Detailed description of the boot process in UEFI systems

Last update: 17/12/2025
Author Isaac
  • UEFI largely replaces BIOS Legacy and the MBR, contributing a Boot Faster, more secure and flexible thanks to the use of GPT and EFI system partitioning.
  • The UEFI boot process initializes hardware, locates the ESP, executes an EFI boot manager and loads the operating system kernel with support for multi-boot and Secure Boot.
  • GPT overcomes the limitations of MBR by allowing more partitions, very large disks, and greater reliability through redundant copies and CRC checksums.
  • Legacy BIOS mode is still necessary for older systems or certain network environments, but whenever possible, it is recommended to use native UEFI with GPT disks.

UEFI boot process diagram

If you enjoy tinkering with your computer and understanding what happens from the moment you press the power button until you see the desktop , sooner or later you'll come across two key terms: BIOS and UEFI . They sound the same, but behind them lies a paradigm shift in how your PC boots up and is protected.

In recent decades, we've gone from very limited firmware to a much more complete environment, capable of managing enormous disks, booting faster, loading multiple operating systems , and implementing advanced security mechanisms like Secure Boot . All of this revolves around the UEFI standard and how it has largely replaced the old BIOS Legacy and the classic MBR partitioning scheme.

BIOS, MBR, UEFI and GPT: bringing order to the concepts

BIOS and UEFI Comparison

Before going into detail about the UEFI boot flow, it is worth clarifying what each part is, because BIOS, UEFI, MBR and GPT are often mixed up and used almost interchangeably, when in reality they are different things that work together during boot.

The Legacy BIOS (Basic Input/Output System) is the classic firmware used by x86 motherboards since the 80s. It resides on a ROM or flash chip on the motherboard and is the first thing the CPU executes upon startup, long before the operating system loads. Its purpose is to initialize the basic hardware and hand control over to a small boot code located on the disk, traditionally in the Master Boot Record (MBR).

The MBR is the first physical sector of a storage drive (the familiar 512 bytes). Within this sector are two critical elements: a small piece of boot code (the first boot manager) and the partition table in MBR format. This code is what the BIOS loads into memory and executes, and from there, second-stage boot managers such as BOOTMGR, GRUB, LILO, or Syslinux are chained together.

The BIOS + MBR combination was sufficient for many years, but it came with significant limitations: a maximum of four primary partitions, a practical limit of about 2,2 TB per disk, and the absence of modern security and integrity verification mechanisms. As disks grew larger and boot management became more complex, these limitations became a problem for both home users and professional environments.

To overcome these barriers, UEFI and GPT emerged . UEFI (Unified Extensible Firmware Interface) is not simply a "new BIOS," but a modern firmware specification, designed by a consortium of manufacturers ( Intel , AMD, Microsoft, HP, and others) to provide a more flexible, extensible, and secure environment. GPT (GUID Partition Table), for its part, is a partitioning system that accompanies UEFI and leaves the limitations of MBR far behind.

What is UEFI and how does it relate to firmware?

UEFI is a specification that describes the interface between the motherboard firmware and the operating system . It doesn't dictate how all the firmware should be implemented internally, but it does define which services it should expose and how they communicate with the operating system and EFI applications.

  AMD Ryzen Z2 AI Extreme: The New AI Laptop Deal

In practice, when people talk about "entering the BIOS" on a modern PC, they're almost always entering a UEFI environment, even though manufacturers still call it BIOS out of habit and to avoid confusing the user. This UEFI environment is stored on a chip on the motherboard and runs from non-volatile memory, usually backed up by a coin cell battery and NVRAM, which retains the configuration even when the computer is turned off.

UEFI acts as a mini operating system layer: it provides boot services, runtime services, its own drivers for certain devices, and the ability to read some file systems, especially FAT in the EFI System Partition (ESP) . Thanks to this, it no longer relies exclusively on the MBR nor does it require the boot code to reside in a 512-byte sector.

It's important to clarify that UEFI doesn't completely "eliminate" the BIOS in the traditional sense: the specification focuses on the interface, not the entire firmware. However, in practice, modern computers are sold with UEFI firmware that also implements compatibility modes to continue supporting legacy booting when needed.

That's why we often talk about UEFI-BIOS or UEFI BIOS : a firmware that exposes the modern UEFI interface, but maintains options to behave like a Legacy BIOS through the famous CSM (Compatibility Support Module).

GPT: the partitioning scheme that accompanies UEFI

The other major component is GPT (GUID Partition Table) . While MBR only allows four primary partitions and has a limit of about 2,2 TB, GPT uses globally unique identifiers (GUIDs) for each partition and stores partitioning information with 64-bit LBA addresses.

With GPT, you can define many more partitions (usually up to 128 by default, expandable depending on the tool), without needing to create extended or logical partitions. Furthermore, the maximum addressable disk size theoretically increases to zettabyte levels, far exceeding current needs, but ensuring that we won't run out of space in the short or medium term.

Another crucial difference is that GPT stores redundant copies of its header and partition table at the beginning and end of the disk, along with CRC checksums. This allows for the detection of corruption and, in many cases, recovery. In MBR, if that 512-byte sector becomes corrupted, you can easily lose access to the entire disk.

In summary, UEFI and GPT form a tandem that provides greater capacity, more partitions, redundancy, and integrity verification , something fundamental in modern disks, high-performance NVMe SSDs , and systems with multiple operating systems.

Description of the boot process in UEFI systems

Now that the concepts are clear, let's get to the heart of the matter: what exactly does a computer with UEFI do from the moment you turn it on until the login screen appears? The workflow is different from the classic BIOS + MBR model, although at first glance the user only perceives that it boots faster and with fewer strange issues.

1. Powering on and firmware initialization

When you press the power button, the CPU locates the firmware code at a fixed memory address and begins running the UEFI environment stored in NVRAM, or the motherboard's flash memory . Unlike legacy BIOS, UEFI can operate in either 32-bit or 64-bit mode from the start, allowing it to use more memory and work with more complex data structures.

  Sony and AMD collaborate on AI hardware development for the long-awaited PS6

In this phase, the POST (Power-On Self-Test) is performed. UEFI checks essential components such as the CPU, RAM, GPU, storage, USB ports , fans, and other critical devices. If something fails, the computer will typically emit error beeps or display codes on the screen before proceeding.

2. Loading UEFI Drivers and Services

Once the basic check is complete, the UEFI firmware loads its own drivers for certain devices (for example, some network or storage controllers) and enables the boot and runtime services . These services will be accessible to both the boot manager and the operating system kernel during the initial stages.

The configuration stored in NVRAM also applies here: boot order , security options such as Secure Boot, lists of authorized EFI applications, etc. Changes such as choosing whether to use UEFI mode or Legacy/CSM mode, or enabling or disabling Secure Boot, are saved precisely in this non-volatile memory.

3. Location of the EFI System Partition (ESP)

In a standard UEFI system, instead of going directly to the MBR, the firmware looks on the storage devices for a partition with a specific GUID, which it marks as the EFI System Partition (ESP) . This partition is usually formatted in FAT and contains EFI executables (.efi files) for boot managers and utilities.

Within the ESP, a default boot path is defined, typically /EFI/Boot/bootx64.efi on 64-bit systems. Additionally, NVRAM stores further boot entries: for example, one entry for Windows Boot Manager , another for GRUB on a Linux distribution , another for diagnostic tools, and so on. The internal UEFI boot manager can display these options in a menu or follow a pre-configured order.

4. EFI boot manager execution

Once the device and boot entry are selected (either by configuration or user choice), the firmware loads the corresponding EFI application into memory, for example the Windows boot manager or GRUB in UEFI mode , and transfers control to it.

The main difference compared to the MBR model is that we are no longer limited to 512 bytes of initial code or a single, rigid partition table. The EFI boot manager can directly read the GPT table, access different file systems on the ESP, and offer a visually appealing and configurable multi-boot menu , all while leveraging the services provided by UEFI.

5. Loading the operating system kernel

The boot manager is responsible for locating the operating system kernel on the disk (and, if applicable, the initrd or initramfs in Unix -like systems ), loading it into memory, and passing it boot parameters and hardware information. At this stage, it may still rely on UEFI services to read disks, consult ACPI tables, etc.

Once the kernel takes over, it begins more advanced device initialization, loading specific drivers, mounting user file systems, and starting system services. Finally, the graphical or console environment you use daily is launched, and the login screen or desktop is displayed.

BIOS Legacy vs UEFI: Functional and Usage Differences

It introduces very specific improvements that go beyond a simple name change: on the surface, both "do the same thing": initialize the hardware and boot the operating system. But UEFI introduces very specific improvements that significantly change the landscape.

On the one hand, UEFI allows for the management of disks larger than 2 TB without any workarounds , which is important for modern PCs, workstations, and servers. On the other hand, the interface is no longer a blue or black screen controlled solely by the keyboard, but rather a more user-friendly graphical environment, typically with mouse support and menu-based navigation.

  Non-root containers: A complete guide to managing permissions and security in restricted environments

Furthermore, UEFI incorporates mechanisms such as Secure Boot, which verifies the signature of boot components to prevent malware from sneaking in during this early stage before the antivirus or firewall starts. Traditional BIOS, on the other hand, was much more vulnerable in this respect, leaving the boot process highly exposed to UEFI bootkits and other low-level attacks.

Another key point is the flexibility in multi-booting. With BIOS + MBR, boot managers had to manage the MBR, the active partition, and various boot chains. With UEFI + GPT, each system can have its own .efi file on the ESP, and the firmware itself can handle multiple clean and well-organized boot entries.

Therefore, if your hardware allows it, the general recommendation is to always use native UEFI mode instead of legacy BIOS mode. You'll gain in capacity, performance, security, and ease of use when managing operating systems.

Specifically boot in UEFI or BIOS mode and control WinPE

If you use Windows PE or custom Windows installers, you can choose which mode your computer will boot in. When you access the firmware boot menu, you'll typically see duplicate entries for the same device: for example, "UEFI: USB Drive" and "BIOS: USB Drive ." Choosing one or the other determines the firmware mode in which WinPE or the installer will run.

Some devices only support one of the two modes. Others allow both, but if you want to force an image to boot only in a specific mode, you can experiment with removing certain files from the media: remove the bootmgr file from the root directory to force UEFI booting, or delete the efi folder to force it to boot only in BIOS.

Another technique used in corporate deployments is to prepare the disk drives beforehand with the appropriate format: GPT for UEFI and MBR for BIOS . If you try to install Windows on a GPT disk by booting in BIOS mode, the installation will fail, and this serves as a safeguard to prevent mixing modes.

In WinPE, in addition to the PEFirmwareType registry value, you can use the wpeutil UpdateBootInfo command and batch scripts that analyze the reg query output to make automated decisions, such as launching one partitioning flow or another depending on the current mode; you can also rely on tools to analyze Windows boot with BootTrace and debug specific steps.

All of this is especially relevant when you want to standardize the use of UEFI + GPT with Secure Boot across all your systems , which is currently the combination recommended by Microsoft and most manufacturers to take full advantage of hardware security and management capabilities.

Restore UEFI defaults
Related articles:
Restore UEFI Defaults: Step-by-Step Guide for All Brands
Isaac

Isaac

Passionate writer about the world of bytes and technology in general. I love sharing my knowledge through writing, and that's what I'll do on this blog, show you all the most interesting things about gadgets, software, hardware, tech trends, and more. My goal is to help you navigate the digital world in a simple and entertaining way.