Upgrade to Win 8 questionable

[Fokke Nauta]

My VM of choice, was VPC2007, and it's only installed in WinXP.
VPC2007 won't run Windows 8. And VPC2007 supports a single core
only.

I wasn't aware of VPC 2007 at all.
I used VMWare.

I don't like VirtualBox, because of the "rigid" rules
in the control panel (usage of GUIDs) - the rules just get
in the way of getting things done. And VMWare never seemed to
be a practical option (I think I installed it once for a trial,
but threw it away).

Funny. I liked VMWare from the beginning.

I'm just set in my ways I guess.

Windows 8 won't run VPC2007 or Windows Virtual PC or WinXP mode.
Windows 8 will run Hyper-V (but only if you have SLAT)

What's SLAT?

and
Windows 8 will run VirtualBox. What a mess. And much of the
"rules" are purely arbitrary and capricious.

Kinda like the Linux kernel bug for the last year, where
Linux improperly determined that VPC2007 was actually
Hyper-V, and then Linux assumed Hyper-V "drivers" were
available, when in fact they're not. (The Linux LiveCD
could then not find a hard drive to install on!)

That's pretty bad :-(

Finally,
that was fixed, and the first evidence of the fix is in
Ubuntu 13 beta. I can actually load Ubuntu 13 in VPC2007 - it's
dog slow on my 3GHz Core2!

As it can only use 1 core as you stated before. No wonder.

It even gets into a loop once
in a while, and pegs the CPU on one core. Cool!

Now, why would I want to scare someone, when I can instead
just convince them to use multi-booting for a few things

Well, as I am used to VMWare, it's way easier than setting up a
multiboot system. And not so risky.

Multi-booting isn't convenient, but it is "clean".

Is it cleaner than using virtual machines?

Paul

Fokke

 

[Paul]

What's SLAT?

SLAT is Second Level Address Translation. On Intel, this
is called Extended Page Tables. Don't ask me what it's for,
it sounded like "feature creep". I just wanted a
software product that would work!

http://en.wikipedia.org/wiki/Page_table#Nested_page_tables

"Nested page tables can be implemented to increase
the performance of hardware virtualization."

A good software, would have a fallback operation, for
systems with only regular page tables.

The latest "craze" in the VM world, is virtual machines
where the virtual RAM provided for them, changes dynamically.
Each VM is like an accordion. Again, I'm guessing this
is "progress". I just want the stuff to work, not spend
an eternity figuring out why it stopped working. I'm
perfectly happy giving a VM a fixed 1024MB allocation.

Is it cleaner than using virtual machines?

The idea is, people understand how to get an OS booted.
Once it boots, it's working. There are fewer technical
hurtles. Imagine if I had to explain all the quirks
of virtual machines, the things they do and don't emulate,
the boot time switches that might be needed to complete
booting. Or, adjustments needed to give acceptable
performance. That would take a few pages of text.
And people hate reading my posts

Paul

 

[charlie]

SLAT is Second Level Address Translation. On Intel, this
is called Extended Page Tables. Don't ask me what it's for,
it sounded like "feature creep". I just wanted a
software product that would work!

http://en.wikipedia.org/wiki/Page_table#Nested_page_tables

"Nested page tables can be implemented to increase
the performance of hardware virtualization."

A good software, would have a fallback operation, for
systems with only regular page tables.

The latest "craze" in the VM world, is virtual machines
where the virtual RAM provided for them, changes dynamically.
Each VM is like an accordion. Again, I'm guessing this
is "progress". I just want the stuff to work, not spend
an eternity figuring out why it stopped working. I'm
perfectly happy giving a VM a fixed 1024MB allocation.


The idea is, people understand how to get an OS booted.
Once it boots, it's working. There are fewer technical
hurtles. Imagine if I had to explain all the quirks
of virtual machines, the things they do and don't emulate,
the boot time switches that might be needed to complete
booting. Or, adjustments needed to give acceptable
performance. That would take a few pages of text.
And people hate reading my posts

Paul

When people start talking about virtual machines, I'm reminded of old HP
Minis and HP "RTE" Various versions of RTE used virtual memory, multi
tasking, multi users, and "rings" of control in the ops system.
Even the quasi "machine language" functions/operations could be
modified! The assemblers produced "machine code" instructions that were
actually locations/addresses into storage registers/ROM/RAM that
contained multi byte true machine language instructions which were then
executed.

We used them in the 70's and into the 90's to control complex electronic
test stations, develop software, and so forth.
There was even a fiber optic link between a building that housed the
"pure" software development systems and the building housing a duplicate
development system and the test stations. You could, once the device or
system was mounted on a test station, control the testing from the test
system operator console or any of the development systems terminals.
Even the terminals were "intelligent" in that they were Z80 based and
could execute programs beyond that contained in a ROM based terminal
operating system.

 

[Fokke Nauta]

SLAT is Second Level Address Translation. On Intel, this
is called Extended Page Tables. Don't ask me what it's for,
it sounded like "feature creep". I just wanted a
software product that would work!

http://en.wikipedia.org/wiki/Page_table#Nested_page_tables

"Nested page tables can be implemented to increase
the performance of hardware virtualization."

A good software, would have a fallback operation, for
systems with only regular page tables.

The latest "craze" in the VM world, is virtual machines
where the virtual RAM provided for them, changes dynamically.
Each VM is like an accordion. Again, I'm guessing this
is "progress". I just want the stuff to work, not spend
an eternity figuring out why it stopped working. I'm
perfectly happy giving a VM a fixed 1024MB allocation.

That's the way I always work. Or sometimes 2048 MB.

The idea is, people understand how to get an OS booted.
Once it boots, it's working. There are fewer technical
hurtles. Imagine if I had to explain all the quirks
of virtual machines, the things they do and don't emulate,
the boot time switches that might be needed to complete
booting. Or, adjustments needed to give acceptable
performance. That would take a few pages of text.
And people hate reading my posts

Do they?

Paul

With multiboot systems you always have to be careful what you are doing.
I tend to believe Windows always ignores existing linux systems so you
have to install Windows first. Linux recognizes a Windows system, but
try to remove a Linux system? It leaves you with a boot manager.

I never had any problems with my VM's. No quirks at all.

Fokke

 

[Paul]

I never had any problems with my VM's. No quirks at all.

Fokke

On VPC2007, these are some things I try on Linux.

Boot options. First option reduces network chatter. Second
option, we can't really be sure it's helping. VPC2007 clock
sources just don't help matters. Third option, helps a
virtual machine shutdown faster. Fourth option had something
to do with getting my mouse scrollwheel to work. Fifth
option is, well, optional.

ipv6.disable=1 clocksource=pit vga=786 psmouse.proto=imps noacpi

I also craft a custom xorg.conf . The framework was generated, by
a physical machine running the NVidia tainted driver package.
The nvidia-xconfig program did a great job creating a template
to start from. And then I added a few things (custom modeline).
On some VMs, the GUI crashes, and you have to bootstrap yourself
in the console text terminal, by entering this crap in VI as
/etc/X11/xorg.conf.

The problem with VPC2007, is the graphics emulation is S3 chipset
based, and the limited graphics memory requires running in
Xserver 16 bit modes. Note - this file does not work right on
Ubuntu 13, and I don't understand why yet. It's like Ubuntu 13
has two rendering planes or methods, and some objects are
horizontally scaled wrong. I don't understand what they've done
to it. Word is, Canonical will eventually replace X, hard as that
is to believe.

*******

# nvidia-xconfig: X configuration file generated by nvidia-xconfig
# nvidia-xconfig: version 256.53 (buildmeister@builder101) Fri Aug 27 21:34:01 PDT 2010

Section "ServerLayout"
Identifier "Layout0"
Screen 0 "Screen0"
InputDevice "Keyboard0" "CoreKeyboard"
InputDevice "Mouse0" "CorePointer"
EndSection

Section "Files"
EndSection

Section "InputDevice"
# generated from default
Identifier "Mouse0"
Driver "mouse"
Option "Protocol" "auto"
Option "Device" "/dev/psaux"
Option "Emulate3Buttons" "no"
Option "ZAxisMapping" "4 5"
EndSection

Section "InputDevice"
# generated from default
Identifier "Keyboard0"
Driver "kbd"
EndSection

Section "Monitor"
Identifier "Monitor0"
VendorName "Unknown"
ModelName "Unknown"
# These are modified so they won't be a limiting factor.
HorizSync 31.5 - 75.0
VertRefresh 50.0 - 100.0
# Modelines via "cvt" - I usually use the first modeline.
# Since the screen is "virtual", the refresh rate is meaningless.
# That's why I can use 50Hz without a problem. The refresh rate
# info likely isn't tied into anything, except some math.
Modeline "1152x864_50.00" 66.25 1152 1208 1320 1488 864 867 871 892 -hsync +vsync
Modeline "1024x768_60.00" 63.50 1024 1072 1176 1328 768 771 775 790 -hsync +vsync
Option "DPMS"
EndSection

Section "Device"
Identifier "Device0"
Driver "s3"
VendorName "Vanilla Corporation"
EndSection

Section "Screen"
Identifier "Screen0"
Device "Device0"
Monitor "Monitor0"
DefaultDepth 16
SubSection "Display"
Depth 16
Modes "1152x864_50.00"
EndSubSection
EndSection

*******

Paul

 

[Roy Smith]

LOL.

The only reason I'm reading this thread, is my Newsreader asked me, do
you want to read this thread? And I answered YES.

Next week I will have a new computer running Windows 8 and I hope you
never hear any complaints from me.

Invest the $5 to buy Start8 and you won't regret it.

 

[Paul]

LOL.

The only reason I'm reading this thread, is my Newsreader asked me, do
you want to read this thread? And I answered YES.

Next week I will have a new computer running Windows 8 and I hope you
never hear any complaints from me.

There's a free introduction to W8 here.

https://marketing.dell.com/Global/FileLib/Windows_8/windows-8-ebook.pdf

Paul

 

[Fokke Nauta]

On VPC2007, these are some things I try on Linux.

Boot options. First option reduces network chatter. Second
option, we can't really be sure it's helping. VPC2007 clock
sources just don't help matters. Third option, helps a
virtual machine shutdown faster. Fourth option had something
to do with getting my mouse scrollwheel to work. Fifth
option is, well, optional.

ipv6.disable=1 clocksource=pit vga=786 psmouse.proto=imps noacpi

I also craft a custom xorg.conf . The framework was generated, by
a physical machine running the NVidia tainted driver package.
The nvidia-xconfig program did a great job creating a template
to start from. And then I added a few things (custom modeline).
On some VMs, the GUI crashes, and you have to bootstrap yourself
in the console text terminal, by entering this crap in VI as
/etc/X11/xorg.conf.

The problem with VPC2007, is the graphics emulation is S3 chipset
based, and the limited graphics memory requires running in
Xserver 16 bit modes. Note - this file does not work right on
Ubuntu 13, and I don't understand why yet. It's like Ubuntu 13
has two rendering planes or methods, and some objects are
horizontally scaled wrong. I don't understand what they've done
to it. Word is, Canonical will eventually replace X, hard as that
is to believe.

*******

# nvidia-xconfig: X configuration file generated by nvidia-xconfig
# nvidia-xconfig: version 256.53 (buildmeister@builder101) Fri Aug 27
21:34:01 PDT 2010

Section "ServerLayout"
Identifier "Layout0"
Screen 0 "Screen0"
InputDevice "Keyboard0" "CoreKeyboard"
InputDevice "Mouse0" "CorePointer"
EndSection

Section "Files"
EndSection

Section "InputDevice"
# generated from default
Identifier "Mouse0"
Driver "mouse"
Option "Protocol" "auto"
Option "Device" "/dev/psaux"
Option "Emulate3Buttons" "no"
Option "ZAxisMapping" "4 5"
EndSection

Section "InputDevice"
# generated from default
Identifier "Keyboard0"
Driver "kbd"
EndSection

Section "Monitor"
Identifier "Monitor0"
VendorName "Unknown"
ModelName "Unknown"
# These are modified so they won't be a limiting factor.
HorizSync 31.5 - 75.0
VertRefresh 50.0 - 100.0
# Modelines via "cvt" - I usually use the first modeline.
# Since the screen is "virtual", the refresh rate is meaningless.
# That's why I can use 50Hz without a problem. The refresh rate
# info likely isn't tied into anything, except some math.
Modeline "1152x864_50.00" 66.25 1152 1208 1320 1488 864 867 871
892 -hsync +vsync
Modeline "1024x768_60.00" 63.50 1024 1072 1176 1328 768 771 775
790 -hsync +vsync
Option "DPMS"
EndSection

Section "Device"
Identifier "Device0"
Driver "s3"
VendorName "Vanilla Corporation"
EndSection

Section "Screen"
Identifier "Screen0"
Device "Device0"
Monitor "Monitor0"
DefaultDepth 16
SubSection "Display"
Depth 16
Modes "1152x864_50.00"
EndSubSection
EndSection

*******

Paul

I replied to this but I can't see my reply!

Fokke

 

[Bob I]

There's a free introduction to W8 here.

Is that pronounced "wait"?

 

[Paul]

Is that pronounced "wait"?

At least _wait_ until you've read the manual.

Paul

 

[Fokke Nauta]

On VPC2007, these are some things I try on Linux.

Boot options. First option reduces network chatter. Second
option, we can't really be sure it's helping. VPC2007 clock
sources just don't help matters. Third option, helps a
virtual machine shutdown faster. Fourth option had something
to do with getting my mouse scrollwheel to work. Fifth
option is, well, optional.

ipv6.disable=1 clocksource=pit vga=786 psmouse.proto=imps noacpi

I also craft a custom xorg.conf . The framework was generated, by
a physical machine running the NVidia tainted driver package.
The nvidia-xconfig program did a great job creating a template
to start from. And then I added a few things (custom modeline).
On some VMs, the GUI crashes, and you have to bootstrap yourself
in the console text terminal, by entering this crap in VI as
/etc/X11/xorg.conf.

The problem with VPC2007, is the graphics emulation is S3 chipset
based, and the limited graphics memory requires running in
Xserver 16 bit modes. Note - this file does not work right on
Ubuntu 13, and I don't understand why yet. It's like Ubuntu 13
has two rendering planes or methods, and some objects are
horizontally scaled wrong. I don't understand what they've done
to it. Word is, Canonical will eventually replace X, hard as that
is to believe.

*******

# nvidia-xconfig: X configuration file generated by nvidia-xconfig
# nvidia-xconfig: version 256.53 (buildmeister@builder101) Fri Aug 27
21:34:01 PDT 2010

Section "ServerLayout"
Identifier "Layout0"
Screen 0 "Screen0"
InputDevice "Keyboard0" "CoreKeyboard"
InputDevice "Mouse0" "CorePointer"
EndSection

Section "Files"
EndSection

Section "InputDevice"
# generated from default
Identifier "Mouse0"
Driver "mouse"
Option "Protocol" "auto"
Option "Device" "/dev/psaux"
Option "Emulate3Buttons" "no"
Option "ZAxisMapping" "4 5"
EndSection

Section "InputDevice"
# generated from default
Identifier "Keyboard0"
Driver "kbd"
EndSection

Section "Monitor"
Identifier "Monitor0"
VendorName "Unknown"
ModelName "Unknown"
# These are modified so they won't be a limiting factor.
HorizSync 31.5 - 75.0
VertRefresh 50.0 - 100.0
# Modelines via "cvt" - I usually use the first modeline.
# Since the screen is "virtual", the refresh rate is meaningless.
# That's why I can use 50Hz without a problem. The refresh rate
# info likely isn't tied into anything, except some math.
Modeline "1152x864_50.00" 66.25 1152 1208 1320 1488 864 867 871
892 -hsync +vsync
Modeline "1024x768_60.00" 63.50 1024 1072 1176 1328 768 771 775
790 -hsync +vsync
Option "DPMS"
EndSection

Section "Device"
Identifier "Device0"
Driver "s3"
VendorName "Vanilla Corporation"
EndSection

Section "Screen"
Identifier "Screen0"
Device "Device0"
Monitor "Monitor0"
DefaultDepth 16
SubSection "Display"
Depth 16
Modes "1152x864_50.00"
EndSubSection
EndSection

*******

Paul

I replied some time ago but my reply was lost :-(
So I'll write it again.

Interesting stuff.
I never had any problems with the X-system or GUI in VMWare. And -
indeed - Ubuntu will replace the X server for a proprietary system. I am
curious as tho what it will be.
They already replaced Gnome for ... "what's it called now? Unity?"
I wasn't so keen on that either.

Haven't tried Ubuntu 13 yet. I'll wait for the final release.

Fokke