SMT (contributing 1 thread per E-core).

Important note on Intel Hyper-Threading: Intel has been refining its approach to SMT. While Hyper-Threading remains standard on AMD's Ryzen processors and older Intel generations, modern Intel architectures — including Lunar Lake and the latest Core Ultra Series 2 (Arrow Lake) — have eliminated Hyper-Threading on their performance cores. This design choice reflects Intel's focus on core count and efficiency rather than relying on SMT to boost thread capacity.

The operating system sees and schedules threads. The practical upshot: more cores and threads improve performance on heavily multithreaded workloads (video encoding, compilation, large spreadsheets, simultaneous applications), but the exact thread count depends on the architecture.

SMT Branding: AMD uses Simultaneous Multithreading (SMT); Intel traditionally used Hyper-Threading (though it's now absent from recent P-cores).

Clock Speed and Cache

Clock speed (measured in GHz — gigahertz) determines how many instructions a CPU can execute per second. A 3.0 GHz processor executes 3 billion instructions per second, all else being equal. A faster clock is faster — but clock speed alone doesn't determine performance. A 5 GHz dual-core will not beat a 3 GHz 16-core on a multithreaded workload, because the cores do more work even at a slower pace.

Cache is ultra-fast memory built into the CPU that stores recently-used data and instructions. When the CPU needs to fetch data from RAM, it's significantly slower than fetching from cache. Modern CPUs have multiple levels:

• L1 cache — smallest and fastest (a few kilobytes per core). Modern L1 data caches typically incur a latency of 4 to 5 clock cycles (not a single cycle, despite popular belief). Still dramatically faster than RAM.
• L2 cache — larger than L1, with higher latency (typically 10–20 cycles). Modern mid-to-high-end architectures feature 1MB to 3MB of L2 cache per core, far beyond the older "hundreds of kilobytes" standard.
• L3 cache — largest of the on-CPU caches (several to tens of megabytes), shared across all cores, with even higher latency (typically 40–75 cycles).

More cache helps when the workload has good spatial locality — when the CPU repeatedly accesses the same region of data. A CPU with more cache will outperform a clock-for-clock identical CPU with less cache on memory-intensive workloads. This is why modern high-performance CPUs dedicate so much die space to large L2 and L3 caches.

For the exam, the takeaway is: cache size and hierarchy matter for performance. Larger cache generally means better performance under memory-intensive workloads, even if the clock speed is slightly lower, because cache hits are orders of magnitude faster than RAM fetches.

Integrated Graphics

Many CPUs include integrated graphics — a simple GPU built into the processor that can handle 2D drawing, video playback, and light 3D gaming without needing a discrete graphics card.

Integrated GPU — included in the CPU, shares system RAM as video memory, uses CPU power budget. It's slower than a dedicated GPU but sufficient for office work, light gaming, and everyday tasks. Intel's integrated graphics have been branded as "Intel HD Graphics," "Intel Iris," and more recently "Intel Arc Graphics" (in newer Core Ultra processors). AMD calls theirs "Radeon Graphics" (in Ryzen) or "Radeon Vega" (older gen).

Dedicated GPU — a separate expansion card (discrete graphics card) that has its own dedicated memory and power supply. Far faster than integrated graphics, and essential for heavy gaming, 3D design, or video encoding. If your motherboard and budget allow, a discrete card is worth it for demanding workloads.

If a CPU lacks integrated graphics (some high-end server CPUs do), you must have a dedicated GPU to see anything on the monitor. Most consumer CPUs include integrated graphics, so you can boot and use the system without a dedicated card.

Virtualization Support

Many CPUs include hardware virtualization support — a set of features that allow the processor to efficiently run virtual machines (VMs). These features are optional extensions, and they're often disabled by default in BIOS.

Intel VT-x — Intel's virtualization extension. AMD-V — AMD's virtualization extension.

If you're running Hyper-V on Windows, VMware, VirtualBox, or other VM software, these extensions dramatically improve VM performance. Older CPUs or those with virtualization disabled will run VMs, but much more slowly (the CPU has to emulate hardware instead of directly supporting it).

For general-purpose desktop work, virtualization support is nice-to-have. For server or development work with VMs, it's critical.

Thermal Considerations and Cooling

CPUs dissipate heat — the faster and more powerful the CPU, the more heat it generates. This is quantified as TDP (Thermal Design Power), measured in watts. However, understanding TDP in the modern context is important:

What TDP represents: TDP is the thermal output a CPU is designed to dissipate at base clock under sustained, typical workloads. For example, a CPU rated at 125 W TDP is expected to output roughly 125 watts of heat when running at its base frequency with a representative workload.

The boosting caveat: Modern CPUs use aggressive boosting algorithms (Intel's Turbo Boost, AMD's Precision Boost, and power management schemes like PL1/PL2 and PPT) that allow the CPU to exceed its base clock significantly. Under these boost conditions, real-world power consumption and heat output can far exceed the rated TDP — sometimes by 50% or more. For example, a 125 W TDP CPU might draw 180+ watts and generate corresponding heat during all-core boost under gaming or rendering.

Practical implications:

• A budget CPU might have a 35 W TDP; a high-end desktop CPU might be 125 W or higher.
• The power supply and cooling system must be sized to handle not just the TDP, but the sustained boost power of the CPU under real-world workloads.
• When choosing a cooler or PSU, check the CPU manufacturer's peak power specifications (not just TDP) for a more accurate sizing.

CPU coolers come in two types:

Stock cooler — shipped with the CPU (if you buy retail, not OEM). Usually adequate for normal use, but limited in noise and cooling capacity.

Aftermarket cooler — a replacement air or liquid cooler that cools better and often runs quieter. Popular for overclocking or quieter systems.

Installing a cooler is straightforward: mount the cooler to the CPU socket using clips or screws, apply thermal paste between the CPU and cooler (or use a pre-applied pad), and ensure the cooler is flush and secure.

Putting It Together

A CPU is the core compute engine. To build or support a system, you need to:

1. Match the socket — CPU socket must match motherboard socket.
2. Check the instruction set — all consumer desktops are x64.
3. Size cores and threads for your workload — more cores = better multithreading, more single-threaded speed = better for light tasks.
4. Pick a cooler — sized to the CPU's TDP.
5. Enable virtualization in BIOS if you plan to run VMs.
6. Trust integrated graphics for everyday use, or add a dedicated GPU if performance matters.

The check questions below test these specifications and compatibility rules.