Lesson 1.1
Computer components and their functions
A computer is a set of parts doing different jobs together. Learn to recognize a part, explain its job, and connect that job to something a user does. Model numbers and specification labels provide evidence; appearance alone does not prove compatibility.
This chapter supports identification and reasoning. Physical inspection needs your instructor’s approval and the exact computer’s service manual. Do not open a power supply, remove a fitted CPU cooler, or open a working storage drive for this reading.
01
The board connects; the processor works
The motherboard provides the connections between major components. A Central Processing Unit (CPU) executes program instructions. Seeing both parts in one photograph does not make them the same part: a large circuit board may contain a socket for a separate processor.
A PCIe Slot provides an expansion connection, such as the connection for an add-in graphics card. Fit is only one check. The system’s documentation must also support the device, its power needs, and its other requirements.
A Heatsink transfers heat away from a hot component. A cooler may hide the CPU beneath it; identifying the visible cooler is different from seeing the processor itself. Leave installed cooling equipment in place unless an approved service task specifically requires removal.
Pause and explain
Two motherboards both have a PCIe slot. Does that prove that the same graphics card is suitable for both computers?
Compare your reasoning
No. The shared connection is a starting point. Check the exact systems’ support, available space, power supply and connections before recommending a card.
Technical references for this section
02
Working on a file is different from saving it
RAM is working memory for active programs and data. A Storage Device retains saved data when power is off. The two capacities answer different questions: how much active workspace is available, and how much saved material can be kept.
A Hard Disk Drive (HDD) uses magnetic storage and moving mechanical parts. A Solid-State Drive (SSD) stores data electronically without that moving drive mechanism. Both are storage devices. “SSD” does not, by itself, identify the connector, capacity, condition or compatibility of a particular drive.
- Open
Saved program and file data are read from storage.
- Work
The CPU works with instructions and data in active memory.
- Save
The application writes the saved result to storage.
Pause and explain
A computer has plenty of free drive space but struggles with many active apps. Why is free space not enough evidence to rule out a memory problem?
Compare your reasoning
Drive space and active memory describe different resources. Inspect actual memory demand and other evidence rather than treating free storage as spare RAM.
Technical references for this section
03
A GPU is not the whole graphics card
A Graphics Processing Unit (GPU) is a processor used for graphics and other parallel workloads. A Graphics Card is an assembly that can include a GPU, memory, cooling and connectors. The GPU itself may be hidden by the cooler in a photograph of the card.
Some systems use integrated graphics instead of a separate add-in card. Accelerated Processing Unit (APU) is AMD’s term for a design combining CPU and GPU capabilities. Therefore, “no graphics card is visible” does not mean that a computer has no graphics-processing capability.
CUDA and ROCm belong to the software side of supported GPU computing. They are not additional physical components to find on a motherboard. A program needs a supported hardware/software combination; a GPU’s appearance does not establish that support.
Pause and explain
Why should an inventory record distinguish “GPU model” from “graphics card model”?
Compare your reasoning
One identifies the processor; the other identifies the complete assembly. Different cards can use the same GPU while differing in other hardware and requirements.
04
Keep hardware, firmware and architecture separate
BIOS is firmware associated with starting and configuring a computer; modern systems commonly use UEFI firmware. It is not the Windows desktop or an application. Observe only: changing firmware is a separate, authorized task.
Reduced Instruction Set Computing (RISC) describes an approach to processor instruction-set design. It is not a motherboard socket or a guarantee that one computer is faster than another. Identify the exact architecture, processor and workload rather than deciding from a label alone.
Pause and explain
You cannot read a RAM module’s label in a photo. What belongs in your inventory?
Compare your reasoning
Record what is actually visible and mark the unreadable specification unknown. Find a reliable model record or ask for a better approved photo; do not guess a part number from its shape.
Technical references for this section
05
RAM: capacity is not speed
RAM capacity, measured in GB, describes how much working data can fit. DDR generation, module form factor and the system’s supported capacity must match. Ordinary system memory uses DRAM, which needs refresh while powered; small processor caches commonly use SRAM. Both are volatile, unlike a saved file on a drive.
A DDR transfer rate is commonly stated in MT/s, not interchangeable with clock MHz. CAS latency (CL) counts clock cycles, so a lower CL number alone does not prove a shorter real delay when rates differ. Supported memory channels can increase the data path available to memory; two sticks do not guarantee the right channel arrangement. Follow the board’s slot-population instructions.
Task Manager → Performance → Memory can show installed capacity, current use and reported speed/slots. Sustained memory pressure can lead Windows to move data through its page file on storage. A freeze alone does not diagnose faulty RAM. Compare the workload, observed use and manufacturer specifications before recommending an upgrade.
Pause and explain
Does DDR5 with a lower CL number always beat every DDR4 configuration?
Compare your reasoning
No. Generation compatibility comes first. Compare rate, actual latency, capacity, channel setup and the workload rather than one label number.
Technical references for this section
06
Architecture, cores and clock
x86 originally describes a 32-bit instruction architecture; x64 extends that family to 64 bits. Arm64 is a different 64-bit architecture. Windows 11 requires supported 64-bit hardware. Matching bitness alone does not make an x64 driver work on Arm64 or make two CPUs fit the same socket.
Clock speed in GHz measures cycles per second, not useful work per cycle. Cores can work on multiple instruction streams, and cache keeps frequently needed data near the processor. Architecture, application support, cooling and power limits all affect performance. An APU includes CPU and graphics functions; it is not a separate third kind of compatible socket.
Task Manager’s CPU view helps observe workload, cores and current speed. Overclocking runs supported hardware beyond its normal settings and can raise heat, power demand and instability; it is not required or authorized in this lab. Compare suitable processors using the actual task and validated system support, not the highest GHz alone.
Pause and explain
Two processors both say 64-bit and 3 GHz. Are they interchangeable?
Compare your reasoning
No. Check architecture, socket, board firmware support and workload. Equal bitness and clock do not establish physical or software compatibility.
07
Specifications answer different questions
Discrete graphics memory (VRAM) holds graphics data. Capacity describes how much fits; bandwidth describes how quickly data can move. GPU cores, architecture, clocks, cooling and software support also matter. Integrated graphics commonly shares system RAM. CUDA and ROCm are platform-specific computing ecosystems, not proof that any program runs on any GPU.
HDDs store data on rotating magnetic media; SSDs use flash. An SSHD combines an HDD with a flash cache, not a full SSD of the same capacity. RPM describes rotation; MB/s describes transfer rate. SSD write-endurance ratings and sustained/random workload behavior are separate from an advertised peak speed. Capacity and an interface name do not prove health.
Disk Management shows disks, partitions and volumes; Task Manager can show observed disk or GPU activity. Neither view alone proves a drive is healthy or a component is the bottleneck. The motherboard must support the CPU socket, chipset/firmware, memory and expansion connections. Also check case form factor, connectors and power supply capacity before buying a part.
Pause and explain
Does an 8 GB graphics card have to outperform every 4 GB card?
Compare your reasoning
No. VRAM capacity is only one constraint. The GPU, bandwidth, workload and system compatibility also matter; compare evidence for the intended application.
The 15 lesson terms
The same terms and definitions used in the site’s study tools. Supporting labels in the reading are not extra required vocabulary.
- Motherboard
The main circuit board that connects the processor, memory, storage connections, and other hardware so they can work together.
Factual sources
- Central Processing Unit (CPU)
The general-purpose processor that carries out program instructions, including calculations and decisions.
Factual sources
- Accelerated Processing Unit (APU)
AMD’s term for a processor combining CPU and GPU capabilities in one device. Integrated graphics can provide video without a separate graphics card.
Factual sources
- RAM
Fast, temporary working memory for programs and data currently in use. Ordinary system RAM needs power to retain its contents.
Factual sources
- Storage Device
Hardware that keeps saved data for later use, such as an HDD or SSD. Saved files can remain after the computer is turned off.
Factual sources
- Hard Disk Drive (HDD)
A storage drive that records data magnetically on rotating platters and uses moving read/write heads.
Factual sources
- Solid-State Drive (SSD)
A drive that stores data in flash memory with no spinning platters. It retains saved files without continuous power.
Factual sources
- Graphics Processing Unit (GPU)
A processor suited to many calculations in parallel, including graphics work. It may be integrated into another chip or located on a graphics card.
Factual sources
- Graphics Card
An expansion board containing a GPU and supporting hardware, such as cooling and usually dedicated memory. The board is more than the GPU chip alone.
Factual sources
- Heatsink
A heat-conducting component, often with metal fins, that moves heat away from a chip and exposes more surface area for cooling.
Factual sources
- PCIe Slot
A motherboard expansion connector used by compatible PCI Express cards, such as many graphics and network cards.
Factual sources
- BIOS
Basic Input/Output System: startup firmware that initializes hardware and begins the boot process. Modern PCs generally use UEFI, though setup is often still called BIOS.
Factual sources
- Reduced Instruction Set Computing (RISC)
A processor-design approach built around comparatively simple instructions. Arm is an example; the label alone does not determine a computer’s speed.
Factual sources
- CUDA
NVIDIA’s parallel-computing platform and programming model for running supported workloads on NVIDIA GPUs, beyond drawing graphics.
Factual sources
- ROCm
AMD’s software platform for GPU computing, used with supported hardware for tasks such as scientific computation and machine learning.
Factual sources




































