The collected works of dclough - Page 1

I would count it a blessing to have even just half of the talent that all of you have shown through your work. These are some of the most incredible CG pieces I've seen yet. And Fiat Lux? My god, how long did that take?

Is anybody here into space/sci-fi graphics? Space vessels, nebulae, stuff like that?

HTTP/rsync mirror operator,
hamei wrote:
dclough wrote:
Is anybody here into space/sci-fi graphics? Space vessels, nebulae, stuff like that?

Just saw Star Wreck : In the Pirkinning ... graphics were good, wonder what they used ?

If I remember correctly they used all commodity x86 junk. Low-spec Celerons, Durons, Athlons, Pentiums, off-the-shelf desktop graphics cards. Even IDE hard drives. :P Only in the middle of their endeavor did they manage to get their hands on some nicer HP workstations, which were still x86.

Poor bastards. Probably could have cut their rendering time in half with some decent kit.

Software-wise they used Lightwave, After Effects, Photoshop and the like.

HTTP/rsync mirror operator, ... re=related (Altix advert [ice])

This makes me sad :( Those are the most awkward bunch of ice skaters I've ever seen. And the graphics effects are absolute shit. Did Rackable sack *all* of SGI's artists during the buyout, or just the good ones?

It's got the quality of an entry-level high school computer graphics class project rendered on a Pentium 3 with a GeForceMX...
In the process of getting my new Octane set up I've gone through two 13w3-VGA adapters. I bought them both on eBay... one I bought because I thought it would work, then I realized the listing said nothing about SGI so I bought another one that said SUN/SGI in the listing, saying it would work with SGI systems.

Well, none of them work. In fact, they're the same damn adapter in the exact same packaging. I tried plugging both in between my Octane and my Acer X193w+ monitor that worked with my O2 (flawlessly might I add), but I can't get any signal. An older, crappier Compaq LCD just says "Input out of range" and doesn't do anything. My AL1913W that works with my O2 and is SOG-capable doesn't work, either.

I've tried stripping out all of the pins on one of them to get it down to the basic R G and B plugs, but no joy. I've also tried looking around on eBay for a seller who *actually* knows what he's talking about and won't just post the same adapter and claim it works. Still no joy.

I know there are a few retailers online like Deepspacecables, but DSC I've heard has kicked the bucket and isn't even responding to any orders, payments or emails. Is there anyone else who actually can give me the right adapter? Is there something I can do to one of my current ones to make it work?
Approximately how much of a surge are you seeing?

My HTTP/rsync mirror at has been bumped up to Primary on the page, but has not yet been added to the wiki. I think this may be why your server is still getting pounded in its final days.
The company I'm going to work for soon (Cerner) is a major user of IBM POWER and AIX, and my job involves a heavy dose of POWER and AIX administration.

Though I'll go through extensive training to acclimate to that environment, I'd like to get a jump start on learning how to work with AIX.

Where should I start? Are there any good resources available to learn AIX concepts without spending thousands on a sandbox system?
Glad to hear everything's alright! For a moment there I was starting to suspect a disk crash, database wipeout, system failure, or all of the above...
nekonoko wrote:
(currently runs me about $3000/yr out of pocket).


Wow. I really, really, really appreciate your dedication!

Just to throw it out there, I have some unused IPs, ample server space, and plenty of bandwidth available - If there's absolutely anything I can do to help you work that number down, I'd be happy to help. Even if it means a Paypal donation button - I'd click it a few times. :)
hamei wrote:
jan-jaap wrote:
Mr Neko, I notice you do your own DNS. How does that work ? Can you just set up a dns server and the other dns servers will pick you up, or did you have to get integrated into the Greater Internet in some way ? I need to better understand dns, since that's the biggest tool they use to mess us up. I can say with some authority that putting a known-good dns server address in your setup doesn't help, they still manage to poison it. Thanks, Cisco.

During the downtime I was doing some snooping on my own to pinpoint the problem, and it looks like both nameservers are run locally by nekonoko on the webserver. So unless they continually update their list of known DNS servers, I wouldn't think they would catch on and poison it for such a small site.

If he were to switch over to one of the free DNS providers like (which is an amazing service, by the way) then I could imagine that various ISPs/state governments would tinker with the records that you're seeing.
The time has come to upgrade my server to handle some beefier workloads, so I hopped over to eBay and picked up a Sun Fire T2000 with 16GB of RAM and the 8-core 1.0Ghz Niagara CPU. Once I slap in four new 73GB SAS drives and load up the best Linux distro of all time (Debian, of course) this new box will replace my current server as the primary Nekoware mirror.

So far, my biggest complaint without actually touching the box is an apparent lack of RAID options beyond 0 and 1. With 4 disks, I'm certainly not going to stripe - I would like to have a RAID5 array at least. However, it looks like the best I can do is use two RAID1 arrays and set them up as LVM PVs to get one large contiguous disk area. Does anyone have a better idea, or know of a way to get RAID5 working on these boxes? I could buy an external controller if necessary.

Also, to avoid the PITA ILOM firmware issues I had with my last Sun box, I've got updated firmware files ready and waiting.

Does anyone here have experience with the Sun/Oracle Niagara-based servers, and if so, do you have any complaints or tips for getting the most out of them?
That's one thing that I was debating (and am now debating once more after reading your post) - which OS to use. I'm in love with apt on Debian, and coming from an existing Debian box it would make the transition very seamless, but at the same time the lack of crypto hardware support for Linux does worry me.

Also I've seen a lot of benchmarks on the T2000 that peg Debian/Ubuntu above Solaris in terms of web server performance, by a pretty sizable amount too. On the other hand, now might be a great time to learn Solaris if it means I get better hardware support in the OS and more configuration options.
Have you been able to try Solaris 11 yet? If I do go the Solaris route I'll probably go 11, but it seems too... eh, "tainted" by Oracle, so 10 might be my fallback option.
It's gorgeous! Absolutely incredible. I want one!

While you've got it on Debian 6, were you looking to run any benchmarks?

Just for fun I'd like to see how this thing would do with UnixBench ( ) and compare it to my weak little dual-socket web server. :lol:
The 32-bit Xeon's score is pretty much what I expected compared to my own. And of course the 8-core Xeon box just blows everything out of the water. :)

Compare to my dual 3.06Ghz Xeon, 2GB RAM, 32-bit webserver:

Code: Select all

BYTE UNIX Benchmarks (Version 5.1.2)

System: atlas: GNU/Linux
OS: GNU/Linux -- 2.6.32-5-686 -- #1 SMP Sun May 6 04:01:19 UTC 2012
Machine: i686 (unknown)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Xeon(TM) CPU 3.06GHz (6130.7 bogomips)
Hyper-Threading, MMX, Physical Address Ext, SYSENTER/SYSEXIT
CPU 1: Intel(R) Xeon(TM) CPU 3.06GHz (6131.9 bogomips)
Hyper-Threading, MMX, Physical Address Ext, SYSENTER/SYSEXIT
CPU 2: Intel(R) Xeon(TM) CPU 3.06GHz (6131.9 bogomips)
Hyper-Threading, MMX, Physical Address Ext, SYSENTER/SYSEXIT
CPU 3: Intel(R) Xeon(TM) CPU 3.06GHz (6131.9 bogomips)
Hyper-Threading, MMX, Physical Address Ext, SYSENTER/SYSEXIT
20:17:20 up 23 days, 10:52,  2 users,  load average: 0.14, 0.03, 0.01; runlevel 2

Benchmark Run: Fri Sep 07 2012 20:17:20 - 20:45:46
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables        6361488.5 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     1379.3 MWIPS (10.6 s, 7 samples)
Execl Throughput                               1551.8 lps   (29.8 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        189908.3 KBps  (30.1 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           66670.3 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        346413.2 KBps  (30.1 s, 2 samples)
Pipe Throughput                              537788.7 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                  52561.2 lps   (10.0 s, 7 samples)
Process Creation                               4687.2 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   3147.4 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                    708.7 lpm   (60.1 s, 2 samples)
System Call Overhead                        1073610.8 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0    6361488.5    545.1
Double-Precision Whetstone                       55.0       1379.3    250.8
Execl Throughput                                 43.0       1551.8    360.9
File Copy 1024 bufsize 2000 maxblocks          3960.0     189908.3    479.6
File Copy 256 bufsize 500 maxblocks            1655.0      66670.3    402.8
File Copy 4096 bufsize 8000 maxblocks          5800.0     346413.2    597.3
Pipe Throughput                               12440.0     537788.7    432.3
Pipe-based Context Switching                   4000.0      52561.2    131.4
Process Creation                                126.0       4687.2    372.0
Shell Scripts (1 concurrent)                     42.4       3147.4    742.3
Shell Scripts (8 concurrent)                      6.0        708.7   1181.2
System Call Overhead                          15000.0    1073610.8    715.7
System Benchmarks Index Score                                         453.4

Benchmark Run: Fri Sep 07 2012 20:45:46 - 21:14:32
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables       12950974.0 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     3534.2 MWIPS (10.5 s, 7 samples)
Execl Throughput                               3290.8 lps   (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        149294.5 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           44277.3 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        355121.2 KBps  (30.2 s, 2 samples)
Pipe Throughput                             1322686.3 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 267227.5 lps   (10.0 s, 7 samples)
Process Creation                              10614.9 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   5665.6 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                    738.6 lpm   (60.2 s, 2 samples)
System Call Overhead                        2157802.7 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   12950974.0   1109.8
Double-Precision Whetstone                       55.0       3534.2    642.6
Execl Throughput                                 43.0       3290.8    765.3
File Copy 1024 bufsize 2000 maxblocks          3960.0     149294.5    377.0
File Copy 256 bufsize 500 maxblocks            1655.0      44277.3    267.5
File Copy 4096 bufsize 8000 maxblocks          5800.0     355121.2    612.3
Pipe Throughput                               12440.0    1322686.3   1063.3
Pipe-based Context Switching                   4000.0     267227.5    668.1
Process Creation                                126.0      10614.9    842.4
Shell Scripts (1 concurrent)                     42.4       5665.6   1336.2
Shell Scripts (8 concurrent)                      6.0        738.6   1230.9
System Call Overhead                          15000.0    2157802.7   1438.5
System Benchmarks Index Score                                         776.1

So from this I can gather that the UnixBench workload obviously wasn't designed with distributed MPI-friendly architectures in mind, but nonetheless it's fun to compare. I'll have to give UnixBench a go when I get my new 8-core UltraSparc T1/16GB RAM box up and running.

How responsive does the Altix system feel compared to working on the newer Core Xeon and the old Netburst?
Solaris 11 is up and running on the T2000. So far it's going well, except for some reason the drive refuses to boot my Solaris 10 installation DVD, so all I could do was use the Solaris 11 Automated Installer to set up RAID1 and load an OS. Solaris 11 is so foreign to me, and Oracle's stupid little command-line utilities (ipadm, svc, etc.) are making my job harder than it has to be, but Google is my friend and together, we shall prevail.

The box threw a fit when I tried to boot the Debian installer straightaway, but a quick firmware update fixed that. Then once I got Debian installed, it would hang during boot-up, and I had no idea what was causing it. Poking around init.d in single user mode proved futile. So I tried Gentoo, which wouldn't even make it past the init stages of the installation CD.

So now I'm patching UnixBench to allow >16 CPUs, and barring any major hiccups, should have some scores posted soon for this temperamental little server.
Turns out the particular server I got either A) hates Linux, B) hates me, or C) is just not a good platform for the things I want to do.

Wouldn't accept the Solaris 10 DVD, benchmarked very slow in Solaris 11 (slower than my existing server), would install Debian but wouldn't boot past the rc.local init phase, wouldn't even boot into the Gentoo installation disc, RAID array kept complaining about geometry, and the ALOM was a royal pain in the ass. :roll:

As plain-jane as HP x86 servers may seem, they're very reliable and predictable machines in my experience. I'll be snatching up an 8-core Core2-based Xeon box very soon and sending this PITA Niagara platform right back out the door.
yetanother**ixuser wrote:
the T's are indeed not made for linux. as a former long time solaris admin, we used the T3/T4 as virt platform. i can not imagine whats the deal running linux on them...when i can not use ldoms i dont need a T. better get some intel box for debian.

Indeed, it sounds like SPARC has become very much a niche platform since the T series. I'll probably just grab a cheap Netra X1 to quench my SPARClust from time to time.

zmttoxics wrote:
Believe it or not, that is something that makes sense these days (as I said in my first post, the T1 series are painfully useless). I just installed a lab of 20 HP servers (new DL120g7s (nodes), a dl320g5p (install server), and a dl365 (vpn)) and I wouldn't change a thing, rock solid env.

If it weren't for the limited memory, slower CPU and now-crowded disk space on my current webserver (a DL360g4) I'd happily stay with it as it's been going strong for a few years now. But the DL360g5s have come down a lot and the newer quad-core Xeons are calling my name. That, and moving from 2 SCSI drives to 6 SAS drives in the same 1RU footprint. :)
zmttoxics wrote:
Alternatively, the Sun X4100M2 and X4150 series can also be had really cheap and come with a full featured iLom vs the license required iLo on the HPs.

Been there with Sun's Fire X series before. What frustrated me most was Oracle's selfishness in locking down all of Sun's old firmware files, so updating to fix ILOM bugs was not happening... also, does HP's ILO require a license? I didn't think it did, because my 360g4 runs ILO1 just fine up to the latest version and I can freely download the latest firmware. Did they change this with ILO2-based systems? Because that will be a big deal breaker for me.
If you're going to hate Apple for that reason alone I also suggest that you hate Nintendo for the Wii.
Apple has a track record of valuing their customers and the customer experience, a hell of a lot more than Microsoft or HP or even SGI. If you feel like you were "treated like dirt" by an Apple employee, you should tell somebody. [email protected] is a good place to start. I've had a few phone calls and some corporate favors thrown my way in the past because I wrote to [email protected] to complain about something.

How did Apple treat you like dirt?
Just got back from the datacenter. The primary mirror has been moved from the old dual Xeon box to a later-generation dual quad-core Xeon E5345 server.

Most importantly, the server now has a full gigabit uplink. Between the new SAS-based storage and the bump to a full gig, there's plenty of bandwidth to go around. Now I just have to find a use for it!

Since this is a new box I had to configure everything from scratch. Please test your connection to the mirror ( over HTTP and rsync, and let me know if you encounter any issues. Enjoy!
diegel wrote:
Do you have any statistics? My mirror is far away from needing a gigabit connection. The utilization of my Origin 300 looks like that:

IRIX64 origin 6.5 07202013 IP35    10/06/12

00:00:05  %usr  %sys %intr  %wio %idle %sbrk  %wfs %wswp %wphy %wgsw %wfif
01:00:03     0     0     0     0   100     0     0     0     0     0     0
02:00:03     0     0     0     0   100     0     0     0     0     0     0
03:00:03     0     0     0     0   100     0     0     0     0     0     0
04:00:03     0     0     0     0   100     0   100     0     0     0     0
05:00:03     0     0     0     0   100     0   100     0     0     0     0
06:00:05     0     0     0     0   100     0     0     0     0     0     0
07:00:03     0     0     0     0   100     0     0     0     0     0     0
08:00:06     0     0     0     0   100     0     0     0     0     0     0
09:00:03     0     0     0     0   100     0   100     0     0     0     0
10:00:03     0     0     0     0   100     0     0     0     0     0     0
11:00:05     0     0     0     0   100     0     0     0     0     0     0
Average      0     0     0     0   100     0   101     0     0     0     0

And the network connection viewed from the switch looks like this:

No utilization statistics except some webserver logs and MRTG. I don't run MRTG on the server interfaces but have access to the central datacenter MRTG. I'll occasionally see brief spikes at 25-30Mbps every now and then, but rarely anything sustained beyond a few minutes. I might set up some more detailed system performance logs when I get a chance.
I want to grab an HPUX or OpenVMS box to have something else to mess around on, but I'm limited on space and power. What's the smallest system you've ever seen running HPUX or OpenVMS?
Holy mother of VAX. That looks... insane. Where's the power? Is this thing part of a larger chassis? It looks like the precursor to a modern blade system, except not... well, bladed. What are all those plugs on the front? What sort of cards go inside this thing? How old is it?
Ask and ye shall receive.

I'm really not looking for anything purpose-driven. I just have limited (read: absolutely no) experience in a VMS environment and figured it would be neat to grab a box to learn on every now and then, hence the small and cheap requirement. :) Nothing I have to devote considerable floorspace or a new circuit breaker to.
ItsMeOnly wrote:
How about XP1000? It's quite powerfull, small footprint workstation.

It looks like a nifty little box. Seems very compact, very utilitarian, and seemed like it would be reasonably inexpensive.

Until I saw this: ... 2c6310d53c


How common are these and what could I expect to pay for one?
Winnili wrote:
Also, what's wrong with the Hobbyist Program? I get my PAKs within 24 hours typically, even during weekends. (Especially since it was in-sourced by HP.)

I second that, I joined DECUSERV and immediately applied for hobbyist IA64 and Alpha licenses and had an email within an hour. Asked for download kits at 11PM Friday and had the instructional email and links by 7AM Saturday. Very speedy!
mia wrote:
How much is the subscrition these days?

SAQ wrote:
If I had enough money and I was looking for potential commercial applications I'd go I64.

My desire for an HP-UX box is still unquenched, and the more I read and look around, the more I want to go the IA64 route for the added benefits of a more modern system, newer software support, and the ability to run HP-UX and OpenVMS on the same box (and as a result, a smaller overall footprint).

Now the next big step is to find an affordable Integrity somewhere!
Seconded. I felt so alone and disconnected today! I wonder why? :)
First you'll need to download and unzip all three of the .22 release archives. Then, throw the contents of all three into the same directory. When Software Manager (or inst) asks for the dist location, just point it to that directory. Then you can continue as you normally would for an installation:

keep *
install standard

P.S. Make sure you apply patch 5086 first!
I have a limited amount of academic exposure to computer architecture and microprocessor design theory, but I'll just mention a few things that came to mind while reading this thread:

First, congrats! The Proliant line is a very reliable and well-built family of otherwise boring x86 servers. I've never had hardware trouble with any of my Proliants, although the ILO software can be a PITA at times.

Second, on a single-threaded program you may or may not see linear improvements in execution time by changing the CPU. There are a number of factors to account for here. A big one is cache size - bigger cache = less clock cycles spent fetching. Floating-point performance also varies between processor families.

I think the best thing you could do at at this point is to parallelize your operations across the entire CPU. Perhaps divide and conquer with your dataset to utilize the unused threads?
System Engineer for the managed hosting wing of a large established software company in the Healthcare IT/EHR industry. I spend my day installing and maintaining proprietary medical applications and performing basic system administration on AIX, HP-UX and RHEL. There's also a little bit of Oracle 10/11g in there, but only when a DBA can't be arsed to check out the problem himself.

Eh, who am I kidding - I spend my day checking e-mail and reading Slashdot.
geo wrote:
BSDero wrote:
Big Blue
IBM :?:

BSDero wrote:
Red Evil
Lucent :?:

no wonder you're working on the USB for our IRIX hehe how is it? sorry wasn't able to do the test coz i disassembled my O2, got too excited for my next project :)

hmm i wonder which one is nice to have in IRIX, USB or BT/BLE?

You're spot-on with the first guess, but I don't think Lucent has an OS or CPU architecture.

The combinations of "red" + "CPU architecture" + "Unix" seems to point to Oracle. They inherited SPARC and Solaris from Sun.
The market will determine whether his way is "the right way" through the financial success of the company. In the meantime though, considering the high level of guidance Steve provided up until the bitter end, I personally think Apple may struggle to adapt successfully to the Cook methodology and will most likely go through a strategy-shift not unlike the Sculley era of the 80's and early 90's.

Not saying that we're going to see (authorized) white-box clones any time soon, but I think some products will be championed while others will fall by the wayside. Of course, that's easy for me to say because it's already happening!
hamei wrote:
dclough wrote:
The market will determine whether his way is "the right way" through the financial success of the company.

Jesus H Keerist, are you guys still gobbling up that pap ? Grow up, would you please ?

Gobbling up what? Basic and obvious principles of economics and business management?

Apple is a publicly traded corporation and as such their goal is to create value for shareholders. If Tim Cook can establish and enforce policies that result in additional value to shareholders, then he is doing "right" by his shareholders.
hamei wrote:
Tim Cook's goal in life is to grab as much money as he can for himself and to hell with the rest of the world. If you don't believe that I have a lovely golden bridge to sell you.

Because you know him and he's personally told you that? Oh, you don't know him, you say? Okay then, come back when you have more to say than just "eek the whole world's out to get me!" And I'm the one who needs to grow up. Heh.
SAQ wrote:
Note that "successful in the market" is not sufficient to define "the right way"

It's not sufficient in most cases, but is necessary in some. My thought on this is, how do we define "the right way" without introducing our own personal agenda for the company? I argue that we can't. In the absence of all other definitions of "the right way", I think it's necessary to fall back on a very basic answer like "Whatever makes us more money."

One person's definition of "the right way" for Apple may be to exit the desktop market and focus on mobility. Another person's definition may be to reenter the server market. Yet another person's definition may be to strengthen technical computing and bolster the desktop line. The OP asked a simple question, "is it the right way?", and I think that begets a simple answer - "It depends."

Also, on the note of engineering, I agree that there is no one right way... but there is a "most correct" way which would be the way that best meets your project's goals. hamei seems to disagree with me that the goal of a company should be to make money, but nonetheless, it's an imperative that sits at the core of modern business so it's actually quite fitting.
skywriter wrote:
Why is the younger generation so attached to theorizing on subjects that have demonstrable practical outcomes?

I could theorize about what side a coin will land on when flipped, but that would be a waste of time because I could just flip one and see. It's instantaneous. What's not instantaneous, however, is making strategic changes to affect the trajectory of a business. Getting an outcome from that takes years. So, us young'ins spending a few hours theorizing is a lot better than spending years testing things out with a "wait and see" attitude.

Plus, sometimes it's fun to play armchair-CEO. I went to business school but didn't go into a management career, so I enjoy opportunities to reflect on things I've learned.

This brings up another important question. Why does it matter what we talk about?
skywriter wrote:
Predictably, you've missed the point.

Your point was? I should assume somebody else is right and I am wrong because they are an "elder"? I should listen blindly to what others say because I've been conscious for a few years less?
skywriter wrote:
When you're wrong? Of course, how else do you learn things you have experience with?

And on whose authority am I "wrong" about anything I've said? Would you care to disprove me?
ClassicHasClass wrote:
Can we get back to how Apple sucks?

I would love to. Some people, on the other hand, would rather derail this discussion and grunt like bitter old men who are fed up with "all them damn young'ins 'n their fancy book smarts."

Anyway, as I was on-track to say earlier, I think Cook may have an uphill battle ahead of him trying to break through the "What Would Steve Do" culture. I don't think he wants to be Steve 2.0, and he's made it obvious that he isn't going to be.