[parisc-linux] H60

[parisc-linux] H60

[Davin Petersen]

I've looked through the available documentation available at
http://thepuffingroup.com/parisc/documentation.html and see nothing
related to the 9000/800 H60.  Am I wrong?  Would it be useful in
developing the HP9K kernel?

Re: [parisc-linux] H60

[Stan Sieler]

Re:
> related to the 9000/800 H60.  Am I wrong?  Would it be useful in
> developing the HP9K kernel?

It's the same as a G60 (96 MHz, PA7100, HP-PB, 2048 KB cache)
Also same as I60.

Of course, I don't know if either of those are in the docs :) 

-- 
Stan Sieler                                          sieler@allegro.com
                                         http://www.allegro.com/sieler/

Re: [parisc-linux] H60

[Alex deVries]

Stan Sieler wrote:
> 
> Re:
> > related to the 9000/800 H60.  Am I wrong?  Would it be useful in
> > developing the HP9K kernel?
> 
> It's the same as a G60 (96 MHz, PA7100, HP-PB, 2048 KB cache)
> Also same as I60.
> 
> Of course, I don't know if either of those are in the docs :)

First, yes, the goal is to have it running on the G60/I60 and H60.

However, HP's only released documentation for a certain subset of the
machines available. 

The 7100 CPU shouldn't be too difficult, it's what's in a lot of 712s. 
What kind of peripheral devices are in the G/H/I60?

It'd be good if we could start a document that contained a list of the
internal chips we're finding in all these machines.

- Alex

-- 
Alex deVries
Vice President of Engineering
The Puffin Group

Re: [parisc-linux] H60

[Stan Sieler]

Re:

> The 7100 CPU shouldn't be too difficult, it's what's in a lot of 712s. 
> What kind of peripheral devices are in the G/H/I60?

SCSI disk, SCSI DDS, SCSI CDROM, and (less commonly) HPIB disk, and (rarely)
HPIB DDS, and (possible, but not seen) HPIB CDROM.

Various terminal muxes.
 
> It'd be good if we could start a document that contained a list of the
> internal chips we're finding in all these machines.
 
This may be of interest...a spreadsheet maintained by Michael Hensley
http://www.allegro.com/HPPA_Systems.xls

It has a number of HP 9000 (and HP 3000) models, with speed and chip info.

-- 
Stan Sieler                                          sieler@allegro.com
                                         http://www.allegro.com/sieler/

Re: [parisc-linux] H60

[Rich Rauenzahn]

Take a look at  /usr/lib/sched.models on HP-UX -- it documents quite a few
models and PA-RISC versions.

Rich

----- Original Message -----
From: Alex deVries <adevries@thepuffingroup.com>
To: Stan Sieler <sieler@allegro.com>
Cc: Davin Petersen <davinp@celltech.com>; <parisc-linux@thepuffingroup.com>
Sent: Tuesday, June 29, 1999 5:43 PM
Subject: Re: [parisc-linux] H60



[...]
> It'd be good if we could start a document that contained a list of the
> internal chips we're finding in all these machines.
>
> - Alex

Re: [parisc-linux] H60

[Alex deVries]

Rich Rauenzahn wrote:
> 
> Take a look at  /usr/lib/sched.models on HP-UX -- it documents quite a few
> models and PA-RISC versions.
> 

Oh, that's pretty interesting.  I found sched.modules in
/usr/sam/lib/mo.

What's the second column?

- Alex

-- 
Alex deVries
Vice President of Engineering
The Puffin Group

Re: [parisc-linux] H60

[Richard J. Rauenzahn]

Apparently it exists in quite a few places!

The second column is the PA-RISC version number.  I'm not sure how
significant the letters are in the version numbers.

> Rich Rauenzahn wrote:
> > 
> > Take a look at  /usr/lib/sched.models on HP-UX -- it documents quite a few
> > models and PA-RISC versions.
> > 
> 
> Oh, that's pretty interesting.  I found sched.modules in
> /usr/sam/lib/mo.
> 
> What's the second column?
> 
> - Alex

Re: [parisc-linux] H60

[Mike Hibler]

If the H sei\bries machines really have HP-PB, then it is going to take
some work.  The HP-PB that I am familiar with (from back in the hp832 days)
is a whole other bus architecture that isn't nearly as pleasent to work with
as the \b\bat on the 700 series machines.

RE: [parisc-linux] H60

[Davin Petersen]

Yep, my H60 is all HP-PB.  The series number on the back is an 827s, but
it had a back plane upgrade several years ago to a 887 (H60).  This
particular box has:

2	10Base-2 w/ AUI			28640-60001
1	FWD SCSI				28696-60001
1	SE SCSI/Parallel/MUX		A1703-60022
1	HP6000 Disk Array (RAID5)	C4230A
5	2G disks for the 6000		C2431A
2	1G internal disks
1	HP35480A	DDS tape
1	HP35470A	DDS tape		(one of these is DDS2,
the other is DDS3)
1	External CDROM (SE)		C2948A
1	700/92 Terminal (green, yech)	

-----Original Message-----
From: Mike Hibler [mailto:mike@fast.cs.utah.edu]
Sent: Tuesday, June 29, 1999 10:20 PM
To: adevries@thepuffingroup.com; sieler@allegro.com
Cc: parisc-linux@thepuffingroup.com
Subject: Re: [parisc-linux] H60


If the H sei\bries machines really have HP-PB, then it is going to take
some work.  The HP-PB that I am familiar with (from back in the hp832
days)
is a whole other bus architecture that isn't nearly as pleasent to work
with
as the \b\bat on the 700 series machines.

------------------------------------------------------------------------
---
To unsubscribe: send e-mail to parisc-linux-request@thepuffingroup.com
with
`unsubscribe' as the subject.

Re: [parisc-linux] H60

[Stan Sieler]

Re:

> Yep, my H60 is all HP-PB.  The series number on the back is an 827s, but
> it had a back plane upgrade several years ago to a 887 (H60).  This

Not that it matters much, but the 887 encompasses the G50/H50/I50 ...
all of which are 96 MHz machines with 512 KB cache, PA-7100 with the
PCX-T chipset).  The G60/H60/I60 appear to be slightly different (2 MB
cache, and the "PCX-T" doesn't seem to appear in HP literature). 
Still, the I/O is the same for the 887/G50/H50/I50/G60/H60/I60.

It would be interesting to see what "uname -i" reports on a "real" G60/H60/I60...
will it be 887 or ?

-- 
Stan Sieler                                          sieler@allegro.com
                                         http://www.allegro.com/sieler/

RE: [parisc-linux] H60

[Davin Petersen]

uname -i returns:
900807201

uname -a returns:
HP-UX max B.10.20 A 9000/887 900807201 two-user license

-----Original Message-----
From: Stan Sieler [mailto:sieler@allegro.com]
Sent: Wednesday, June 30, 1999 10:27 AM
To: davinp@celltech.com
Cc: parisc-linux@thepuffingroup.com
Subject: Re: [parisc-linux] H60


Re:

> Yep, my H60 is all HP-PB.  The series number on the back is an 827s,
but
> it had a back plane upgrade several years ago to a 887 (H60).  This

Not that it matters much, but the 887 encompasses the G50/H50/I50 ...
all of which are 96 MHz machines with 512 KB cache, PA-7100 with the
PCX-T chipset).  The G60/H60/I60 appear to be slightly different (2 MB
cache, and the "PCX-T" doesn't seem to appear in HP literature). 
Still, the I/O is the same for the 887/G50/H50/I50/G60/H60/I60.

It would be interesting to see what "uname -i" reports on a "real"
G60/H60/I60...
will it be 887 or ?

-- 
Stan Sieler                                          sieler@allegro.com
                                         http://www.allegro.com/sieler/

RE: [parisc-linux] H60

[Kirk Bresniker]

The letter in the [GHI][567] systems refers to the number of I/O slots
on the backplane. The number designates the processor board. The

50 = Uni 96MHz PA-7100, 256K I$, 256K D$
60 = Uni 96MHz PA-7100, 1024K I$, 1024K D$
70 = Dual 96MHz PA-7100, 1024K I$, 1024K D$ per processor

I wrote an article in the HP Journal about the I70 development, which
has block diagrams which descibe the system at a high level.

PA-RISC Symmetric Multiprocessing in Midrange Servers 
  http://www.hp.com/hpj/94jun/jun94_31.pdf

KMB
--
+============================================================+
|       Kirk Bresniker    	(916) 785-5677		     |
|       8000 Foothills Blvd                                  |
|       Roseville, CA 95747-5649                             |
|       kirkb@rose.hp.com                                    |

Re: [parisc-linux] H60

[Kevin Carson]

As I understand it,

second column consists of the PA-RISC instruction set revision (1.0, 1.1, or
2.0) and the letter indicates what type of CPU is running that revision in
the respective
box (CPU models being also the third column).

The point of all this is that the HP ANSI compiler takes both a instruction
revision the object code is targeted towards (the most portable being the
older revision 1.0 since all the revisions are backward compatible) and a
hint about which implementation the instruction scheduling optimizations
should be targeting (which does affect compatability in any way).

How much difference does the letter portion really make? From practical
experience not much, but it could!  Which is the point of maintaining this
file in the first place.  The instruction set revision makes a BIG
difference for later chips.

Kevin

-----Original Message-----
From: rrauenza@cup.hp.com <rrauenza@cup.hp.com>
To: adevries@thepuffingroup.com <adevries@thepuffingroup.com>
Cc: rrauenza@cup.hp.com <rrauenza@cup.hp.com>;
parisc-linux@thepuffingroup.com <parisc-linux@thepuffingroup.com>
Date: Tuesday, June 29, 1999 11:55 PM
Subject: Re: [parisc-linux] H60


>
>Apparently it exists in quite a few places!
>
>The second column is the PA-RISC version number.  I'm not sure how
>significant the letters are in the version numbers.
>
>> Rich Rauenzahn wrote:
>> >
>> > Take a look at  /usr/lib/sched.models on HP-UX -- it documents quite a
few
>> > models and PA-RISC versions.
>> >
>>
>> Oh, that's pretty interesting.  I found sched.modules in
>> /usr/sam/lib/mo.
>>
>> What's the second column?
>>
>> - Alex
>
>---------------------------------------------------------------------------
>To unsubscribe: send e-mail to parisc-linux-request@thepuffingroup.com with
>`unsubscribe' as the subject.
>