Биос видеокарты Intel

[Jmd]

Скачал обновленные драйвера для встроенного видеоадаптера Intel GM965 и обнаружил несколько зипчиков в главном архиве.

По названиям понятно что это биос для видеокарты.

Но ни слова описания: для чего и как (как - еще вроде понятно, но вот зачем?)

А вдруг станет хуже и не откатишь?

Файл имеет вид vbiosgm965.zip

с ним еще несколько подобных файликов для других чипов.

[#Nii]

Единственое что я на этот счет нашел

Where is documentation about the dependencies between the VideoBIOS (VBIOS) and a particular GMA driver revision for XP and Vista available? Or the much simpler question: What's the minimum VideoBIOS revision supported by any GMA driver?

I do not find this information in any distributed readme/changelog/install guide. I can only see that the zip-Versions of GMA drivers intended for developers contain VBIOS images for different chipsets (e.g. vbiosGM965.zip). However this dependency exists (see calling conventions related to the System Management Interrupt). Furthermore the driver itself does not reject to install if it finds a too old VBIOS revision.

In the long run this dependency should be removed altogether. It's true that the OS might be interested in some state information to initialize it's drivers appropriately (e.g. connected displays, users preferred display, available power saving modes etc.).

Even your own IEGD 8.0 drivers are much further developed in this respect. See 4.4.1 "Data Dependencies Between VBIOS and Intel Graphics Driver":

The Intel Embedded Graphics Drivers do not depend on any data from the VBIOS, and

will either use driver settings or select default values for the attached displays. This

allows the driver to properly operate with incompatible BIOS or BIOS replacements.

The Linux folks (in particular intel's OSS group) are also successfully moving in this direction (source), because they suffered from distributed and redundant responsibilities regarding graphics mode control.

There are muliple ways to resolve the VBIOS <-> GMA driver dependancy one of them being:

Retrieve some state information from VBIOS during OS driver initialization and from this point on only use OS driver code paths. If the latter one is not possible install own VBIOS by patching software interrupts and transfer state information to make sure VBIOS is always compatible and tested with current OS driver.

I'm pretty sure this would lower the number of bugs intel drivers have in respect to power management: Just look at your changelogs regarding issues related to ACPI mode switching (S0,S3,S4). Even the latest (v15.9) driver produces warnings in Vista's performance monitoring related to resume/suspend.

[Jmd]

спасибо.

теперь можно найти и это наше обсуждение. :)

я так понял и статьи - чем старее, тем лучше -меньше ошибок + все равно система им не пользуется...