The collected works of bri3d - Page 4

SGI even sold remarketed Tezro prototypes directly for a very brief time, if I recall correctly.

I'm not sure about Indigo/Indigo2 prototypes - there were several rebrands/case styles for both that I've seen confused as prototypes though.

I've heard of Octanes/Octane2s with "prototype" labels but I'm not sure a different (clear?) case was ever made for them.

http://en.wikipedia.org/wiki/Wikipedia: ... ion_to_SVG

At least some grouping information is preserved - just watching one render in any standard PDF reader reveals that the shapes are z-ordered (back-to-front) so that the final image appears properly.

Since the diagram is of a 3D object, the order is a form of grouping - this is what I was trying to say

It can be assumed that shapes of a similar order can be grouped as they are located "near" each other in a 3D representation of the system - thus, simply by determining the group boundaries one could approximately separate the image into layers. It'd still be quite the task but I believe this approach would be beneficial over manually grouping everything via click-selection or whatever other method.

This will only work in a 68k Mac and probably with ProTools 4. The breakout box you want is called a "441 I/O," if I recall correctly.

There were other components in a real Pro Tools setup but I have no idea if they're required (I've seen a SCSI accelerator, a DSP board, and some others).

RAID-Z sucks. In terms of speed and reliability it's not really close to RAID-6 on dedicated controllers, especially as the number of drives grows.

I just picked one up (I had a software idea and couldn't resist - so far development is going quite well with the exception of the hackiness of my scheme to avoid paying Apple another $99 for a developer account). I just grabbed the $499 WiFi model - $629 + $20/month is definitely a lot more than I can spend on an iPad.

I must say it grows on me with every second I use it. I'm still not enamored with it, nor do I think it's a "magical revolution" (more an obvious evolution) or something everyone should have. But I've spent $500 in a lot of stupider ways, and it is without a doubt the best "laying around in bed checking out the web / watching movies / listening to music / reading books" device I've ever owned or seen.

The biggest minuses so far have been the packaging and what was (or rather wasn't) included, and its appeal to everyone around me. I was sorely disappointed by the packaging - my 2010 Macbook Pro had a lot less wasted space in the box and came with at least a cleaning cloth. Analysts have pegged the per-device margin for Apple at over $200 even for the lower-end iPads - at that rate, was $5 for a cleaning cloth and stand or something really too much? I know Apple love selling overpriced accessories, but there's a balance between what you force people to buy and making your consumer feel shafted when they unpack their new device, and I think Apple swung the balance a little too minimalist with the iPad.
And the iPad is *way* too attractive to people around me (even ones who have already used one, own one, or have decided never to buy one) - it's proven near-impossible to get anything done with it myself so far as I've been letting all my friends take it for a spin instead.

At any rate, my buyer's remorse is wearing off as my code gets better (I'm hoping to make at least the cost of the iPad back) so all is well

re: reliability: http://www.mail-archive.com/zfs-discuss ... 28907.html

Of course the Thumper docs don't say anything about hardware RAID, because they were designed, implemented, and executed to run only RAID-Z. I don't think you could get enough ports off of any commercially available low-profile SATA RAID controllers to run hardware RAID on one (the disks are SATA, by the way, not SAS).

Sun have been switching between LSI and Marvell parts for a while (leaning towards Marvell lately). The X4500 is Marvell while the X4540 is LSI, if I recall correctly.

how about an automatic Fuel and Tezro

The colors, man, the colors! It's all about the colors. Or, for SkyWriter's benefit, should I say the colors !

iSSH isn't perfect but it is really useful.

If you wanted a serial console, it'd probably be easier to use the dock serial port APIs, as Bluetooth is too locked down and your app wouldn't get approved.

I've used either Keynote or prezi.com for my last few presentations. Keynote can do that classic, super lame "PowerPoint" presentation that everyone's looking for without being too much of a royal pain (it's way easier to place text/images/charts properly in Keynote, in my experience, than it is in Powerpoint). Prezi is awesome in that it delivers similar information to a slide presentation, but has a neat format (zooming sheet) that makes you think about how the information you're presenting fits together. Plus their online creation and viewing tools have worked pretty well, in my experience. Even when I can't use it for an actual presentation, I'll sometimes use it just to brainstorm.

Another possible scenario is that something in the Qube was shorted to the Qube's case and the Qube's case was not properly grounded. When you plugged the serial cable from the I2 to the Qube, the I2's case (thru the serial port border/shield) provided a ground path for the current, but whatever was shorted to the Qube's case fried.

I've seen this in a few PCs from time to time - someone will plug in a peripheral and the peripheral and/or PC will release magic smoke. People instinctively blame the peripheral or bus controller but it's often times the case grounding combined with a short.

I've also seen somewhat similar behavior out of ground differentials (this is why you *always* tie two Origin racks together with something metal before trying to power them up) - if both systems are grounded separately there can be a potential difference of up to a few V between the grounds, and the differential can manifest in a lot of ways (current flowing through cable shields when attaching the systems together, current ending up on random data pins and frying things, etc.). Based on how quickly and dramatically you describe the issue as happening I doubt this is it, though, especially if you had them plugged into the same circuits (or even different circuits in a house). It's rare to see too much ground potential difference in common household wiring as the grounds aren't usually isolated in any way.

The evo is getting panned for poor battery life - considering that I'm already disappointed in my Nexus One for that reason I wouldn't get one. It's not "Android" as a platform that causes poor battery life like a lot of dumb tech bloggers seem to think though - it's more advanced, faster chipsets found in new Android phones combined with some iffy kernel engineering on some devices. Once the device has no wake locks and goes into power collapse mode (which most phones spend most of their time in), the kernel combined with the hardware is responsible for battery life, not any Android bits. My HTC Vogue gets several days out of a charge running Android because the hardware is old and the community developed kernel has matured well. Newer Snapdragon phones fare less well because the radio on Snapdragon is power hungry and Qualcomm enjoy shipping half-baked kernels (plus these phones tend to have either huge or OLED screens, both which suck power).
We'll see how the iPhone 4 fares in this regard - the iPad has impressive run time but also weighs a ton and is mostly battery internally. On the other hand, Apple tend to excel at battery life on the software side lately, even when getting drivers from 3rd parties (take a look at how well OS X does on a recent MacBook pro).
And hopefully Qualcomm and all the big phone integrators get their act together and start focusing on battery life rather than raw speed and ship dates on the Android side.

That was probably one of the dumbest and most overblown "hacks" ever, and I don't know why the FBI are even bothering with it. All that happened is some idiot ex-Bantown kids found a servlet AT+T forgot to require auth on that maps IMEI to email, and proceeded to write a script to make GET requests against the entire iPad IMEI range. Then, being idiot ex-Bantown kids, they went and spammed it everywhere and drummed up as much drama as possible, even sending weev and co. on the interview circuit. It exposes essentially no information and exposes users to exactly zero danger

You'd have to fool the system's environmental management into thinking your fans are connected and working - otherwise, it should be just like water-cooling any other computer system (replace the existing heat sinks with water blocks, build in a radiator and pump, check for leaks, go). Fabricating entirely new water blocks that are safe and watertight is usually not easy, but I suspect you could just take a water block for something else and make it fit without too much work.

There's not much "trickery" to water-cooling - it really is a very physical, "what you see is what you get" pursuit, so if you can make everything fit together and not leak all over it'll certainly work.

A Tezro is a mighty expensive system (still :/ ) to be playing around with this sort of thing on though - if you've never done a watercooling project before it might not be the place to start (in case you accidentally introduce a leak, etc).

I would pursue the newest, quietest fans before I went "all out" and built a water-cooling rig.

Nice SGI marketing material

This reminds me of quite possibly the best post ever on a driving forum:

"Why would I use turn signals? You have to take the Americans by surprise!"

SheepShaver is definitely the best solution for PowerPC at this point - PearPC was designed and implemented to run PowerPC OSX, not Classic - and once the Intel Mac came out, it became less relevant (most of its users and developers just wanted OSX on a PC). Plus one of their main developers died in an accident a few years ago and the project has been nearly stagnant since.
I don't think Apple are too concerned about support commitments as their revenue is mostly from hardware and the consumer market, not software and the enterprise market like Microsoft. Plus Microsoft's release schedule is much slower, so each release is supported longer because its successor comes out later.
At any rate, if you're using Classic applications on newer hardware for anything serious at this point you're crazy and should stop, and if you're just using it as a hobby you might as well just keep some of the hardware it was designed for (old PPC Macs) around and run OS9.

Yeah - this is just a bunch of Atom systems attached via PCIe rather than via some other interconnect. SiCortex was much more interesting in that the interconnect was a Kautz graph and it shipped with MPI libraries (in addition to being MIPS), and Tilera is interesting in that they're doing a MIPS-on-microcode approach over a VLIW CPU, similar in some respects to what Transmeta did for x86 (although hopefully less broken).

I think it's the latter, and I think it's mostly because they had OS/2 installed and shipped on a vast number of internal systems and management systems for other machines they wanted to continue collecting licenses on (mainframes).

On the other hand Apple nowadays make money hand-over-fist doing something that is completely not server or enterprise related (iPhone, iPod, iPad), so I think they really couldn't care less about that market right now.

I think there were a couple third-party apps, SoftRAID being the one that springs to mind first. I wouldn't bother.

Hardware support. With the source to IRIX and Xsgi, a single talented individual could easily expand and improve the support for SGI systems in [Linux|NetBSD|OpenBSD|whatever open-source OS of choice]. There's still a lot about IRIX systems (especially the most recent ones) that nobody's bothered to reverse (or that was too difficult to reverse). Plus it'd be an excellent learning experience - things like the bringup process on ODYSSEY (which currently is understood at a generic level such as "this brings up RAM somehow, we copied it from the assembly dump of PROM") would no longer be a mystery.

The Dallas battery does not recharge itself. Search the forums for "Dallas battery" or even "indy battery" and you'll find plenty of topics regarding replacing it, either via a new Dallas chip or by dremeling the existing one to install external battery leads.

Kicking SGIs to boot them is usually a bad idea - while it works okay on the Indy, systems like the Octane have fragile compression connectors that are easily unseated, so kicking will do them more harm than good.

That's definitely a set of dials next to the monitor... but where are the buttons and what are they attached to?

Keep in mind that the O350 came out long after the O3200 series - so the option wasn't even available for a lot of the O3200's lifetime. I'm imagining the O3200 was almost entirely superseded in sales by O350 racks (especially considering how many O350 racks we see on the market) once the O350 came out.

You get a huge power density bonus over O300/O350 by using power bays and CX-bricks in large O3x00 configs, but for the 3200 that's not really relevant.

I'm guessing that it was also probably easier to get SGI to sell you an O3200 with high-end options (V-brick, etc.) than to sell you an O350 configuration with similar options (InfinitePerformance), but as I've never been an SGI salesperson I'm not sure about that one (it just seems to me that that'd be smart business on SGI's part - mark people up to the bigger iron whenever possible).

There are two different basic kinds of "Onyx2/O2k" chassis - The Onyx2 Deskside (has the Infinite Reality plane built into the enclosure, takes IR boards in the same chassis as compute nodes) and the Onyx2 Rack/Origin 2000 (takes only compute nodes, needs an XIO <->KTown board to connect to a seperate G-Brick for graphics).
And beyond these, there are several (tons) of different skins (rackmount, deskside, new-logo, old-logo, Cray-branded, weird flat late-model Onyx2 front panels) - should probably specify if you're picky on those as well (or do you even want any)?

I have an Origin2000 chassis I might be willing to part with but it's only plausible if you're nearby.

I would agree with the third machine assessment - at least on Linux, it's impossible to really tell if your TCP checksumming is correct locally because checksumming can be offloaded into the driver/card if they support it. Try firing up Ethereal/Wireshark on a Linux box with any of the Intel PRO networking adapters and you'll see what I mean - it'll say all locally sent packets have incorrect checksums even when they don't on the wire.
So get a third system in the mix and you can see exactly what the two machines are sending each other over the wire, rather than what they *think* they're sending each other from userspace.

Nice, I'll bite, because this one is obvious.

What makes OS X a "modern platform" compared to IRIX is active development and support for hardware sold after 2006. The technical merits of any OS are up for debate - but the fact that IRIX doesn't run on any new hardware and probably never will return makes it no longer a "modern platform."

And that's why you'd recreate something on a "modern platform" - to take advantage of the speed and power of new hardware. Plus, while you were at it "recreating," the software would probably emerge from the "1870s" you speak of with modern features as well, possibly by accident (for example support for more modern container, compression, and interchange formats).

At any rate, Matador is an awesome app that I'd love to see stick around in some form or another.

viewtopic.php?f=4&t=16723644

The prices are pretty... serious, though.

Canyon Runner looks many kinds of awesome... it's a shame the PC so handily trounced video arcades (at least in the US) because that kind of multiplayer gaming with unique controls isn't really viable for PC games. I shudder to think how much that must have cost, though - an Onyx RE2 *and* all of those custom seats/displays/controls...
It's impressive that the Onyx can drive all six at once - that's certainly one of the coolest things I've seen an Onyx do yet!

If you read through the Magic Edge link syberfreq provided, there's mention of a couple of other insanely-priced Onyx-based games as well.

No "Format: Windows" on the 3G means it's got the HFS+ Mac firmware on it - you were right on that.

What you need to do is plug it into a Windows PC with iTunes and run a "Restore" to get the FAT-aware firmware. If it's not recognized by iTunes, there's something deeper wrong with it.

I'm not sure it's possible to get the Mac iTunes to restore a FAT firmware or vice versa.

If you want 3D (which I'm assuming you do, because you say "modern") the answer is none, at least on the Linux side. Both NVidia and AMD/ATI cancelled their public Linux IA-64 drivers 3-5 years ago. Plus, I'm pretty sure that the only supported cards were PCI-E (FireGL X1/Z1 for ATI, some Quadro stuff for NVidia) anyway.

Windows I'm not sure about - anyone else more up to date?

It's pretty much a lost cause - the closest you'll get is the AGP FireGL X1/Z1 but the drivers for them won't work on any new Linux, since they're years-old now (and never actually worked on anything besides Red Hat to start with). The last public NVidia Linux driver is the same way - I believe the highest card they supported was the AGP Quadro FX line, but there are special stability issues (probably bus-related, although I don't know) with Quadro FX cards in HP workstations so most people had to fall back to lower-end cards.

While I'm talking R300, though, the open-source R300 drivers will work on a recent Linux distribution and provide 3D to a very, very, very limited (slow and buggy) extent. I would assume they would work with a PCI card although you'd probably have to goof around with KMS and early-stage boot stuff to get the system up (as an off-the-shelf PCI R300 won't have an EFI BIOS).

Most people nowadays run their Linux IA-64s headless, judging from the Linux-IA64 mailing list. Proprietary video drivers strike again!

Using a Windows-formatted iPod on a Mac does indeed work (or at least it did with older versions of iTunes). However, what I was saying is that I'm not sure it's possible to use a Mac to install the FAT firmware on an iPod ("convert a Mac iPod to Windows"), as all the iPod Firmware Updaters for Mac OS X are the Mac firmware (obviously). There might be a workaround with some trickery or iTunes might have changed since last time I used it but that's been my experience.

At any rate, your Windows PC should pick up on the iPod regardless of which format it's currently in, and should offer you the option to convert it to a Windows iPod if it's currently a Mac iPod. The fact that it's not suggests that either the iPod is broken in some way, you're having typical Windows driver issues, or you just don't have the right software.

I'm not sure how to do it with the latest iTunes stack, but in the past it was necessary to install a package called "iPod for Windows Software" and then run the "Updater" application which it installed.