The times when you had to fight for the basic settings of a computer by text-based BIOS screens are with a modern UEFI firmware happily over. With graphical user interface, mouse operation, and sometimes even additional tools, the setup screens are more reminiscent of Windows programs. The features include everything you can find in a traditional BIOS setup: most mainboards also have extensive overclocking options.
Modern Processor Architecture
Unlike a traditional BIOS, UEFI supports 32-bit and 64-bit modes of modern x86 processors, as well as other processor architectures, such as ARM (32-bit). Thus, a UEFI firmware can also address more than one MByte of the main memory, which allows for more comfortable and comprehensive functions as well as setup screens.
UEFI comes with a new graphical user interface
By using a device driver, a UEFI firmware can access components such as the graphics card and the mouse or a touchpad or even a touchscreen, which allows comfortable operation with a graphical user interface with high display resolution.
HDDs and SSDs on the road
Instead of VGA, a newly developed Graphical Output Protocol (GOP) is used for the graphical display of the setup screens, which supports higher resolutions, color depths and, above all, multiple graphics adapters in one computer, but does not require special drivers for different graphics chips. Another benefit of UEFI is GPT support. With GUID PartitionTable, hard disks with full capacity can be used as boot devices larger than 2.2 TB.
Turbocharger for the CPU
This is not possible with the previous master boot record (MBR). The starting point and the size of a partition are described by 32-bit values in MBR, GPT uses 64-bit values, theoretically a GPT disk could be up to 9.4 Zettabyte. The limitation of MBR to four primary partitions per disk is also omitted for GPT partitioning. In principle, many partitions can be created without restrictions.
UEFI is currently supported on Linux and 64-bit versions of Windows Vista with SP1 and Windows 7. However, practically all UEFI implementations contain a Compatibility Support Module (CSM), which emulates the previous BIOS to allow the installation of older or alternative operating systems.
Most manufacturers use the capabilities of the graphical user interface to split into a setup screen for all the necessary settings and information as well as an expert mode for more advanced options such as clock and voltage settings for overclocking
This is how current Asus motherboards start with the EZ Mode, which provides at a glance all important information such as the clock rate of the CPU, processor and motherboard temperature, voltages and fan speeds. You can also drag-and-drop the boot drive and choose between three performance settings for the system.
MSI also provides a start screen for the most important settings and information on the ClickBIOS II, which branches to different subpages. A special feature is that an Internet access with an integrated browser or the loading of a current BIOS update is possible from the UEFI screen. However, a hard drive and a functioning Windows installation are necessary, since a Winki software must be installed for the Internet access.
Gigabyte's 3D BIOS displays an interactive 3D representation of the mainboard with click-points, which allow you to enter sub-menus. A mouse click on the CPU calls the clock settings, one to the SATA sockets for the SATA controllers. This looks nice, but it does not offer any advantages in practice, since a central overview is missing.
The graphical user interface of a modern UEFI firmware also facilitates the initial setup of a mainboard. For example, the boot drive can often be simply drag-and-drop. For the 64-bit version of Windows you should use the UEFI option during installation. Afterwards, you should take a look at the settings for the SATA controller (s). It should be for a hard drive or SSD in any case to AHCI instead of IDE mode.
The abbreviation AHCI stands for Advanced Host Controller Interface and is a standard introduced by Intel for the communication between SATA controllers and drives. Only the AHCI option makes SATA 3GBit features of modern hard disks such as Hot-Swap or Native Command Queuing (NCQ) usable for the host controller. NCQ increases the transfer rate on traditional hard disks by up to 10 percent.
Because the disk controller 32 places instructions into a buffer and optimizes their sequence, the heads need to travel less, but at the cost of a slightly increased access time. For SSDs, AHCI should also be enabled, as most models benefit from NCQ. Here, buffering the commands is primarily to keep the SSD busy while the CPU is dealing with other tasks.
Feintuning
For Windows 7 systems, AHCI is also a prerequisite for the operating system to use the TRIM command. When using Windows 7, Vista or a Linux distribution with kernel 2.6.19 or later, the operating system provides the necessary drivers. An AHCI driver must be installed for Windows XP.
In the case of several identical hard disks, it is also possible to run these in a RAID network on most of the mainboards in order to increase the performance or ensure a higher data security. To do so, the RAID option must be selected for the SATA controller settings in the UEFI setup. The RAIDS is not set up in the UEFI setup, but either in the RAID controller setup screen or in the driver dialog in Windows.
Install Windows 7 in UEFI mode
In Windows 7, additional drivers must also be installed. Native Command Queuing also works on disks in the RAID array, TRIM is currently not supported by any driver, which should be the case with Intel chipsets with the next version 11.5 of the Windows driver for Rapid Storage Technology (RST) Code>
With the overclocking functions UEFI differs in nothing from a traditional BIOS, the difference is only in the operation by mouse click. Even if the overclocking is easy, you drive the CPU outside its specifications and therefore at your own risk. In principle, the overclocking options for the majority of the mainboards and for both Intel and AMD processors consist of three function blocks: clock rates and multipliers of the processor, memory and, as a third, the power settings of the mainboard.
Many of the current processors offer a whole range of additional overclocking options compared to previous models, resulting from the variable clock frequencies that the Intel Core processors or AMD Llano and Bulldozer work with. They lower their clock when the CPU has little to do to throttle the power consumption. An Intel Core i7-3770K with a clock frequency of 3.5 GHz runs in idle mode with only 1.6 GHz.
Installation using a USB stick
Unlocked
The clock rate of a processor is determined by two factors: the fixed reference clock and the variable clock multiplier. For older processors with memory controllers in the Northbridge, the reference clock was still referred to as FSB (Front Side Bus). For the mainboards for current Intel processors with sockets 1155 or 2011, he calls CPU base clock and can also be found under the acronym BLCK in the UEFI setup.
It also depends on the frequencies of the memory, the PCI Express ports as well as the SATA and USB controllers. The normal BLCK frequency is 100 MHz, but an increase is only possible by a few percent, without the computer running unstably. Therefore, overclocking on Intel computers with Sandy Bridge or Ivy Bridge processors is practically only via the clock multiplier and only for processors whose serial number ends with a "K" and with the Extreme Edition CPUs >
In the CPU or overclocking screen of a UEFI firmware, depending on the manufacturer and mainboard, various overclocking possibilities are available. Here you should take a look at the mainboard manual.
Thus, the multiplier for the normal operation of the CPU can be increased first. This increases performance and power consumption when the CPU is not idle. Alternatively, you can intervene in the settings for the Turbo Boost mode and increase the multipliers that depend on the number of active processor cores.
For AMD processors, the clock multiplier is free for most models, while overclocking is recommended for a similar procedure as for Intel CPUs. For AMD systems the base clock is at 200 MHz, the automatic overclocking function is called Turbo Boost. In the case of the FX processors, ie bulldozer CPUs, the possibility of overclocking the Northbridge integrated in the CPU by a few percent has been overclocked. This increases the clock rate of the Level 3 cache.
If you overclock a processor, then the cores voltage for the CPU should be increased. Here, a plus of 0.1 volts is enough, more has a detrimental effect on the service life of the processor. In addition, the load-line calibration should be set to an at least average value. This minimizes the reduction of the core voltage with a high processor load.
This results in a higher power consumption and processor temperature, but increases the stability of an overclocked CPU. Mainboards with digital VRMs provide several profiles, one should set a performance profile for overclocking, which provides more power, but allows a higher temperature of the voltage regulators.
For Intel mainboards you should use XMP memory. The abbreviation stands for "Extreme Memory Profile", the EEPROMs of the modules contain profiles prepared here, which facilitate the overclocking. AMD also has AMD smart profiles, which, in addition to the FX, also support the A series processors. Of course, you can also set the memory multiplier and the timings manually in the UEFI setup.
The 64-bit versions of Windows 7 already support UEFI. A UEFI installation is useful when you want to install Windows on a hard disk that is larger than 2.2 Terabytes, or if you plan to upgrade your operating system to Windows 8.
The files required for a UEFI installation are already present on the installation DVD. For mainboads with UEFI firmware the optical drive appears two times under the offered boot drives, once with the addition "UEFI". This should be assigned the boot disk priority in the UEFI setup screen.
The Windows installation using a prepared USB stick is somewhat more complex as the installation medium has to be adapted. A USB flash drive belongs to a device class other than an optical drive, so a different EFI startup file is required. On the Windows 7 DVD or ISO file, the BOOTX64.EFI file is the same as CDBOOT.EFI and must be replaced.
To do this, navigate to the% SYSTEMROOT \ BOOT \ EFI directory on a 64-bit Windows 7 computer and copy the BOOTMGFW.EFI file to the \ EFI \ BOOT directory on the stick. This file is the appropriate start file for USB mass storage devices. Then rename this file to BOOTX64.EFI. After that, the fast installation via USB stick is no longer an obstacle.
Important: The hard drive is partitioned as a GPT disk when setting up Windows 7. At least three partitions are created: an EFI system partition, a Microsoft Reserved partition, and a Windows system partition. 32-bit versions of Windows can not access GPT disks.
A memory module is constructed logically as a grid. It consists of row and column addresses similar to the fields on a chessboard. When the processor accesses a memory cell via the memory controller, the row and then the column are determined. The memory cell is defined by the intersection point. Row address strobe and column address strobe are differentiated.
When accessed, the memory controller sends the line to be addressed to the memory logic, which completely reads out its contents and stores it in a temporary memory. The time between the reading of the line and the storing of the entries is called RAS-to-CAS latency, in short tRCD.
The controller then sends the CAS (Column Address Strobe) signal to the memory. This selects the element of the series defined in this way and writes its contents into the output register. The time between these operations is the CAS latency (CAS).
This allows the cell to be accessed and the memory reads the next cell. Ideally, it is in the same line as the cell just read. This hit, the so-called page hit, causes the content of the cell to be stored without delay into the output register. For continuous page hits, the memory can reach the optimal data throughput.
If this does not work (Page Miss), the entire line has to be read again. Previously, however, the store must restore and deactivate the row. The time for this is called RAS Active Time (t_RAS). The data lines are then charged with a reference voltage (RAS Precharge-Time -t_RP), the process begins anew.
No comments:
Post a Comment