Re: Does anyone have aarch64 main FreeBSD working under aarch64 WIndows 11 Pro's Hyper-V?

From: Mark Millard <marklmi_at_yahoo.com>
Date: Sat, 09 Nov 2024 18:04:51 UTC
Hello Robert,

Thanks for the reporting the possibility. It seems to make no
observable difference in behavior.

I use the Windows Dev Kit 2023 natively too (USB3 media).

The idea for Hyper-V use was to be able to to test with a 
controlled amount of RAM available for a controlled number of
"cpus" (both from FreeBSD's point of view). I've done such for
amd64 on a 7950X3D amd64 system, for example. I had no access
to amd64 systems with the combinations of RAM and FreeBSD CPU
counts and the 7950X3D took less time to run tests than such
small systems would.

I was able to show building, for example, lang/rust via
poudriere(-devel) was just a matter of appropriate
configuration for RAM spaces much smaller than some folks
were reporting as "impossible", such as with 8 GiBytes of RAM.
(Yea, part of that configuration is use of RAM+SWAP and part
of it is avoiding configuring competition for the RAM+SWAP
space. For example, I used UFS and had little or no tmpfs use
involved. I also limited how many cores would be in active
use for the smallest RAM sizes.)

The speed ratio would be less for the Windows Dev Kit 2023,
but the general types of advantages for such testing would
apply if Hyper-V use worked. Using <= 4 cores would take less
time than using total_mem= in config.txt on a RPi4B for when
actively using the same number of "cores" for such
testing.


Thanks,
Mark


On Nov 9, 2024, at 02:47, Robert Clausecker <fuz@fuz.su> wrote:

Hi Mark,

I also run a Windows Dev Kit, but natively.  Try

   hw.pac.enable="0"

in /boot/loader.conf.

Yours,
Robert Clausecker

Am Fri, Nov 08, 2024 at 11:49:21PM -0800 schrieb Mark Millard:
> The specific Win11 Pro context that I have access to involves a
> Windows Dev Kit 2023.
> 
> The Hyper-V boot output stops at the end of (no serial
> port to capture, so not literal text):
> 
> EFI framebuffer information:
> addr, size     0xe0000000, 0x800000 
> dimensions     1024 x 768
> stride         1024
> masks          0x00ff0000, 0x0000ff00, 0x000000ff, 0xff000000
> 
> 
> Trying boot loader selections for Video vs. Serial vs. both made
> no visible difference.
> 
> Hyper-V indicates 12% usage. (8 cores so this likely is how 1
> core being busy would display.)
> 
> 
> One form of experiment was via a download and expansion of:
> 
> http://ftp3.freebsd.org/pub/FreeBSD/snapshots/VM-IMAGES/15.0-CURRENT/aarch64/20241107/FreeBSD-15.0-CURRENT-arm64-aarch64-ufs-20241107-5036d9652a57-273504.vhd.xz
> 
> (Listed with "2024-Nov-07 06:22" in the web browser.)
> 
> ===
> Mark Millard
> marklmi at yahoo.com
> 
> 
> 

-- 
()  ascii ribbon campaign - for an encoding-agnostic world
/\  - against html email  - against proprietary attachments


===
Mark Millard
marklmi at yahoo.com