https://bugs.documentfoundation.org/show_bug.cgi?id=172456

V Stuart Foote <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEEDINFO                    |NEW
           See Also|                            |https://bugs.documentfounda
                   |                            |tion.org/show_bug.cgi?id=16
                   |                            |4788
                 CC|                            |[email protected],
                   |                            |[email protected]

--- Comment #9 from V Stuart Foote <[email protected]> ---
(In reply to roadsideyard.ohi from comment #8)
> <item oor:path="/org.openoffice.Office.Common/VCL"><prop
> oor:name="ForceSkiaRaster" oor:op="fuse"><value>false</value></prop></item>
> 
> As you instructed earlier, I replaced the line with "false" with "true" and
> it started up without any problems.
> Thank you for your help.

Great! Thanks for sticking it out.

But, there could still have an issue with nVidia Optimus and assignment of
LibreOffice to one or the other GPUs.

I  seems the nVidia 582.28.0.0 (2441543680) driver on device 0x1c8c is fairly
recent, but it has known issues affecting Vulkan on Windows 10/11--mitigated by
updates that are needed to Windows 11.  This note on reddit for example: 
https://www.reddit.com/r/nvidia/comments/1qqgx9q/58228_security_update_driver_for_maxwell_pascal/
 So your laptop may need Windows 11 updates.

Latest would be the 582.66 driver build found here:
https://www.nvidia.com/en-us/drivers/details/272769/  you could check if that
is functional for Vulkan acceleration.

Also, that listed Intel driver 0.16.2 (65538) for the 0x3e9b Intel(R) UHD
Graphics 630 (8th/9th gen) is hopelessly out of date. You should update that
graphics driver, but even then probably won't provide functional Vulkan
acceleration, we had previously black listed it for the OpenGL era. None the
less, I would update the driver directly from Intel here, and not depend on
Windows for update:

https://www.intel.com/content/www/us/en/download/776137/intel-7th-10th-gen-processor-graphics-windows.html

@Julien, I thought we were defaulting ForceSkiaRaster to "true" after a failed
graphics test? As done for bug 164788, with Mike's
https://gerrit.libreoffice.org/c/core/+/181290 other than a simple denylist for
the nVidia 0x1c8c, and Intel 0x3e9b is something more needed?

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to