On 8 Jul 2011, at 01:09, MG wrote:
Your point about the limited amount of storage space is what surprises many people.
In a way it also makes a (tiny) bit of sense, especially as the bigger players usually
tend to boot off SAN and larger, centralized, disk systems anyway.
Apropos SANs: Has anyone played with the iSCSI initiator in 8.4? That could be a pretty killer feature for me, I'm sure for others too.
Sampsa
On 7-7-2011 10:04, Mark Wickens wrote:
http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=350475648545&ssPageNam…
Fully loaded RX4640 up for GBP 640, that would put a dent in your electricity bill.
Always amazed me that these came with only two drives built in, given the size of the case!
That's a great find, two available even. Thanks for sharing! Beware though, this is not as
loaded out as one of these can be. In fact, for the power it must consume and heat it that
it will dissipate, it's perhaps not a very wise choice (especially for a Hobbyist) and you
might as well max. it out, so at least you'll get a good performance benefit/return on the
power usage.
The rx4640 is interesting as it makes use of so-called "cell"-architecture, unlike the single-
"cell" "entry-level" systems (like the rx1600/rx1620, zx6000/rx2600/rx2620 and so forth), thus
a bit more like the 'big iron' Integrity systems (like the rx7620, rx8620 and the Superdome).
I believe the rx4640 can take two, which can be fitted with four single-core processors each
or two "mx2" dual-processor modules (as seen in the first Integrity Superdome systems).
Then there are also "memory expansion boards", I'm not sure how many DIMMs go on each of these,
but I vaguely recall it can take a maximum of 128 GB. (In the case of the "cell"/processor
boards, it seems maxed out, but the auction doesn't appear to mention too much on the 'memory
boards'. it might be worth to ask the seller about, before making an offer or buying.)
A friend of mine has one of these (with lesser specifications), he bought it a couple of
years ago, for 1300 at the time! He still hasn't gotten around to installing VMS yet, I
hope he will soon.
Your point about the limited amount of storage space is what surprises many people.
In a way it also makes a (tiny) bit of sense, especially as the bigger players usually
tend to boot off SAN and larger, centralized, disk systems anyway.
- MG
On 8-7-2011 1:17, Sampsa Laine wrote:
Marco,
Is it possible to run VMS in a virtual machine (I think 8.4 supports
being a guest on Itanium) on a zx6000? What about a rx2600?
Unfortunately, no. The version is right, but the rx2600 is not of a recent enough
architecture to support HPVM (or "IntegrityVM"). One must at least have a 9000
(Montecito), 9100 (Montvale) or 9300 (Tukwila) series IA-64 processor. The rx2600
supports the Madison (at 1500 MHz, with 6 MB cache) at most. Then, as Mark just
said, there's also the hypervisor and licensing issue.
On the HP public/anonymous FTP, there's a HP-UX .depot-file that contains "HPVM"
in the filename. Maybe it's worth taking a look at? I guess it can't hurt to try?
But, don't count too much on it.
- MG
On 8-7-2011 1:03, Sampsa Laine wrote:
I _think_ this just means that OPA0: will be allocated to the serial port,
not that DECWindows won't run - am I wrong?
So once it's all installed and running, you'll get a X session up on the
graphics card - you just can't use it as a VMS OP console..
That's right, though, in his specific case DECwindows wouldn't work as there
is no support for ATi FireGL graphics adapters.
I heard HP is considering to support some new graphics adapters, like the
AMD/ATi FireMV 2250 (for which there should already be HP-UX support, though
only 'accelerated 2D'), because of some of requests and demand from several
big customers. Demand for graphics support for these platforms is still
there, which is both interesting and nice, especially as we're told that it
is 'dead' and that 'nobody' allegedly 'cares' for it.
The only problem is that these modern cards are PCIe (PCI-Express) models,
thus won't be of much use for someone with a rx2600 with only PCI/PCI-X
slots...
- MG
On 8 Jul 2011, at 00:17, Sampsa Laine wrote:
Marco,
Is it possible to run VMS in a virtual machine (I think 8.4 supports being a guest on Itanium) on a zx6000? What about a rx2600?
Yes, 8.4 Supports that but no I don't know how they hell to do it - I assume it requires a copy of/license for the HP VM hypervisor to do it?
--
Mark Benson
My Blog:
<http://markbenson.org/blog>
Follow me on Twitter:
http://twitter.com/mdbenson
"Never send a human to do a machine's job..."
On 8 Jul 2011, at 00:03, Sampsa Laine wrote:
I _think_ this just means that OPA0: will be allocated to the serial port, not that DECWindows won't run - am I wrong?
So once it's all installed and running, you'll get a X session up on the graphics card - you just can't use it as a VMS OP console..
I had a dig and only the Radeon 7500 is supported in OpenVMS - mine has a FireGL X1.
Fun fact: If you install Debian Squeeze IA64 on a zx6000 with a FireGL X1 in Xorg thinks it's an IBM GXT6000P and configures it completely wrongly. So much for the magic of Xorg - bring back X11R6 and manual config files :)
--
Mark Benson
My Blog:
<http://markbenson.org/blog>
Follow me on Twitter:
http://twitter.com/mdbenson
"Never send a human to do a machine's job..."
Marco,
Is it possible to run VMS in a virtual machine (I think 8.4 supports being a guest on Itanium) on a zx6000? What about a rx2600?
Sampsa
On 8 Jul 2011, at 00:11, MG wrote:
On 8-7-2011 1:01, Mark Benson wrote:
Ahh to hell with that then - I'm not running a Serial Console to a machine with a ATi FireGL card in :P
Unfortunately, you won't be able to take advantage of that particular graphics
adapter under OpenVMS anyway, unless you make it happen (I also happen to have
one of those cards and maybe I'll give that a try some day, to attempt writing
a 'simple' driver, perhaps based on the Linux open source driver).
The 'last' and 'best' card that is available for OpenVMS, which will even do
OpenGL (or "Open3D") is the ATi Radeon 7500. Beware, most Radeon 7500 cards
offered for sale nowadays are of the so-called "Mobility" type and it's often
not mentioned or not very clearly. One of the best ways of identifying one is
the fact that they typically only came with 64 MB memory.
- MG
On 8-7-2011 1:01, Mark Benson wrote:
Ahh to hell with that then - I'm not running a Serial Console to a machine with a ATi FireGL card in :P
Unfortunately, you won't be able to take advantage of that particular graphics
adapter under OpenVMS anyway, unless you make it happen (I also happen to have
one of those cards and maybe I'll give that a try some day, to attempt writing
a 'simple' driver, perhaps based on the Linux open source driver).
The 'last' and 'best' card that is available for OpenVMS, which will even do
OpenGL (or "Open3D") is the ATi Radeon 7500. Beware, most Radeon 7500 cards
offered for sale nowadays are of the so-called "Mobility" type and it's often
not mentioned or not very clearly. One of the best ways of identifying one is
the fact that they typically only came with 64 MB memory.
- MG
I _think_ this just means that OPA0: will be allocated to the serial port, not that DECWindows won't run - am I wrong?
So once it's all installed and running, you'll get a X session up on the graphics card - you just can't use it as a VMS OP console..
Sampsa
On 8 Jul 2011, at 00:01, Mark Benson wrote:
On 7 Jul 2011, at 23:47, MG wrote:
For the zx6000/rx2600 there's unfortunately no 'glass terminal' (only the rx2620
has that, since it supports newer firmware).
Ahh to hell with that then - I'm not running a Serial Console to a machine with a ATi FireGL card in :P
--
Mark Benson
My Blog:
<http://markbenson.org/blog>
Follow me on Twitter:
http://twitter.com/mdbenson
"Never send a human to do a machine's job..."
On 7 Jul 2011, at 23:47, MG wrote:
For the zx6000/rx2600 there's unfortunately no 'glass terminal' (only the rx2620
has that, since it supports newer firmware).
Ahh to hell with that then - I'm not running a Serial Console to a machine with a ATi FireGL card in :P
--
Mark Benson
My Blog:
<http://markbenson.org/blog>
Follow me on Twitter:
http://twitter.com/mdbenson
"Never send a human to do a machine's job..."
