Technology 23

Murphy’s Law: Murphy was an optimist

Well, I have to hand it to you, Murphy. I really wasn’t expecting it this time. But you managed to do it. I thought I was free, but clearly I was not.

It’s 5:44am, the birds are starting to chirp, and here I am, finally with a working computer. I am still not entirely sure why it works now. I believe that the clips on the HT800 CPU cooler push down to hard on the CPU or the socket, and are causing instability problems. It doesn’t surprise me. Right now I have the clips off and it boots every time. I put the clips on and it doesn’t boot. I’m pretty sure it’s not coincidental. It’s also the only thing that has changed.

In any event, this has significantly slowed down my work on the summer project, but I’m going to make it up by working all weekend. angry

The question becomes, how am I gonna clip this CPU cooler to the motherboard without causing the instability? I tried only using 2 clips but that seems not to work either. And I can’t even lift up the HT800 since the thermal paste is acting as an adhesive (maybe I did too good a job?). In fact, I’m starting to think perhaps I put too much thermal paste, and that’s why the CPU is getting crushed. I guess things like that really make a difference when you’re talking millimeters.

Throughout all of this, I thought the reason for POST failure was: (a) thermal paste hotspots again; (b) BIOS issues with the installed video card, causing me to do CMOS resets about 40 times with mixed results; (c) bad memory, becuase sometimes taking a DIMM out helped; (d) bad IDE controller, because sometimes removing the drives helped. In all a-d, I believe I was wrong. This was Murphy playing games with me. He gave me four red herrings. The problem, I’m sure, was the clips all along, but Murphy took me for a ride and provided the coincidences to make it possible. All-expense paid trip into my own personal hell: a malfunctioning computer with all my data on it.

I am still not entirely sure of my theory. If Murphy breaks it here, though, I will be quite upset. What a bastard he can be.

Tomorrow I have to do UAC work all day. It’s imperative. Hopefully it’ll keep booting so I don’t have to use the Mac.

SSL and high school reunion

Been reading the SSL Certificates HOWTO to learn a bit for my summer employment (which will require SSL communications). Nice because here my college education is actually helping, in that things I learned in my Discrete Math class are showing up.

Yesterday in the wee hours of the morning, I was getting bored researching this stuff, and so I googled one of my good friends from high school, David Krauss, an extremely talented computer engineer who knew PPC assembler and was fluent in C (and when I say fluent, I mean it) at the tender age of 15. A die-hard Mac lover with so much knowledge, I always wished and hoped he would get a job at Apple. And, it turns out, he did. That makes me happy. The world is somewhat just, it seems.

If you want something done right…

You know the rest.

Well, I got my video card, and definitely needed new holes drilled in my side panels. Unable to find a usable drill press, I decided to bring it to a local glass shop, as I mentioned before.

The guy acted like he could handle it, and said it was completely “feasible,” but that he’d have to “look it over tomorrow.” So I left the panels with him. I called him twice the next morning to see if he still thought it was possible, and to suggest I reduce the drill pattern to make it easier for him. No responses. He disappeared until 5pm or so. When I called him back he said he was “done.” I couldn’t believe it. Could the side panels be finished already? He said they “looked good,” there was just “a little shelling.” (Ok, so I can’t expect something like this).

Holy shit. A little shelling? He ruined my side panels!…

Hardware instability

Well, I’m annoyed now. My hardware’s been unstable since the start of the summer. In WinXP I experience random BSOD on mysterious STOP codes like 0x0000008E and 0x0000007E (which are like page faults and unhandled system thread exceptions). I’m suspecting bad RAM. I pulled out one of my memory cards and left 3DMark03 running in loop, and indeed the computer still crashed. I have the other RAM DIMM out and will do the same kinda test to see what happens. If this crashes as well, the only thing I have left to suspect is this damn AIW 8500DV, which’ll be out of here soon anyway.

I hate stuff like this.

Leadtek Winfast A360 (GeForce FX5700)

So, I decided I’m going to move away from my ATI hardware (an 8500DV AIW) and on to a cheap GeForce card by Leadtek. According to reviews, my 3D performance should increase significantly (from 1000 3DMarks with my current card to ~3500), and this way I should be able to use the Nvidia Linux drivers which are all the rave lately. Frankly I’m tired of ATI’s piss-poor graphics support for Linux.

The other thing is that now I will move my TV capture to a separate card in the PCI slot. A Hauppauge WinTV PVR Media Center Edition (MCE) 250 is my current choice. What’s nice is that then I’ll also get the capture functionality in Linux, which I’ve been missing with my ATI card.

The only thing I fear is that the Leadtek’s fan won’t fit in my Ecube because the AGP slot is near the side panel. But if that happens I’ll RMA to NewEgg.

libranet-archive

About a year ago, I bought a commercial version of Debian from a Canadian company called Libra Systems. It’s called Libranet, and it’s become pretty popular. Basically, it is a 2-CD distro that gets you up-and-running with a Debian sarge system in no time. Pretty soon, Libranet may be obsolete what with the debian-installer efforts and better and better configuration tools being released for free, but I’m still thankful for a commercial Debian distro that gets out of your way and allows you to get up and running with ease, and all the while maintaining 100% compatibility with the official Debian branches. I have since upgraded to sid without a problem, and even run Gnome 2.6 from experimental. All of this upgrading has never been met with problems, despite the custom packages produced by the Libranet maintainers.

In any event, libranet-archive is a project that is not entirely mine. Someone on the libranet mailing list (a great place to get general Debian Linux help) said that he had made archives of the mailing list from February 2001 to December 2003. So I opted to host them on my server. You can view the libranet-archive here. I plan on continuing his effort by turning my mboxes of the libranet mailing list since December 2003 into html using the same tool, MHonArc. For now, though, this may be useful for people searching for answers to the questions.