?

Log in

No account? Create an account

Previous Entry | Next Entry


[slash] how many interrupts does a digium card occupy?
[AmishOne] One, but it wants it all to it's self.
[paul] 5: 1659556326 XT-PIC t4xxp
[paul] from /proc/interrupts
[AmishOne] It's better when you put that in context
[paul] 1.6 trillion interrupts
[paul] in 19 days
[slash] ahh, i was under the impression it did some wacky consumption of I/O addresses and IRQ's, being a multiport card and all
[pjustice] Damn, begins to sound like political debates
[paul] nope
[AmishOne] heh
[slash] the days of irq conflicts and setting jumpers still haunt me
[paul] my challenger here, he neglects to tell you how amazing his single IRQ is used. He says the card takes only one IRQ. What he doesn't tell you is that it generates 1.6 trillion, now listen carefully to that, not 1.6 million, 1.6 TRILLION interrupts in 19 days. That's not a very conservative card to me, and I think you'll agree.
[pjustice] "Now see, he hasn't answered the question again. You can't call them interrupts if they didn't suspend some other running task doing actual work. But I have a plan to allow visiting hungry puppy processes to use the bus during off hours, making the CPU load more even. This will amortize the interrupt load over more actual jobs, and put more american transistors back in the workforce."
* AmishOne runs in fear.
[AmishOne] Ow. ow.. ow.. this hurts. ow.
* pjustice emulates a tacacs and spind liberal
[paul] now let me address this here. This liberal use of interrupts is without precident in current systems. We may have a PCI bus that allows overlapping IRQs, but his card lets 80% of the IRQs go to less than 1% of the hardware. That's 80%. It's my plan to bring the interrupt load down and still provide excellent service. I know I can if you continue to allow me to work on the situation.
[slash] Thunderbolt and lightning-very very frightening, me!
[slash] Galileo!
[paul] and he supports a card that can change genders, not just being a CPE, but also being able to emulate a switching system. I want to enact a rule that reaffirms the past that allows a T1 to exist between a switch and a CPE.
[paul] and I'll never allow a call to be prematurely terminated, especially after it's supervised.
[pjustice] It's just a spacecraft in a dense gravity
[paul] my opponent supports call termination after supervision, and I think that's a fundamental difference between us
[pjustice] The JPL has a planet put aside for me
[slash] for me, for me!!!
[AmishOne] I, for one, will, if elected, free the mallocs.
[paul] my opponent says this, but his plan just doesn't add up. we need some of those mallocs.
[AmishOne] my opponent stands against garbage collection. I will impliment public memory sanatation, both freeing the mallocs, and relieveing the developers from memory managment.
[paul] he claims to have a plan to relieve developers. I am telling you I deal with developers every day, and I know it just wont work, it won't work!
[AmishOne] And with that, I abandon this bad idea and welcome blissful sleep.
[paul] lol