On Sun, 2023-04-16 at 16:32 -0500, John H Reinhardt wrote:
> On 4/16/2023 3:37 PM, Single Stage to Orbit wrote:
> > On Sun, 2023-04-16 at 12:06 -0400, fsword007@gmail.com wrote:
> > > Many modern CPU processors has more then 6 cores.  Does it affect
> > > VMS
> > > clusters?  Mine has 12 cores (i7-8700K).
> >
> > I can confirm mine has 8 logical codes assigned to the OpenVMS VM.
> > It
> > boots up and sees all 8. I think I could add even more if I wanted
> > to.
>
> How many license units does it use? My virtual machine with 4 vCPU
> used 4 of the 6 available.
>
> $ show lice/unit
> ACMS               VSI                 6          4          2  Yes
>
> *snip*

You are right. Although it does see 8 cores, it shows this:
$ sh lic/usage

View of loaded licenses from node GHOST1 17-APR-2023
08:41:44.15

------- Product ID -------- ---- Unit usage information -------------
---
Product Producer Loaded Allocated Available
Compliance
BASIC VSI 6 0 6 Yes
C VSI 6 0 6 Yes
COBOL VSI 6 0 6 Yes
CXX-V VSI 6 0 6 Yes
FORTRAN VSI 6 0 6 Yes
PASCAL VSI 6 0 6 Yes

<snip>

(not sure why it's not showing VMS BOE as licenced when it is!)
$ sh proc

17-APR-2023 08:41:55.49 User: SYSTEM Process ID: 00000425
Node: GHOST1 Process name: "SYSTEM"

Terminal: OPA0:
User Identifier: [SYSTEM]
Base priority: 4
Default file spec: SYS$SYSROOT:[SYSMGR]
Number of Kthreads: 1 (System-wide limit: 8)

So it does see all the cores ...
--
Tactical Nuclear Kittens

On 4/17/2023 2:45 AM, Single Stage to Orbit wrote:
> On Sun, 2023-04-16 at 16:32 -0500, John H Reinhardt wrote:
>> On 4/16/2023 3:37 PM, Single Stage to Orbit wrote:
>>> On Sun, 2023-04-16 at 12:06 -0400, fsword007@gmail.com wrote:
>>>> Many modern CPU processors has more then 6 cores.  Does it affect
>>>> VMS
>>>> clusters?  Mine has 12 cores (i7-8700K).
>>>
>>> I can confirm mine has 8 logical codes assigned to the OpenVMS VM.
>>> It
>>> boots up and sees all 8. I think I could add even more if I wanted
>>> to.
>>
>> How many license units does it use? My virtual machine with 4 vCPU
>> used 4 of the 6 available.
>>
>> $ show lice/unit
>> ACMS               VSI                 6          4          2  Yes
>>
>> *snip*
>
> You are right. Although it does see 8 cores, it shows this:
> $ sh lic/usage
>
> View of loaded licenses from node GHOST1 17-APR-2023
> 08:41:44.15
>
> ------- Product ID -------- ---- Unit usage information -------------
> ---
> Product Producer Loaded Allocated Available
> Compliance
> BASIC VSI 6 0 6 Yes
> C VSI 6 0 6 Yes
> COBOL VSI 6 0 6 Yes
> CXX-V VSI 6 0 6 Yes
> FORTRAN VSI 6 0 6 Yes
> PASCAL VSI 6 0 6 Yes
>
> <snip>
>
> (not sure why it's not showing VMS BOE as licenced when it is!)

Mine is:
OPENVMS-X86-BOE VSI 6 4 2 Yes
OPENVMS-X86-HAOE VSI 6 4 2 Yes

I have not loaded any of the compilers yet and they are all like yours above, Loaded: 6, Allocated: 0, Available: 6 I guess because it knows they haven't been installed?

>
> $ sh proc
>
> 17-APR-2023 08:41:55.49 User: SYSTEM Process ID: 00000425
> Node: GHOST1 Process name: "SYSTEM"
>
> Terminal: OPA0:
> User Identifier: [SYSTEM]
> Base priority: 4
> Default file spec: SYS$SYSROOT:[SYSMGR]
> Number of Kthreads: 1 (System-wide limit: 8)
>
> So it does see all the cores ...

Interesting. Mine has a system wide limit of 2. I would have expected 4. But when I first install VMS I had a vCPU of 2. Maybe something is set upon install, not startup.

Try a SHOW LICE/UNIT since that should also show active CPU count

$ show lice/unit
This is an VMware, Inc. VMware7,1, with 4 cores active
Type: PCL, Units Required: 4 (X86_64 Per Core)
$

If you do a SHOW CPU does it show all 8 as being active?

$ show cpu

System: SCALZI, VMware, Inc. VMware7,1

CPU ownership sets:
Active 0-3
Configure 0-3

CPU state sets:
Potential 0-3
Autostart 0-3
Powered Down None
Not Present None
Hard Excluded None
Failover None
$

----------------
John H. Reinhardt

On Mon, 2023-04-17 at 03:22 -0500, John H Reinhardt wrote:
> > > *snip*
> >
> > (not sure why it's not showing VMS BOE as licenced when it is!)
>
> Mine is:
> OPENVMS-X86-BOE    VSI                 6          4          2  Yes
> OPENVMS-X86-HAOE   VSI                 6          4          2  Yes
>

Looks like I forgot to load the licence for BOE! It's now loaded and
all looks good.

>
> teresting.  Mine has a system wide limit of 2.  I would have expected
> 4.  But when I first install VMS I had a vCPU of 2.  Maybe something
> is set upon install, not startup.
>
> Try a SHOW LICE/UNIT since that should also show active CPU count

> $ show lice/unit
> This is an VMware, Inc. VMware7,1, with 4 cores active
> Type: PCL,   Units Required: 4  (X86_64 Per Core)
> $
>
> If you do a SHOW CPU does it show all 8 as being active?
>
> $ show cpu
>
> System: SCALZI, VMware, Inc. VMware7,1
>
> CPU ownership sets:
>     Active               0-3
>     Configure            0-3
>
> CPU state sets:
>     Potential            0-3
>     Autostart            0-3
>     Powered Down         None
>     Not Present          None
>     Hard Excluded        None
>     Failover             None
> $

$ sh lic/unit
This is an innotek GmbH VirtualBox, with 8 cores active
Type: PCL, Units Required: 8 (X86_64 Per Core)
$ sh cpu

System: GHOST1, innotek GmbH VirtualBox

CPU ownership sets:
Active 0-7
Configure 0-7

CPU state sets:
Potential 0-7
Autostart 0-7
Powered Down None
Not Present None
Hard Excluded None
Failover None

Looks like the licences limit the software but not the hardware?
--
Tactical Nuclear Kittens

I have a few questions. Does OpenVMS support IOMMU for developing device
drivers like GPU, etc?

Tim

-----Original Message-----
From: Info-vax <info-vax-bounces@rbnsn.com> On Behalf Of Scott Dorsey via
Info-vax
Sent: Sunday, April 16, 2023 10:44 AM
To: info-vax@rbnsn.com
Cc: Scott Dorsey <kludge@panix.com>
Subject: Re: [Info-vax] VSI Community License Program - x86

Andrew Back <andrew@carrierdetect.com> wrote:
>On Sunday, 16 April 2023 at 04:19:03 UTC+1, Mark Berryman wrote:
>> I do not believe the issue is CPU related. My understanding is that
>> it=20 is only Gen 9 and later that use UEFI. Previous generations
>> are=20 apparently BIOS based. VMS requires UEFI.=20
>
>In that case I should think you could run it on earlier than Gen 9,
>albeit = via Qemu + KVM, where the virtual machine is configured to use
>OVMF bios (T= ianoCore UEFI port).

Yes. As I pointed out, it's possible to run it under a VM on some pretty
slow and obsolete hardware, so the issues with running on bare metal have
nothing to do with CPU demands but the boot process and the device drivers.

>Interesting also to see someone has had bare metal success with an Asus
>mai= nboard. Has me wondering how much effort would be involved in
>porting the A= MDgpu open source driver, plus others as required, to
>get OpenVMS up and ru= nning on a modern laptop.

You need to provide device drivers for all the disk controllers, for the USB
controller, for the wifi controller, for the ethernet controller, and in the
case of the laptop there's a bunch of stuff for power management that is
very
problematic. The laptop is EXACTLY the sort of application where
virtualization is a clear win.

Right now I am thinking the primary limitations for support have to do with
VSI supplying only the drivers for the devices available with the Gen 9
server and for a few virtual devices. With more drivers and a little bit of
sysgen stuff, I am sure it would be possible to run it on the bare metal of
other machines... but the number of different ethernet chipsets alone makes
for difficulties in supporting a wide number of machines. This is how the
PC world is.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
_______________________________________________
Info-vax mailing list
Info-vax@rbnsn.com
http://rbnsn.com/mailman/listinfo/info-vax_rbnsn.com

On 4/15/2023 11:54 PM, Volker Halle wrote:
> John H Reinhardt schrieb am Sonntag, 16. April 2023 um 06:33:01 UTC+2:
>
>> The answer to your question is none since as of the moment, it's not supported nor able to run on bare metal anything. The only drivers are for virtualized devices and hosts.
>
> There is a report over in the VSI Forum of a success story running VSI OpenVMS x86-64 E9.2-1 on baremetal:
>
> https://forum.vmssoftware.com/viewtopic.php?f=21&t=8621
>
> And that forum is currently very busy with posts relating to OpenVMS x86-64 ;-)
>
> Volker.

Thanks Volker! I had missed that. Looks like it was posted about 15 minutes before I did here. Of course, Scot asked for "supported" and we know what that means. ;) I did a quick Google search and that ASRock motherboard looks to be about 8 years old and out of stock most places. Used on Ebay for prices between $150 and $500. Could be an interesting experiment. Have to watch the # of cores in the CPU you use as the VSI license only allows 6.

But, yes, the VSI Forums have been very busy lately. Which is good!

John H. Reinhardt

Many modern CPU processors has more then 6 cores. Does it affect VMS
clusters? Mine has 12 cores (i7-8700K).

Tim

-----Original Message-----
From: Info-vax <info-vax-bounces@rbnsn.com> On Behalf Of John H Reinhardt
via Info-vax
Sent: Sunday, April 16, 2023 11:51 AM
To: info-vax@rbnsn.com
Cc: John H Reinhardt <johnhreinhardt@thereinhardts.org>
Subject: Re: [Info-vax] VSI Community License Program - x86

On 4/15/2023 11:54 PM, Volker Halle wrote:
> John H Reinhardt schrieb am Sonntag, 16. April 2023 um 06:33:01 UTC+2:
>
>> The answer to your question is none since as of the moment, it's not
supported nor able to run on bare metal anything. The only drivers are for
virtualized devices and hosts.
>
> There is a report over in the VSI Forum of a success story running VSI
OpenVMS x86-64 E9.2-1 on baremetal:
>
> https://forum.vmssoftware.com/viewtopic.php?f=21&t=8621
>
> And that forum is currently very busy with posts relating to OpenVMS
x86-64 ;-)
>
> Volker.

Thanks Volker! I had missed that. Looks like it was posted about 15 minutes
before I did here. Of course, Scot asked for "supported" and we know what
that means. ;) I did a quick Google search and that ASRock motherboard looks
to be about 8 years old and out of stock most places. Used on Ebay for
prices between $150 and $500. Could be an interesting experiment. Have to
watch the # of cores in the CPU you use as the VSI license only allows 6.

But, yes, the VSI Forums have been very busy lately. Which is good!

John H. Reinhardt
_______________________________________________
Info-vax mailing list
Info-vax@rbnsn.com
http://rbnsn.com/mailman/listinfo/info-vax_rbnsn.com

On Saturday, April 15, 2023 at 3:07:28 PM UTC-4, Arne Vajhøj wrote:
> On 4/15/2023 11:17 AM, Chris Townley wrote:
> > On 15/04/2023 16:10, Scott Dorsey wrote:
> >> John H Reinhardt <johnhre...@thereinhardts.org> wrote:
> >>> Robert Brooks and others (I think) have in past times stated
> >>> repeatedly that a Gen 9 or later HP Proliant will be required for
> >>> bare metal. Since the CPU characteristics are usually propagated
> >>> through by the VM hypervisor, that would seem to mean that VM Hosts
> >>> must be Gen 9 or higher as well, but I'm just guessing at that..
> >>
> >> Oh no, I know folks who are running a test VM server on some junk
> >> Optiplex
> >> that's more than a decade old, and it's slow but working. I am guessing
> >> it's a matter of driver support if it really does require the Gen 9.
> >
> > The advice from VSI was fairly woolly:
> >
> > Most Intel CPUs from 2016 and later support these processor features
> I thougth VSI recommended running vmscheck.py to verify if the
> CPU will do?
>
> Arne

Is that script (vmscheck.py) publicly available?

Neil Rieck
Waterloo, Ontario, Canada.
https://neilrieck.net

On 03/11/2023 14:37, Neil Rieck wrote:
> On Saturday, April 15, 2023 at 3:07:28 PM UTC-4, Arne Vajhøj wrote:
>> On 4/15/2023 11:17 AM, Chris Townley wrote:
>>> On 15/04/2023 16:10, Scott Dorsey wrote:
>>>> John H Reinhardt <johnhre...@thereinhardts.org> wrote:
>>>>> Robert Brooks and others (I think) have in past times stated
>>>>> repeatedly that a Gen 9 or later HP Proliant will be required for
>>>>> bare metal. Since the CPU characteristics are usually propagated
>>>>> through by the VM hypervisor, that would seem to mean that VM Hosts
>>>>> must be Gen 9 or higher as well, but I'm just guessing at that..
>>>>
>>>> Oh no, I know folks who are running a test VM server on some junk
>>>> Optiplex
>>>> that's more than a decade old, and it's slow but working. I am guessing
>>>> it's a matter of driver support if it really does require the Gen 9.
>>>
>>> The advice from VSI was fairly woolly:
>>>
>>> Most Intel CPUs from 2016 and later support these processor features
>> I thougth VSI recommended running vmscheck.py to verify if the
>> CPU will do?
>>
>> Arne
>
> Is that script (vmscheck.py) publicly available?
>
> Neil Rieck
> Waterloo, Ontario, Canada.
> https://neilrieck.net

See https://vmssoftware.com/openkits/alpopensource/vmscheck.zip

--
Chris

On 11/3/2023 10:37 AM, Neil Rieck wrote:
> On Saturday, April 15, 2023 at 3:07:28 PM UTC-4, Arne Vajhøj wrote:
>> On 4/15/2023 11:17 AM, Chris Townley wrote:
>>> On 15/04/2023 16:10, Scott Dorsey wrote:
>>>> John H Reinhardt <johnhre...@thereinhardts.org> wrote:
>>>>> Robert Brooks and others (I think) have in past times stated
>>>>> repeatedly that a Gen 9 or later HP Proliant will be required for
>>>>> bare metal. Since the CPU characteristics are usually propagated
>>>>> through by the VM hypervisor, that would seem to mean that VM Hosts
>>>>> must be Gen 9 or higher as well, but I'm just guessing at that..
>>>>
>>>> Oh no, I know folks who are running a test VM server on some junk
>>>> Optiplex
>>>> that's more than a decade old, and it's slow but working. I am guessing
>>>> it's a matter of driver support if it really does require the Gen 9.
>>>
>>> The advice from VSI was fairly woolly:
>>>
>>> Most Intel CPUs from 2016 and later support these processor features
>> I thougth VSI recommended running vmscheck.py to verify if the
>> CPU will do?
>
> Is that script (vmscheck.py) publicly available?

Try:

https://vmssoftware.com/openkits/alpopensource/vmscheck.zip

Arne