23 Oct
2007
23 Oct
'07
2:37 a.m.
Sridhar Ayengar wrote:
David Griffith wrote:
Didn't the O2, O20 (Octane), O200 and O2000 come out at roughly the same
time?
More or less. Besides that, the Indigo2 and Indy represent a split in the low-end line,
both being offspring of the Indigo, with O2 being Indy's successor and Octane being
Indigo2's successor. The O200 was the offspring of the Challenge S (and M) low-end
server offering. The Fuel and Tezro are also pretty much complementary machines.
Also, since Octane sported the Cube logo, I seriously doubt that 1999 introduction date.
The logo was changed in 1997/1998 and debuted on the Octane2/O2+ in the IRIS workstation
line.
Here's the family tree, then:
/-> Indigo2 -> Octane -> Octane2 -> Tezro
Personal Iris -> Indigo -<
\-> Indy -> O2 -> O2+ -> Fuel
,xtG
tsooJ
I think
it goes like this, someone please correct me if I'm wrong:
1984 IRIS (terminal/workstation)
1988? Personal Iris
1990 Indigo
1992 Indigo^2
1993 Indy
1996 O2
1999 Octane
2000 Octane2
2002 Fuel
2003 Tezro
I seem to recall the O2 coming out in 1999, roughly at the same time as
the Octane. 
23 Oct
23 Oct
3:02 a.m.
On 10/23/07, jvdg at sparcpark.net <jvdg at sparcpark.net> wrote:
Sridhar Ayengar wrote:
> David Griffith wrote:
>>> I think it goes like this, someone please correct me if I'm wrong:
>>>
>>> 1992 Indigo^2
>>> 1993 Indy
More or less. Besides that, the Indigo2 and Indy
represent a split in the low-end line, both being offspring of the Indigo, with O2 being
Indy's successor and Octane being Indigo2's successor.
I've had some interest, off and on, of fiddling with some SGI
hardware. I really wanted an Indy when they came out, but couldn't
come close to affording it (I ended up buying a SPARC 1 for $800 the
following year, and many, many job opportunities and good things came
from that). I did get to fiddle with an Indigo^2 in
the mid-1990s,
during the Virtual Reality bubble, but couldn't afford one of
those at
the time, either.
I had the chance to pick up a discarded Indy about 4 years ago, but by
the time I returned with wheels, someone else had carted it away.
Given the descendent machines in the family tree in this thread, I'm
beginning to wonder if I shouldn't go after an Octane or an O2 rather
than an Indigo or an Indy, at least if I wanted to do more than take
the demo programs for a spin.
So for the more experienced SGI folks - do you feel that an Indy or
Indigo is responsive enough to be reasonable to use, or is it worth
holding out for something newer and most likely more expensive to
acquire? Also, something I don't know much about, do the older
personal graphic workstations use odd or impossible-to-find memory, or
are they easy/cheap to load up?
-ethan
7:13 a.m.
Some thoughts:
I have an Indigo2, O2, and an Octane. (Well, and an RE2 Onyx) The only one I
use with any frequency is the Indigo2. Specs-wise, it's equivalent to the
Octane, 200mhz R10k cpu, Solid Impact graphics. However, the Octane can be
upgraded much farther and the Indigo2 is "maxed out" (at least CPU-wise).
The Octane (and Octane2) use a crossbar architecture instead of the bus that
is in the Indigo2 series, but I haven't a clue how much real-world
difference that makes. Since my Octane and I2 are very similarly spec'ed,
running them side-by-side I prefer the I2. I can't really detect a speed
difference between them in normal use.
The O2's use a semi-rare memory type unique to the O2. This makes my poor O2
fairly useless with its 64mb ram. External appearance-wise, the O2's are by
far the cutest SGI's, imho. The Indigo2's use standard parity SIMMs
(72-pin). Octanes use a unique 200-pin DIMM, but ram for them is easier to
find via ebay, etc. than it is for an O2.
If I were to imagine starting over in SGI's, I'd probably get an Indigo2 or
an Octane2. An Indigo2 is cheap, but heavy if you're going to ship. They're
basically indestructible. An Octane 2, well equipped, is the best performing
workstation short of the Fuel ($$$), but still very heavy to ship. Octanes
have a higher incidence (at least anecdotally - have no statistics) of
mainboard/frontplane issues, compared to analogous parts in the Indigo2.
I'm not a big fan of the compression connectors used in the Octane/Octane2,
but if you are aware of the handling concerns, they can be dealt with.
I've no real experience with Indy or Indigo (sans 2), but the Indigo2 is
fun. I hacked a pair of stereo lcd glasses to the stereo-driver port on the
Solid Impact. Gotta love BZFlag in 3d.
Mike
On 10/23/07, Ethan Dicks <ethan.dicks at gmail.com> wrote:
On 10/23/07, jvdg at sparcpark.net <jvdg at sparcpark.net> wrote:
Sridhar Ayengar wrote:
> David Griffith wrote:
>>> I think it goes like this, someone please correct me if I'm wrong:
>>>
>>> 1992 Indigo^2
>>> 1993 Indy
More or less. Besides that, the Indigo2 and Indy
represent a split in
the low-end line, both being offspring of the Indigo, with O2
being Indy's
successor and Octane being Indigo2's successor.
I've had some interest, off and on, of fiddling with some SGI
hardware. I really wanted an Indy when they came out, but couldn't
come close to affording it (I ended up buying a SPARC 1 for $800 the
following year, and many, many job opportunities and good things came
from that). I did get to fiddle with an Indigo^2 in the mid-1990s,
during the Virtual Reality bubble, but couldn't afford one of those at
the time, either.
I had the chance to pick up a discarded Indy about 4 years ago, but by
the time I returned with wheels, someone else had carted it away.
Given the descendent machines in the family tree in this thread, I'm
beginning to wonder if I shouldn't go after an Octane or an O2 rather
than an Indigo or an Indy, at least if I wanted to do more than take
the demo programs for a spin.
So for the more experienced SGI folks - do you feel that an Indy or
Indigo is responsive enough to be reasonable to use, or is it worth
holding out for something newer and most likely more expensive to
acquire? Also, something I don't know much about, do the older
personal graphic workstations use odd or impossible-to-find memory, or
are they easy/cheap to load up?
-ethan
7 Nov
7 Nov
8:18 a.m.
Mike Shields wrote:
If I were to imagine starting over in SGI's,
I'd probably get an Indigo2 or
an Octane2. An Indigo2 is cheap, but heavy if you're going to ship. They're
basically indestructible. An Octane 2, well equipped, is the best performing
workstation short of the Fuel ($$$), but still very heavy to ship. Octanes
have a higher incidence (at least anecdotally - have no statistics) of
mainboard/frontplane issues, compared to analogous parts in the Indigo2.
The Fuel isn't *that* expensive. A lot of places are dumping their IRIX
hardware nowadays, since it's no longer supported.
Peace... Sridhar
9:20 a.m.
On Wed, 2007-11-07 at 11:18 -0500, Sridhar Ayengar wrote:
Mike Shields wrote:
And one of those places would be? :-)
If I were to imagine starting over in SGI's,
I'd probably get an Indigo2 or
an Octane2. An Indigo2 is cheap, but heavy if you're going to ship. They're
basically indestructible. An Octane 2, well equipped, is the best performing
workstation short of the Fuel ($$$), but still very heavy to ship. Octanes
have a higher incidence (at least anecdotally - have no statistics) of
mainboard/frontplane issues, compared to analogous parts in the Indigo2.
The Fuel isn't *that* expensive. A lot of places are dumping their IRIX
hardware nowadays, since it's no longer supported.
Peace... Sridhar
9:35 a.m.
David Holland wrote:
And one of those places would be? :-)
None of which I know of personally, but there has been a definite
increase in the amount of recent SGI stuff showing up on ebay.
Peace... Sridhar
If I were to imagine starting over in SGI's,
I'd probably get an Indigo2 or
an Octane2. An Indigo2 is cheap, but heavy if you're going to ship. They're
basically indestructible. An Octane 2, well equipped, is the best performing
workstation short of the Fuel ($$$), but still very heavy to ship. Octanes
have a higher incidence (at least anecdotally - have no statistics) of
mainboard/frontplane issues, compared to analogous parts in the Indigo2.
The Fuel
isn't *that* expensive. A lot of places are dumping their IRIX
hardware nowadays, since it's no longer supported. 
9:53 a.m.
In article <4731F74F.50109 at gmail.com>,
Sridhar Ayengar <ploopster at gmail.com> writes:
None of which I know of personally, but there has been
a definite
increase in the amount of recent SGI stuff showing up on ebay.
dovebid and gov't auction sites, too.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download
<http://www.xmission.com/~legalize/book/download/index.html>
Legalize Adulthood! <http://blogs.xmission.com/legalize/>
9:38 a.m.
In article <4731E549.1090009 at gmail.com>,
Sridhar Ayengar <ploopster at gmail.com> writes:
The Fuel isn't *that* expensive. A lot of places
are dumping their IRIX
hardware nowadays, since it's no longer supported.
Again, prices are all relative... I see them selling on ebay for like
$300 lately.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download
<http://www.xmission.com/~legalize/book/download/index.html>
Legalize Adulthood! <http://blogs.xmission.com/legalize/>
23 Oct
23 Oct
9:26 a.m.
New subject: Indy or Indigo reasonable to use? [Was: SGI workstation sequence]
On 23/10/07 11:02, "Ethan Dicks" <ethan.dicks at gmail.com> wrote:
So for the more experienced SGI folks - do you feel
that an Indy or
Indigo is responsive enough to be reasonable to use, or is it worth
holding out for something newer and most likely more expensive to
acquire? Also, something I don't know much about, do the older
personal graphic workstations use odd or impossible-to-find memory, or
are they easy/cheap to load up?
I think that you'd find them a little sluggish / limited for *real* use.
That said, they're both fantastic computers in their own right. Beautifully
made and very stylish. To my mind, nothing says "90's computing" like an
SGI box and their current highly affordable status surely can't last.
Lawnmower man and VR chic must be due a revival. ;-)
I have a 300Mhz R12000 Octane with 1.5GB of RAM and a couple of 72GB SCSI
disks running IRIX 6.5.22. It isn't quite on-topic being c. 1999 vintage
but it runs the latest Nekoware build of Firefox 2 and with the adblock
extension to remove Flash crapware (IRIX only supports Flash up to version
5) it feels every bit as responsive as, say, a late model PowerBook G4.
Surely not many other computers of that vintage can make such a claim?
-Austin.
P.S. The Octane in question was a freebie - you can get Octanes and O2's for
next-to-nothing.
10:23 a.m.
New subject: Indy or Indigo reasonable to use? [Was: SGI workstation sequence]
In article <C343DF46.21BAD%austin at ozpass.co.uk>,
Austin Pass <austin at ozpass.co.uk> writes:
P.S. The Octane in question was a freebie - you can
get Octanes and O2's for
next-to-nothing.
I suppose that depends on where you are; they get bid up on dovebid
and on ebay.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download
<http://www.xmission.com/~legalize/book/download/index.html>
Legalize Adulthood! <http://blogs.xmission.com/legalize/>
10:48 a.m.
New subject: Indy or Indigo reasonable to use? [Was: SGI workstation sequence]
On 23/10/07 18:23, "Richard" <legalize at xmission.com> wrote:
I suppose that depends on where you are; they get bid
up on dovebid
and on ebay.
I'd be astonished if being UK-based put me at an advantage in the classic
collecting stakes! Whereabouts are you?
For example, there's a really basic Octane for ?35 buy-it-now on ebay.co.uk.
I obtained mine through comp.sys.sgi.marketplace - it's a super-low volume
group but items are available through there every once in a while. Likewise
nekochan.net has many items passing through their "For Sale" forum, of which
at least 80% are US-based. That said, I'm aware that the US is *slightly*
larger than blighty ;-), and SGI build-quality definitely works against them
when one is seeking cheap postage. I made a weekend trip of collecting
Octanes and a Crimson, although having a van at my disposal makes the task
simpler too.
An Octane is well worth the extra couple of dollars though IMHO. A
daily-use classic capable of modern tasks is a joy - plus mine looks ace on
the desk at work next to my MacPro!
-Austin.
11:02 a.m.
New subject: Indy or Indigo reasonable to use? [Was: SGI workstation sequence]
In article <C343F28C.21BB4%austin at ozpass.co.uk>,
Austin Pass <austin at ozpass.co.uk> writes:
On 23/10/07 18:23, "Richard" <legalize at
xmission.com> wrote:
I'm in the states. I suppose it also depends on what you mean by
"cheap". For me, "cheap" means < $50. 35 BP = $75, not what I
would
call "cheap", but I suppose it is by SGI standards.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download
<http://www.xmission.com/~legalize/book/download/index.html>
Legalize Adulthood! <http://blogs.xmission.com/legalize/>
I suppose that depends on where you are; they get
bid up on dovebid
and on ebay.
I'd be astonished if being UK-based put me at an advantage in the classic
collecting stakes! Whereabouts are you? 
12:21 p.m.
On Oct 23, 2007, at 6:02 AM, Ethan Dicks wrote:
So for the more experienced SGI folks - do you feel
that an Indy or
Indigo is responsive enough to be reasonable to use, or is it worth
holding out for something newer and most likely more expensive to
acquire? Also, something I don't know much about, do the older
personal graphic workstations use odd or impossible-to-find memory, or
are they easy/cheap to load up?
Indigo2, yes...Indigo or Indy, no. Just my opinion.
I've used a few Indigo2 systems as my main desktop machines over
the years. They are fantastic in many ways. The last one, about
four years ago, was an Indigo2 R10K MaxImpact...It was wonderful,
although the R10K is a bit constrained in the Indigo2 design. It's
still very responsive and more than fast enough for everyday modern-
style work.
With prices the way they are, though, I'd probably find an Octane
if I were looking to set up an SGI desktop workstation. They are
considerably faster than the R10K Indigo2 systems at the same clock
rate, and expandable to truly juicy configurations. (think dual-
R12K, etc)
-Dave
--
Dave McGuire
Port Charlotte, FL
Farewell Ophelia, 9/22/1991 - 7/25/2007
12:37 p.m.
On Tue, 23 Oct 2007, Dave McGuire wrote:
On Oct 23, 2007, at 6:02 AM, Ethan Dicks wrote:
For a laugh, a few years ago I installed Linux on my R4600 133mHz Indy.
The machine took 8 hours to compile a Linux kernel, 21 hours to compile X
- and 34 hours to compile KDE. <chuckle>
Not much of a benchmark, since Linux isn't as efficient on the MIPS
processors as IRIX, for obvious reasons, but... for what it's worth :)
JP
So for the more experienced SGI folks - do you
feel that an Indy or
Indigo is responsive enough to be reasonable to use, or is it worth
holding out for something newer and most likely more expensive to
acquire? Also, something I don't know much about, do the older
personal graphic workstations use odd or impossible-to-find memory, or
are they easy/cheap to load up?
Indigo2, yes...Indigo or Indy, no. Just my opinion.
1:45 p.m.
On 2007-10-23 21:37, JP Hindin wrote:
For a laugh, a few years ago I installed Linux on my
R4600 133mHz Indy.
The machine took 8 hours to compile a Linux kernel, 21 hours to compile X
- and 34 hours to compile KDE. <chuckle>
Heh. My Mac SE/30 (16 MHz 68030) took two days (trying) to compile a NetBSD
1.5 kernel, before keeling over and dying with a failed hard drive. Not
surprising, really; the HD was flaky to begin with, and it had been
thoroughly thrashed during that compile session, making the poor SE/30 sound
like a coffee maker.
,xtG
.tsooJ
--
Never underestimate the power of stupid people in large groups
--
Joost van de Griek
<http://www.jvdg.net/>
2:21 p.m.
Joost van de Griek wrote:
I once compiled a NetBSD kernel (1.4-something) on a VAXstation 2000. I
don't remember how it took, but it was a long, long time.
Peace... Sridhar
For a laugh, a
few years ago I installed Linux on my R4600 133mHz Indy.
The machine took 8 hours to compile a Linux kernel, 21 hours to compile X
- and 34 hours to compile KDE. <chuckle>
Heh. My Mac SE/30 (16 MHz 68030) took two days (trying) to compile a NetBSD
1.5 kernel, before keeling over and dying with a failed hard drive. Not
surprising, really; the HD was flaky to begin with, and it had been
thoroughly thrashed during that compile session, making the poor SE/30 sound
like a coffee maker.
3:29 p.m.
On 10/24/07, Sridhar Ayengar <ploopster at gmail.com> wrote:
I once compiled a NetBSD kernel (1.4-something) on a
VAXstation 2000. I
don't remember how it took, but it was a long, long time.
Well... I'm not worried about how long an Indy or an Indigo or
anything made in the past 15 years is going to take - my first kernel
compile was on an 11/730... I think it was around 8 hours, possibly
10. Pretty much anything I do these days is going to be faster.
I'm more concerned about interactive responsitivity. Sounds like
unless I get a free Indy (including local pickup), I'll probably keep
my eye out for an Indigo2 or Octane.
Thanks for the advice.
-ethan
4:37 p.m.
Ethan Dicks wrote:
On 10/24/07, Sridhar Ayengar <ploopster at
gmail.com> wrote:
My compile on the VAXstation 2000 was at least a few days. Maybe a week.
Peace... Sridhar
I once compiled a NetBSD kernel (1.4-something)
on a VAXstation 2000. I
don't remember how it took, but it was a long, long time.
Well... I'm not worried about how long an Indy or an Indigo or
anything made in the past 15 years is going to take - my first kernel
compile was on an 11/730... I think it was around 8 hours, possibly
10. Pretty much anything I do these days is going to be faster.
5:40 p.m.
New subject: kernel compile times (was: SGI workstation sequence)
In article <471E85C0.4050703 at gmail.com>,
Sridhar Ayengar <ploopster at gmail.com> writes:
My compile on the VAXstation 2000 was at least a few
days. Maybe a week.
Are we talking *just* the kernel here, or are we talking compiling a
full distribution (all the utility programs, shell programs, X server,
X clients, etc.).
I have a really hard time believing it takes a few days *just* to
compile the kernel. Toss in the huge pile of additional programs
accessible from the shell or the X Window System and now we're talking
believable.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download
<http://www.xmission.com/~legalize/book/download/index.html>
Legalize Adulthood! <http://blogs.xmission.com/legalize/>
6:01 p.m.
New subject: kernel compile times (was: SGI workstation sequence)
On 10/24/07, Richard <legalize at xmission.com> wrote:
In article <471E85C0.4050703 at gmail.com>,
Sridhar Ayengar <ploopster at gmail.com> writes:
Sridhar can answer for the details of his VS2000, but in my case, the
two machines I have the most experience with are the 11/750 and
11/730, and Ultrix 1.1 (4.2BSD under the hood, IIRC). It may be that
I didn't do a complete compile from source, just a kernel rebuild of
drivers and affected areas from the installation process, and it may
be that the disk on a VS2000 is vastly slower than Unibus and IDC
disks on the 11/750 and 11/730, respectively. I do stand by my 8-10
hour kernel compile time after an initial install of the OS on a blank
disk for those machines - if my memories have not entirely
disintegrated, it took all of a long work day to install Ultrix, but
not more than one.
-ethan
My compile on the VAXstation 2000 was at least a
few days. Maybe a week.
Are we talking *just* the kernel here, or are we talking compiling a
full distribution (all the utility programs, shell programs, X server,
X clients, etc.).
I have a really hard time believing it takes a few days *just* to
compile the kernel. Toss in the huge pile of additional programs
accessible from the shell or the X Window System and now we're talking
believable.
6:10 p.m.
New subject: kernel compile times (was: SGI workstation sequence)
In article <f4eb766f0710231801p53c62cd0o434e9ea87338e7c4 at mail.gmail.com>,
"Ethan Dicks" <ethan.dicks at gmail.com> writes:
Sridhar can answer for the details of his VS2000, but
in my case, the
two machines I have the most experience with are the 11/750 and
11/730, and Ultrix 1.1 (4.2BSD under the hood, IIRC). It may be that
I didn't do a complete compile from source, just a kernel rebuild of
drivers and affected areas from the installation process, and it may
be that the disk on a VS2000 is vastly slower than Unibus and IDC
disks on the 11/750 and 11/730, respectively. I do stand by my 8-10
hour kernel compile time after an initial install of the OS on a blank
disk for those machines - if my memories have not entirely
disintegrated, it took all of a long work day to install Ultrix, but
not more than one.
I thought those unix kernels were supposed to be little itty bitty
simply pieces of software. I had no idea they were so bloated.
As for the X Window System, I've compiled the whole thing and worked
on the server itself. I know that's a huge honkin pile of stuff and
it woudln't surprise me that compiling any distribution after X11R4
from scratch would take you all day. In 1990 I was
regularly
compiling chunks of the X Window System, usually the server, on a MIPS
R3000 CPU @ 25 MHz. It wasn't horrible, but it wasn't blindingly fast
either. There's simply a huge pile of code in X11, or even just the
server.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download
<http://www.xmission.com/~legalize/book/download/index.html>
Legalize Adulthood! <http://blogs.xmission.com/legalize/>
6:21 p.m.
New subject: kernel compile times (was: SGI workstation sequence)
On 10/24/07, Richard <legalize at xmission.com> wrote:
In article <f4eb766f0710231801p53c62cd0o434e9ea87338e7c4 at mail.gmail.com>,
"Ethan Dicks" <ethan.dicks at gmail.com> writes:
Bloated? For the day, we certainly thought so, but remember, this was
the era of monolithic kernels. All the devices you wanted to use were
compiled in and loaded at boot time. When you fiddled with your
hardware, the final step was to re-do a kernel config and re-compile.
For most changes, you could get away with compiling the affected
drivers and re-linking, but that first compile on install took hours.
I ran Ultrix 1.1 on a 5MB 11/730 and an 8MB 11/750. ISTR the kernel
was between 1MB and 2MB, but that's a fuzzy memory.
Sridhar can answer for the details of his VS2000,
but in my case...
it took all of a long work day to install Ultrix, but not more than one.
I thought those unix kernels were supposed to be little itty bitty
simply pieces of software. I had no idea they were so bloated. As for the X Window System, I've compiled the
whole thing and worked
on the server itself. I know that's a huge honkin pile of stuff and
it woudln't surprise me that compiling any distribution after X11R4
from scratch would take you all day.
Perhaps Sridhar was doing that on his VS2000, but on a headless
VAX-11, we never installed the workstation packages.
I didn't get into the "modern" UNIX era until 1994 when I got that
first $800 SPARC 1 I mentioned the other day (SunOS 4 / Solaris 1.1.1,
etc.) For the first 9 years, I was dumb-terminals only, except for
the few workstations I got to play with at DECUS conventions.
-ethan
6:59 p.m.
New subject: kernel compile times
Richard wrote:
Nope. It wasn't the full tree. I wouldn't even try compiling the
entire tree on that machine. Admittedly, the machine was under a small
amount of other load at the time, but it was just the kernel. Part of
it was because it was swapping like crazy during the compile, and
swapping to the same disk where the sources were stored. That machine
is S*L*O*W. It would have probably gone a whole lot faster if it had
enough RAM.
Peace... Sridhar
My compile on
the VAXstation 2000 was at least a few days. Maybe a week.
Are we talking *just* the kernel here, or are we talking compiling a
full distribution (all the utility programs, shell programs, X server,
X clients, etc.).
I have a really hard time believing it takes a few days *just* to
compile the kernel. Toss in the huge pile of additional programs
accessible from the shell or the X Window System and now we're talking
believable.
7:19 p.m.
New subject: kernel compile times
On 10/24/07, Sridhar Ayengar <ploopster at gmail.com> wrote:
Oh, yeah. The CPU is nominally as fast as a MicroVAX II (and over 50%
faster than an 11/750), but if you were swapping, that'd be a *huge*
difference in compile times. I was doing this on an 8MB 11/750 with
nothing else going on (or on a 5MB 11/730 in a similar state), which
is, BTW, fully loaded, thus the difference.
I haven't had the pleasure to use a fully-loaded VS2000 - mine are
around 6MB - enough to boot and run, but not a lot of empty RAM
sitting around.
I think a VS2000 disk vastly underperforms compared to an RA81 on a
UDA50, so the swapping makes it much, much worse.
-ethan
> My compile
on the VAXstation 2000 was at least a few days. Maybe a week.
...it was swapping like crazy during the compile, and
swapping to the same disk where the sources were stored. That machine
is S*L*O*W. It would have probably gone a whole lot faster if it had
enough RAM.
24 Oct
24 Oct
5:25 a.m.
New subject: kernel compile times
In article <f4eb766f0710231919g381c908cjb76a7a2ebabeeb35 at mail.gmail.com>,
"Ethan Dicks" <ethan.dicks at gmail.com> writes:
This reminds me of when I was fresh out of college and was working on
a consulting contract. They were writing 6809 assembly code using a
cross-compiler than ran on something like a PDP-11/34. The project
engineer insisted that using a linker was slower than just
reassembling all 100,000 lines (I kid you not). It would take all day
just to assemble the code so you could test it the next day. I
modified everything to declare symbols and use a linker. That got the
assembly process down to a few minutes and the link process down to a
few hours. You could get 4 builds done in a day. The guy's ego was
so fragile he had a shit fit because I made things 4x more productive!
--
"The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download
<http://www.xmission.com/~legalize/book/download/index.html>
Legalize Adulthood! <http://blogs.xmission.com/legalize/>
On 10/24/07, Sridhar Ayengar <ploopster at
gmail.com> wrote:
> >> My compile on the VAXstation 2000 was at least a few days. Maybe a week
.
...it was swapping like crazy during the compile, and
swapping to the same disk where the sources were stored. That machine
is S*L*O*W. It would have probably gone a whole lot faster if it had
enough RAM.
I think a VS2000 disk vastly underperforms compared to an RA81 on a
UDA50, so the swapping makes it much, much worse. 
1 Nov
1 Nov
2:07 a.m.
On Tue, 23 Oct 2007 19:37:36 -0400
Sridhar Ayengar <ploopster at gmail.com> wrote:
My compile on the VAXstation 2000 was at least a few
days. Maybe a
week.
See TOP 9:
http://vaxarchive.pimpworks.org/swdist/bsd/bdc/vaxgeektop10.html
--
tsch??,
Jochen
Homepage: http://www.unixag-kl.fh-kl.de/~jkunz/
2:35 a.m.
On Tue, 23 Oct 2007 23:02:02 +1300
"Ethan Dicks" <ethan.dicks at gmail.com> wrote:
So for the more experienced SGI folks - do you feel
that an Indy or
Indigo is responsive enough to be reasonable to use, or is it worth
holding out for something newer and most likely more expensive to
acquire?
At least here in .de prices are down to around 300,- EUR for a 600 MHz
Fuel or 100,- EUR for a single R12k CPU Octane with (E)SI GFX. If I
where in the business for spendig money on a SGI workstation I would go
for a Fuel. The Fuel allways comes with VPro and the VPro does textures
in hardware. On the older (E)SI GFX you need special TRAM for textures
and this TRAM is rare, thus expensive.
An Octane can get expensive if you want to upgrade it to the level of a
Fuel. I.e. 500..600 MHz and VPro GFX. Price will explode if you want
dual 500..600 MHz CPU on the Octane. Dual 300 MHz R12k is affordable,
but for desktop use a fast single CPU system is more effective then dual
CPU of half the speed. (At least for me.)
A R10k Indigo2 is still usefull. But the Indigo2 is a bit limited in IO
and memory bandwith. An R10k Octane will outperform it.
Conclusion:
If you can spend money: Fuel.
If you can't: mid-range Octane.
Beware:
I don't know about the Fuel, but the Octane as loud as a vacuum cleaner!
--
tsch??,
Jochen
Homepage: http://www.unixag-kl.fh-kl.de/~jkunz/
5:21 a.m.
On Thu, 2007-11-01 at 10:35 +0100, Jochen Kunz wrote:
On Tue, 23 Oct 2007 23:02:02 +1300
<snip>
in hardware. On the older (E)SI GFX you need special
TRAM for textures
and this TRAM is rare, thus expensive.
I think ESI likes to cook itself, and let the magic smoke escape. I've
got ESI in my Octane, and it runs _HOT_, with what appears to be only
passive cooling.
David
3 Nov
3 Nov
1:41 a.m.
On Thu, 01 Nov 2007 08:21:06 -0400
David Holland <dholland at woh.rr.com> wrote:
I think ESI likes to cook itself, and let the magic
smoke escape.
I've got ESI in my Octane, and it runs _HOT_, with what appears to be
only passive cooling.
I have ESSI + ESI (dual head) in my Octane and it works well.
Make sure
that there is no dust in the air inlets. IIRC there is also a PROM
setting for fan speed...
It is quite common for SGI GFX to run hot. The SolidIMPACT in my Indigo2
R10k runs almost hot enough top burn your fingers if you touch the heat
sinks.
--
tsch??,
Jochen
Homepage: http://www.unixag-kl.fh-kl.de/~jkunz/
6037
days inactive
6052
days old
28 comments
10 participants
participants (10)
-
austin＠ozpass.co.uk -
dholland＠woh.rr.com -
ethan.dicks＠gmail.com -
jkunz＠unixag-kl.fh-kl.de -
jplist2007＠kiwigeek.com -
jvdg＠sparcpark.net -
legalize＠xmission.com -
mcguire＠neurotica.com -
ploopster＠gmail.com -
shieldsm＠gmail.com