Discussion:
Free Alpha emulators running under Linux ?
Add Reply
Simon Clubley
2022-01-02 09:09:42 UTC
Reply
Permalink
At the moment, I run my home Alpha system on a FreeAXP instance
running under Windows.

It would be nice to switch to running an Alpha emulator under Linux
while switching to the VSI version of Alpha VMS.

What is the current state of play for free Alpha emulators running
under Linux ?

When I last looked at this a number of years ago, the only Linux based
emulator I could find at the time was es40 and that wasn't very usable
at the time.

There's some basic Alpha support in simh, but the last time I looked,
it was missing many of the features that would have made it a viable
full system emulator.

Thanks,

Simon.
--
Simon Clubley, ***@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.
Phil Howell
2022-01-03 07:17:24 UTC
Reply
Permalink
Post by Simon Clubley
At the moment, I run my home Alpha system on a FreeAXP instance
running under Windows.
It would be nice to switch to running an Alpha emulator under Linux
while switching to the VSI version of Alpha VMS.
What is the current state of play for free Alpha emulators running
under Linux ?
When I last looked at this a number of years ago, the only Linux based
emulator I could find at the time was es40 and that wasn't very usable
at the time.
There's some basic Alpha support in simh, but the last time I looked,
it was missing many of the features that would have made it a viable
full system emulator.
Thanks,
Simon.
--
Walking destinations on a map are further away than they appear.
Check out the AXPBox wiki on the GitHub page
https://github-wiki-see.page/m/lenticularis39/axpbox/wiki
and these from Remy
https://raymii.org/s/blog/Installing_the_es40_AlphaServer_emulator_0.18_on_Ubuntu_16.04_and_trying_to_install_openVMS_8.4_on_es40.html#google_vignette
https://raymii.org/s/tutorials/Installing_OpenVMS_8.4_Alpha_in_AXPbox_with_networking.html

Phil
^P
2022-01-03 08:37:20 UTC
Reply
Permalink
On Monday, January 3, 2022 at 8:17:26 AM UTC+1, Phil Howell wrote:
https://raymii.org/s/blog/Installing_the_es40_AlphaServer_emulator_0.18_on_Ubuntu_16.04_and_trying_to_install_openVMS_8.4_on_es40.html#google_vignette
Post by Phil Howell
https://raymii.org/s/tutorials/Installing_OpenVMS_8.4_Alpha_in_AXPbox_with_networking.html
Phil
There is also DECaxp:
https://github.com/JonathanBelanger/DECaxp

^P
Simon Clubley
2022-01-03 10:37:58 UTC
Reply
Permalink
Post by ^P
https://github.com/JonathanBelanger/DECaxp
That's been archived on GitHub and the GitLab page isn't showing for
me for some reason right now. Has anyone tried this and how robust
is it ?

Thanks,

Simon.
--
Simon Clubley, ***@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.
^P
2022-01-03 21:21:46 UTC
Reply
Permalink
Post by Simon Clubley
Post by ^P
https://github.com/JonathanBelanger/DECaxp
That's been archived on GitHub and the GitLab page isn't showing for
me for some reason right now. Has anyone tried this and how robust
is it ?
Thanks,
Simon.
--
Walking destinations on a map are further away than they appear.
Original project is here:
https://gitlab.com/osf1/DECaxp
Simon Clubley
2022-01-04 18:34:05 UTC
Reply
Permalink
Post by ^P
https://gitlab.com/osf1/DECaxp
This is not the original repository. It's my port to get the (then) current version of DECaxp compiled and running on macOS 10.15. The original project is still a work in progress, but there was no activity for quite a while. Not sure, if it's still under development by Jonathan Belanger.
I've now had a look at the original GitLab repository. Last worked on
9 months ago with unanswered issues dating back to a year ago about it
not compiling, so I would assume the answer is no.

I also see the question about whether it can actually run VMS is
still unanswered after 7 months.

Has anyone managed to get the current version to compile and, if so,
have you been able to get it to run VMS ?

Just asking before I spend days trying to get it to work and only then
discover that it's known to be broken or doesn't run VMS properly. :-)

Simon.
--
Simon Clubley, ***@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.
osf1
2022-01-05 08:01:31 UTC
Reply
Permalink
Post by Simon Clubley
I've now had a look at the original GitLab repository. Last worked on
9 months ago with unanswered issues dating back to a year ago about it
not compiling, so I would assume the answer is no.
I also see the question about whether it can actually run VMS is
still unanswered after 7 months.
Has anyone managed to get the current version to compile and, if so,
have you been able to get it to run VMS ?
Just asking before I spend days trying to get it to work and only then
discover that it's known to be broken or doesn't run VMS properly. :-)
Simon.
--
Walking destinations on a map are further away than they appear.
It will not run VMS at this stage. It's missing several parts, to emulate a complete system. I don't remember now, what exactly, but I don't think, the disk I/O parts (ATA/ATAPI) where complete and I don't think the serial ports as well, but haven't checked this.
Simon Clubley
2022-01-05 13:41:04 UTC
Reply
Permalink
Post by osf1
It will not run VMS at this stage. It's missing several parts, to emulate a complete system. I don't remember now, what exactly, but I don't think, the disk I/O parts (ATA/ATAPI) where complete and I don't think the serial ports as well, but haven't checked this.
Thanks for the feedback.

It looks like the last changes were in these areas before the project
was apparently abandoned 9 months ago on GitLab. I may give it a brief
tryout at some point just to see if there's now enough to run VMS.

If I do try it at some point, I will report success or failure back here.

Simon.
--
Simon Clubley, ***@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.
^P
2022-01-05 15:12:09 UTC
Reply
Permalink
Post by Simon Clubley
Post by osf1
It will not run VMS at this stage. It's missing several parts, to emulate a complete system. I don't remember now, what exactly, but I don't think, the disk I/O parts (ATA/ATAPI) where complete and I don't think the serial ports as well, but haven't checked this.
Thanks for the feedback.
It looks like the last changes were in these areas before the project
was apparently abandoned 9 months ago on GitLab. I may give it a brief
tryout at some point just to see if there's now enough to run VMS.
If I do try it at some point, I will report success or failure back here.
Simon.
So it's more or less only AXPBOX left then
cao...@pitbulluk.org
2022-01-05 20:38:20 UTC
Reply
Permalink
Post by ^P
Post by Simon Clubley
Post by osf1
It will not run VMS at this stage. It's missing several parts, to emulate a complete system. I don't remember now, what exactly, but I don't think, the disk I/O parts (ATA/ATAPI) where complete and I don't think the serial ports as well, but haven't checked this.
Thanks for the feedback.
It looks like the last changes were in these areas before the project
was apparently abandoned 9 months ago on GitLab. I may give it a brief
tryout at some point just to see if there's now enough to run VMS.
If I do try it at some point, I will report success or failure back here.
Simon.
So it's more or less only AXPBOX left then
I've noticed that axpbox is more stable when compiled -O2 (rather than the default -O3) though that probably depends on the linux platform - Arm in this case.
Strangely enough, lowering the amount of optimisation like that was also required to get KLH10 to work correctly.

K
Simon Clubley
2022-01-06 18:48:09 UTC
Reply
Permalink
Post by ***@pitbulluk.org
I've noticed that axpbox is more stable when compiled -O2 (rather than the default -O3) though that probably depends on the linux platform - Arm in this case.
Strangely enough, lowering the amount of optimisation like that was also required to get KLH10 to work correctly.
Actually, that's a common enough problem with code in general when the
programmer makes unjustified assumptions about the environment the code
is operating in and those assumptions become invalid because the compiler
applies fully legal optimisations which then break the code.

It's quite possible for example, that code which is accidentally working
using the default settings with the existing GEM based VMS compilers may
actually break when the same code is compiled under the new upcoming
LLVM-based compilers.

Simon.
--
Simon Clubley, ***@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.
John Reagan
2022-01-06 21:51:48 UTC
Reply
Permalink
Post by Simon Clubley
It's quite possible for example, that code which is accidentally working
using the default settings with the existing GEM based VMS compilers may
actually break when the same code is compiled under the new upcoming
LLVM-based compilers.
Simon.
Since that happened when porting from Alpha to Itanium (both GEM) or even
when we added features to GEM, it is guaranteed to happen again.
David Turner
2022-01-07 19:35:07 UTC
Reply
Permalink
Artem Alimarin offers a CHEAP version of AlphaVM
Meant for Educators and Hobbyists. It runs on Linux/Windows and he has
built a MAC variant too
Best emulator out there

www.emuvm.com
Post by John Reagan
Post by Simon Clubley
It's quite possible for example, that code which is accidentally working
using the default settings with the existing GEM based VMS compilers may
actually break when the same code is compiled under the new upcoming
LLVM-based compilers.
Simon.
Since that happened when porting from Alpha to Itanium (both GEM) or even
when we added features to GEM, it is guaranteed to happen again.
Arne Vajhøj
2022-01-07 19:50:07 UTC
Reply
Permalink
Post by David Turner
Post by John Reagan
Post by Simon Clubley
It's quite possible for example, that code which is accidentally working
using the default settings with the existing GEM based VMS compilers may
actually break when the same code is compiled under the new upcoming
LLVM-based compilers.
Simon.
Since that happened when porting from Alpha to Itanium (both GEM) or even
when we added features to GEM, it is guaranteed to happen again.
Artem Alimarin offers a CHEAP version of AlphaVM
Meant for Educators and Hobbyists. It runs on Linux/Windows and he has
built a MAC variant too
Best emulator out there
www.emuvm.com
The product is great. I have used it a bit and it works
perfect.

But cheap is 500 dollar for first year and 250 dollar per
year after that.

Compared to DEC prices in the past that is very cheap.

But compared to VMS x86-64 in the future it is not
so cheap.

You can run a 1 VCPU 2 GB memory instance on AWS EC2
24 x 7 for less than 250 dollar per year. And if just
run occasionally then it will cost nothing.

Arne
John Dallman
1970-01-01 00:00:00 UTC
Reply
Permalink
It does raise one interesting question however: what does the
concept of licencing units even mean when it comes to x86-64 VMS?
Oracle deals with this by charging for their database by the number of
processor cores you use, with a scaling factor based on the processor
model.

<https://www.oracle.com/assets/databaselicensing-070584.pdf>

<http://www.oracle.com/us/corporate/contracts/processor-core-factor-table-
070634.pdf>

It isn't a perfect system, as shown by the number of companies offering
to help optimise Oracle license costs, but it seems to work well enough.

John
Arne Vajhøj
2022-01-03 15:38:53 UTC
Reply
Permalink
It does raise one interesting question however: what does the
concept of licencing units even mean when it comes to x86-64 VMS?
I imagine they will do it the way they do Integrity licensing - probably
per socket but maybe per core, or as with Integrity, they may give you a
choice of one or the other when you buy the license.
I don't think per socket will make much sense in the future. Per
core or actually VCPU is what makes sense.

Examples:

You have VMS running on a 4 VCPU VM on ESXi that run 19 other
VMS on a 2s32c64t box and then you move it all to a new 2s48c96t box.

You have VMS running on a 4 VCPU EC2 instance and have no idea
what the HW is.

In those cases sockets does not have any relevant meaning.

Arne
Subcommandante XDelta
2022-01-08 04:11:55 UTC
Reply
Permalink
On Sun, 2 Jan 2022 09:09:42 -0000 (UTC), Simon Clubley
Post by Simon Clubley
At the moment, I run my home Alpha system on a FreeAXP instance
running under Windows.
It would be nice to switch to running an Alpha emulator under Linux
while switching to the VSI version of Alpha VMS.
What is the current state of play for free Alpha emulators running
under Linux ?
When I last looked at this a number of years ago, the only Linux based
emulator I could find at the time was es40 and that wasn't very usable
at the time.
There's some basic Alpha support in simh, but the last time I looked,
it was missing many of the features that would have made it a viable
full system emulator.
Thanks,
Simon.
Whilst strictly off-topic, relative to your specific query about free
AXP/Linux emulators, the VLF hate to see good digits suffer needlessly
- this should help you potter about, happily enough, in your
temperature controlled, false floor, backyard shed, with your hobbyist
VMS systems:

Stromasys.CHARON-AXP.v4.2.Build.142-01.68704122.Retail.GNU.Linux.x64.Cracked-VLF

File-hosts:

https://www.mirrored.to/files/0UAHF0R6/

Usenet Binaries Newsgroups:

https://binsearch.info/?q=Stromasys.CHARON-AXP.v4.2.Build.142-01.68704122.Retail.GNU.Linux.x64.Cracked-VLF

Old (2013), but solid gold.

We don't like to see good digits suffer, nor do we think that
Stromasys, historically, have suffered - no principals, in their right
minds, of any business or institution, would risk their enterprise, or
organisation, by running their business critical, or mission critical,
production systems, on a cracked hardware emulator - you would want
established business relationships with the hardware emulator
developers, so that there are no delays, whatsoever, in getting
technical support.

We only formally, warrant that our work is sufficient for hobbyist
systems - however, at a pinch, you could also duct tape implement them
in a nuclear power station, to get things working again, if need be,
and if there was no other timely alternative.

Best regards,

VLF: The VMS Liberation Front
In VMS We Trust.

PS: A shout out and salute to Gérard Calliet and the team at the
VMSgenerations Working Group - the VLF greatly admire their above
ground work, from afar - industry association, institutional response,
aggregate activism, in the relation to VSI is the way to go -
individual petitioning will not cut it.

$!------------------------- damn straight --------------------------!$
$ opprobrium/level=kittens/mode=conniptions/input=VLF:/output=NL: !$
$!-------------------------- enough said ---------------------------!$
Loading...