0x7D6 March 0x18:

Yarm was built aeons ago, by a race of three-eyed, three-legged giants who inhabited the twisted surfaces of a volcanic island now sunk beneath the Pacific and spoken of only in whispers by the descendants of the Inca, who bide their time amongst the Polynesians.

Yarm is, in addition to a town in Britain (perhaps in Qwghlm; I am not certain), a computation device. An electronic brain, if you will. He is based around a K6/2-300. One of the ones with a 66*4.5 multiplier, not one of the 100*3 ones.

There was no real reason to upgrade it. Sure, Gimp would be lugubrious working with 300dpi full-page images, but I don't do that often enough to _need_ something faster. And all of Konqueror's lagginess (to say nothing of Mozilla's) could be blamed on loose coding practices.

That being the case, I held off on upgrades until something absolutely dirt cheap dropped on me.

This droppage happened recently. A PIII/933 with 256megs of PC133 RAM. Plus various and sundry minor cards that weren't factors in the decision to obtain.

Selection modus: Because it had MORE SLOTS than the next one up speedwise in the same pile (A Celeron 1100 or 1200 with everything onboard). I like to throw cards into my hardware in copious profusion. (Unfortunately, it has no ISA slots. Another reason for holding off. I couldn't really justify the cost of a new hardware modem. This one comes with a PCI hardware modem, though, which I hadn't known when picking it.)

I've been testplaying Hanako's new game (Since I don't see anything about it on their website, I'll be completely cryptic about it, though). Which testplay is erratic, because I need to wedge Monde off her machine to run GameMaker apps.

So. I need to do some huge compiles anyhow to test my new board, and the Wine website said something about Direct3D-related fixes, if I recalled correctly. So a stack of patches download, apply, compile...

And there's a typo. In the released version of Wine. I change the $ to a space, compile it again. And there's a segmentation fault.

Which, on a second make, isn't.

Buggery. Something is Wrong.

Many more compiles later, I determine that turning on the L2 cache doesn't help, that putting the drive I've been compiling on on a 40-pin IDE cable by itself doesn't help, that there's another spurious typo in which a C has become a G, that turning off the L2 cache doesn't help, and neither does turning on ECC for the L2 cache; that bad sectors fail to show up on non-destructive read-write tests, and that a cdefs.h has become a cdefs.l in a Makefile. Grr. So it is looking to me like something likes to set bit 4. Sometimes. It might be the CPU. Might be the motherboard. Might be my sole PC133 DIMM.

On most recent swap, it appears that that last might well be the problem. *crosses his toes and compiles Wine twice more.


Monde from on 0x7D6 March 0x18:
Wedge ME off ZAX?


CHORONZON would never EVEN allow it, my precioussss. Not whilst I'm in the midst of any Manifestation, or practise thereof, or for. For that matter, *I* would not allow it.

Except, perhaps, for certain specific things, which might necessitate your presence at the box.

Message ID: 'comment-0x7D6_March_0x18-0x0'; spamfilter sequence: {%turing_match%}-{%turing_response%}

m_leprae from on 0x7D6 March 0x19:
You make WINE with your toes crossed?!?

Does that crush the GRAPES faster?

Message ID: 'comment-0x7D6_March_0x19-0x0'; spamfilter sequence: {%turing_match%}-{%turing_response%}


{ Add Comment }