BIOS and UEFI: What They Are and How to Enter Them

Two pieces of software sit between the power button and your desktop, and you will never notice either one. Until the morning the machine refuses to start. They do not appear in the spec sheet the shop likes to brag about, no review mentions them, and yet they are the first code that runs when you press that button. When they work, nobody cares. When they fail, you remember them for years.

Here is my starting point. If you plan to open your machine and touch the hardware, sooner or later you are going to need to get into that screen. Knowing what each one does, and which key to press, saves you a pile of panic the first time the screen goes black and the board starts beeping.

I am not going to stop at the dictionary definition. You are going to see what happens from the moment you hit the button to the moment the desktop appears, where BIOS and UEFI really part ways, which settings are worth changing, and which ones are best left alone. The line between a clean boot and a lost Sunday usually hides on a screen almost nobody looks at.

What is the BIOS?

BIOS stands for Basic Input/Output System. It is firmware, software burned into a chip on the motherboard, and it runs before any operating system. Its job is twofold:

• Start and check the hardware: at power-on it verifies that memory, disks and peripherals answer. That is the famous POST.
• Act as a bridge between that hardware and the operating system, so the boot can continue.

The BIOS leans on a chip called CMOS, a small memory that stores the machine’s configuration (time, boot order and so on) and stays alive thanks to the motherboard battery. When that battery dies, you get the classic complaint: the clock resets every time I turn the machine off.

And yes, the BIOS talks. Its beeps, different for each error, are how it tells you something is wrong when there is still no screen to explain it.

You will meet all of this in a configuration screen the manuals call Setup. It is the same thing people mean when they say they are going into the BIOS. Nothing there runs after Windows starts, and nothing you change there is saved to the disk. It lives in the CMOS chip, and the coin cell on the board is what keeps it remembered between bootups.

Now for the problem: age. The classic BIOS carries the design of the first IBM PCs and runs in 16-bit real mode. That one detail explains almost every limitation. It can only address 1 MB of memory, it reads from disk through ancient interrupts such as int 13h, and it expects to find the bootloader in the first 512 bytes of the drive, the MBR (Master Boot Record).

Those 512 bytes are all the room it has to know what to do next. The MBR only understands four primary partitions and drives up to 2 TB. The BIOS was designed for drives that would look like toys to us today, and that explains the ceiling.

What is UEFI?

UEFI (Unified Extensible Firmware Interface) is the evolution of the BIOS. It does the same job, booting the machine, with real upgrades: faster startup, more security and compatibility with modern hardware. Intel developed it in the late 1990s and today the UEFI Forum maintains it, a consortium that includes Intel, AMD, Microsoft, Apple and the big manufacturers. It is an open standard shaped by all of them.

The real change is technical. UEFI runs in 32 or 64 bits, has no 1 MB ceiling, understands GPT partition tables and boots through .efi files it keeps in a dedicated partition, the ESP (EFI System Partition), formatted as FAT32. That lets it load disk, network or graphics drivers before the operating system, boot natively from NVMe drives and hold several boot entries in NVRAM instead of depending on one magic sector.

In practice UEFI is more than a modern BIOS. It is a small operating system living on the board, with its own boot manager, its own applications and its own environment. That is why it can show a graphical interface with mouse support, reach the network to boot a remote image and enforce security rules the old BIOS never had.

BIOS vs UEFI: how they differ

The essential split is this. The BIOS is old firmware that boots in 16 bits and leans on MBR. UEFI is a modern 32 or 64-bit standard that uses GPT and adds security. Everything else, speed, big drives, the interface, hangs off that trunk.

	BIOS (classic)	UEFI
Startup	Slower, step by step	Faster, initialises hardware in parallel
Execution mode	16-bit (real mode)	32 or 64-bit
Partition table	MBR, 4 primaries	GPT, up to 128 partitions
Large disks (over 2 TB)	Cannot boot from them	Yes, no problem
Interface	Keyboard only, dated look	Graphical, mouse support
Security	Basic	With Secure Boot
Boot management	One sector on the disk	Entries in NVRAM and .efi files
Network boot	Limited	Its own network stack

The 2 TB limit is the one you feel in practice. If you have a bigger drive and the system will not boot from it, it is almost always because you are in legacy BIOS mode. And UEFI’s Secure Boot allows only signed software to boot, which shuts the door on certain malware that would otherwise install itself before the operating system.

There is a common trap when you compare machines. The same PC can boot in UEFI mode or in legacy (BIOS) mode depending on how the firmware is configured. Many boards offer both at once, the so-called CSM (Compatibility Support Module), so older systems are not left behind. The catch: if you install Windows in one mode, you have to boot it in the same one. Mixing them is one of the most common causes of “no bootable device”.

How a computer boots, step by step

Answering this question is the best way to understand what the firmware is for. From the button to the desktop there is a very short relay race, and the firmware runs the first leg.

1. The power supply steadies its rails. The board gets power and releases the processor’s reset line.
2. The CPU wakes at a fixed address. It does not run your operating system. It jumps to a memory position that points at the firmware. That is the first code to execute.
3. The firmware starts and runs POST. It checks processor, memory, keyboard and basic devices. If something fails, it warns with beeps or a message.
4. It initialises the hardware. It configures memory, disks, video and peripherals from the stored settings. UEFI does much of this in parallel and loads its own drivers. The BIOS does it one step at a time.
5. It looks for something to boot from. It walks the boot order (disk, USB, network…) until it finds a valid device.
6. It hands over control. The BIOS passes the MBR. UEFI locates the bootloader’s .efi file. From there the operating system takes charge.

UEFI adds a detail the BIOS never had. It does not scan blindly for a magic sector. It reads an ordered list of boot entries from NVRAM and tries them one by one. If none of them works, it falls back to the default path on the ESP, the file EFIBOOTBOOTX64.EFI on 64-bit systems. That is why a cloned disk sometimes boots fine on one board and refuses on another. The entry points at a drive that is not the one you moved.

Almost every key and menu you will meet later comes out of that route. The Fast Boot many machines ship with, for example, skips checks and shortens POST to start sooner. The price is that it detects newly connected devices worse, like that installation USB that “does not show up”.

Secure Boot: what it is and why it sometimes gets in the way

Secure Boot is a UEFI feature that verifies the digital signature of every component before running it. If the signature does not check out, the machine refuses to start that piece. It is a chain of trust that runs from the firmware to the bootloader and from there to the kernel. Each link validates the next.

The firmware keeps a set of keys to decide what is signed and what is not. With Secure Boot on, a bootkit, malware that tries to load before the operating system, has a much harder time. It cannot slip in without a signature the firmware recognises. That is why this mechanism has been required on machines certified for Windows since 2012, and also why many Linux distributions sign their bootloader with the so-called shim.

Linux users meet the practical side of this early. Most distributions ship a small signed bootloader, the shim, that hands control to the real one. When a driver or a kernel module is not signed with a key the firmware trusts, you can enrol your own with a tool such as mokutil. That is a controlled exception. Switching Secure Boot off for good is a separate decision.

The other side: if you want to boot an “unofficial” system, a modified build or an old recovery USB, you may have to turn it off. I would not leave it off without a reason. If a boot only fails while Secure Boot is on, the sensible move is to look for a signed image or update the bootloader before you leave the door open. It is the same logic we use when we talk about how a computer virus gets in. The earlier you cut the chain, the better.

How to enter the BIOS or UEFI

You go in right at power-on, before the system starts. The key depends on the manufacturer and, almost always, the boot screen itself tells you:

• Del (Delete): the most common on desktop motherboards.
• F2: typical on laptops and many desktops.
• F10: common on HP machines.
• Esc or F1: on some Lenovo, Dell and older laptops.

A trick that works: if you miss the moment, restart and start tapping the key before the logo appears, a couple of times per second. On modern Windows you have a shortcut that never fails: Settings, System, Recovery, Advanced startup, Restart now, then from the blue menu, “Troubleshoot, Advanced options, UEFI Firmware Settings”.

One warning that saves time. On Windows, shutting down does not always power the machine off fully, and fast startup can keep the previous session warm. If the keys do not respond after a shutdown, restart instead of powering off.

On Linux you do not have to guess. systemctl reboot --firmware-setup reboots straight into the firmware. On an Intel Mac, hold Option at boot to pick a drive. Macs with Apple Silicon have no BIOS or UEFI as such, they use their own boot manager, so that screen does not exist there the way we know it.

And if all you want is to boot once from a USB, do not fight the whole menu. Almost every manufacturer keeps a key for the boot menu, F12, F11 or F8 depending on the machine. If you are going to build the installation media, it pays to know your USB connector types and their speeds.

What you can change in the BIOS (and what to leave alone)

The firmware is not just a place where you fix the boot. It is where the hardware’s behaviour is decided. These are the settings that get used day to day.

• Boot order: the first thing you will touch. It forces the USB, the system disk or the network, and it also tells you why something you did not want is booting.
• Virtualisation (Intel VT-x, AMD-V): essential if you use virtual machines, Docker or Windows’ Linux subsystem. If it is off, you will see errors even when the hardware supports it.
• Memory profiles (XMP/EXPO): they enable the speed your RAM was designed for. If it runs below the advertised figure, this is usually why.
• Secure Boot: on or off depending on what you need to boot. Remember the price of leaving it off.
• Fast Boot: starts sooner, but detects hardware changes worse.
• Fan curves: useful if the machine sounds like a hairdryer. Do not invent values without understanding the temperature.

And here is the list of things I would not touch without knowing exactly what I am doing: processor voltages, power limits and anything with “override” in the name. A bad value rarely kills the hardware on the spot. The real risk is a machine that turns unstable in ways that are hard to diagnose. If you get lost, there is almost always a way back. Load Optimized Defaults restores the factory values, and clearing the CMOS, by pulling the battery or using the board jumper, wipes the configuration and lets you start from scratch.

Should you update the BIOS?

Short version, the same thing I would tell a friend: if everything works, leave it alone. Updating the BIOS can fix specific problems, such as faults with new hardware, security microcode or incompatibilities. A half-finished or interrupted update, though, can brick the board. If you do it, keep the machine plugged in, avoid interruptions and follow the manufacturer’s instructions to the letter. Many boards let you run it from the UEFI itself, with no operating system involved.

When would I do it? When the board does not recognise a new processor it should support, when the manufacturer publishes microcode that fixes a serious vulnerability, or when a documented incompatibility is ruining your day, say memory that will not run at its rated speed. The manufacturer usually publishes the changelog for each version. Read it before you flash, not after.

Some boards make the job easier with a second BIOS chip. If the first one fails mid-flash, the backup takes over and lets you recover without a service centre. It is not the norm on budget models, so check your manual before you rely on it.

The modern procedure is safer than its reputation suggests. Many boards ship utilities such as EZ Flash or USB BIOS Flashback that flash from a USB stick, and some models can recover from a failed update without ever booting. Even so, the golden rule does not change: never cut the power or switch the machine off halfway. If the process stalls, wait a few minutes before writing the board off.

How to tell if your machine uses BIOS or UEFI

The fastest way on Windows is to open System Information. Press Win + R, type msinfo32 and look at “BIOS Mode”. It will say UEFI or “Legacy” (the classic BIOS). If you prefer the command line, Confirm-SecureBootUEFI in PowerShell answers directly whether the machine booted in UEFI with Secure Boot available.

On Linux, just check whether the firmware directory exists. If /sys/firmware/efi is there, you booted in UEFI. If not, you are in BIOS mode. To find out how the disk is partitioned, diskpart on Windows or lsblk on Linux tell you whether the table is GPT (UEFI) or MBR (BIOS). That pair, firmware mode and partition type, has to point the same way or the boot fails.

Common boot problems (and how to get out of them)

These are the cases you will actually run into. Almost none is fixed by reinstalling, so it is worth looking before you format.

• “No bootable device”: check the boot order and, above all, that the firmware mode (UEFI or legacy) matches how the system was installed. Mixing GPT with BIOS mode is the classic.
• I installed a 4 TB drive and it will not boot: in BIOS mode there is no way to start from it. Convert to GPT or install in UEFI.
• The machine beeps and shows no image: badly seated memory, a loose graphics card or a peripheral blocking POST. The beeps are a code. The board manual translates them.
• The installation USB does not show up: Fast Boot is on, Secure Boot is blocking it, or the media was created badly. Try turning off fast startup and checking how it was written.
• The clock resets every time I power off: the CMOS battery is dead. Replace it and you are done.
• I changed the motherboard and Windows will not boot: Secure Boot, TPM or a different storage mode can change the boot signature. Matching the firmware to the old machine usually brings it back to life.
• An update stopped halfway: if the board has a USB recovery feature, use it. If not, it is a service job. That is why we insist so much on not cutting the power.

Frequently asked questions

Are BIOS and UEFI the same thing?

Not exactly. UEFI is the evolution of the BIOS. It does the same job, booting the machine, with more speed, more security and support for big drives. Out of habit we still say “entering the BIOS” for entering either one.

What is POST?

Power-On Self-Test. It is the check the firmware runs at power-up to verify that the hardware answers. If it fails, the BIOS warns with beeps or a message.

What is Secure Boot?

A UEFI feature that only allows signed software to boot, so nothing malicious loads before the operating system. The downside: to boot an “unofficial” system you will have to turn it off.

How do I know if my machine uses BIOS or UEFI?

On Windows, open System Information (msinfo32) and look at “BIOS Mode”. It will say UEFI or “Legacy” (the classic BIOS).

Does updating the BIOS erase my files?

No. The update rewrites the board’s firmware and leaves the disk alone, so your data and your system stay where they are. What can happen is that firmware settings reset, such as the boot order or the memory profiles, and you will have to review them afterwards.

What is the difference between MBR and GPT?

Two ways of organising a disk’s partitions. MBR is the old one, the one the BIOS uses: four primary partitions and a 2 TB limit. GPT is the modern one, the one UEFI expects: up to 128 partitions and no size ceiling.

The bottom line

BIOS and UEFI are the starting gun of any computer. They are the first thing to run and the last thing you remember. The BIOS steered the ship for more than thirty years. UEFI has taken over with more capability and, above all, with the security the old one never had. Understanding that handover, where each one comes from, what holds the old one back and what the new one wins, is what turns a black screen into a problem with a solution.

You do not need to master it to use your computer. You only need to remember where it is the day the machine will not boot, or the day you want that new drive to be first in the list. And if the problem is that it runs slow once the system is up, start by understanding what role RAM plays. If it is the full handover from firmware to operating system that interests you, you have the history of operating systems.