By the end of this year, the Intel Pentium IV will reach and probably pass
the 3GHz mark.  By the end of 2003, about the time a platform emulator would
be ready/available/possible to sell  (I don't see this happening before
EOSales,) the high-end Pentium IV should be beyond 5GHz, approaching 6GHz.
This is a very, very far cry from the 110 (ha!) MHz of the A-class and the
400 or 550 MHz of the N-class by at least 3 orders of magnitude.  Couple
this with faster disks (arrays) and faster memory, even a somewhat
inefficient platform emulator would be tremendously fast.

By YE2004, a high-end system could have 10GHz Pentium IV or probably some
Itanium procs in it.  Four orders of magnitude.

The laptops are now 2GHz in speed and should reach about 4 by YE2003.
Couple this with serial ATA disk and they too will be cooking.  :)

All this to say, that even with a single CPU, the combination should be a
lot faster than current MPE machines or even what they would be at the
snail's pace of development exhibited by HP over the last decade.  So, even
if you looked at the number of CPU's it would be irrelevant.  I sure hope we
are not going to do some fancy crippling here.  :)


Denys...

-----Original Message-----
From: OpenMPE Support Group [mailto:[log in to unmask]]On Behalf Of Gavin
Scott
Sent: Tuesday, October 01, 2002 5:42 PM
To: [log in to unmask]
Subject: Re: PA-RISC Emulator - Distribution

Denys writes:
> On a PA-RISC system, the CPUNAME sorta, kinda reflected the power of the
> system.

When the emulator runs so fast that this is a problem then I'll be very
happy to worry about the problem :-)

Initially it could reflect the number of CPUs in the box, plus which "model"
of the emulator you're running (Hobbyist/Developer/Production).

> Also, before the PCI systems, there were various levels of user
> licenses.

But all MPE systems sold by HP today carry an unlimited user license.

> Let's say somehow the emulator can detect that it is running on
> an "HP" box, (difficult but perhaps possible by interrogating the BIOS
> directly,)

It's been determined that this is completely possible, yes.

> how does one control having the platform emulator run on say an
> 8-way Pentium IV VS say a 2 way Pentium III or a Laptop?

:SHOWVAR HPCPUNAME
HPCPUNAME = SERIES 999 (Emulator, Windows, Intel, P4, 8-Way)

or something like that.

> Does one say "if you have an A-class type license, you must run on a
> 486/100.  If you have an N-class license, you can run on a Pentium IV
> multiprocessor system."?

Nobody has ever suggested that there would be more than one "class" of
MPE/iX license for an emulator.  So if you transfer a 917 license to an
emulator or you transfer an N-Class license to an emulator, you'll likely
end up with the same thing.

Now there might be more than one "model" of emulator, such as a cheap
Developer version that is somehow limited in comparison to an unlimited
Production version.  This limitation could take many forms (memory, disk
space, users, connections, etc.) but the goal would be to provide a cheap
version that lets everyone have that laptop 3000 they've always wanted and
yet be able to charge real money in the case that the user is going to run
real production processing on it.

We continue to press HP to offer both a Developer (not Hobbyist!) license
and a Production license.  Of course having both one unrestricted free
license would be optimal from *our* point of view :-)

> Should the platform emulator make a connection between the
> underlying system and the license?

No?

Gavin