Ok, so I may try a stress test if I get a chance, but I'd much rather try an older kernel version first. As far as specs, Turion64 dual core, nvidia GeForce go 6150, 2gb ram (of some kind, not sure, I know, it's deplorable), 120 gb hard disk, other than that I'm not sure, unless you're looking for asthetics, which I assume you aren't. Can anyone tell me how to emerge an older kernel? I'm not that good with portage, and I'm not sure how to emerge older versions. I think I knew at one point, but that was a while ago. Thanks.
-Peter On 5/15/07, Wil Reichert <[EMAIL PROTECTED]> wrote:
> Now wait a minute, not everyone has $100k to spend on a brand new > laptop. I'm a student, and I have a single computer to last me through > two years of highschool and and at least a few years of college, and > there's no way I'm going to screw up my computer without some > insurance, ok? Before I run anything on this machine, I'm going to > make sure that I'm still under warrantee, whether the parts are user > servicable or not. Now if you call that being silly, then that's your > choice, but it's my choice if I want to be cautious, even overly so. Keep in mind the programs mentioned are not supposed to break your hardware but to discover if its already got problems. Yes, they do put stress on various components, but thats the entire point - a lot of issues don't show themselves under 'normal' usage. > On that note, I did buck up and run memtest86+ from a Ubuntu livecd, > and after several loops (about 1h 30 min of straight testing) I didn't > get a single error. It was on Test #6 when I stopped, so I think the > memory's chill. Besides, as I said before, when I run anything GUI > (enlightement, right now), it's fine. I just have to jump in and out > of terminal really quickly. The fact that it likes to crash after > starting x server twice makes me think I might have a few damaged > portions on my harddrive. Does that sound about right? Of course, that > sounds like it could be a kernel issue too. If I can figure out how to > "downgrade" my kernel, maybe that will solve it. Try 'badblocks' from a livecd. Its got a read-only mode which will not harm your existing data. This is sounding more and more like a kernel issue. You haven't mentioned the specs on your laptop, but its a recent core 2 model, you'll find its devices are poorly supported in anything less that 2.6.19 / 2.6.20. Wil > > I just clicked the "<<plain text" button and the setting has held for > this entire thread. Come to think of it, I may have actually converted > it back to Rich Text a few weeks back. > > -Peter > > On 5/15/07, Peter Hoff <[EMAIL PROTECTED]> wrote: > > > > > > > > ----- Original Message ---- > > From: Peter Davoust <[EMAIL PROTECTED]> > > To: [email protected] > > Sent: Monday, May 14, 2007 7:11:20 PM > > Subject: Re: [gentoo-amd64] Gentoo crashing? > > > > I know it doesn't actually burn the cpu, but I'd rather not cook any > > components if I don't have to. From what I know of torture tests, they run > > the cpu so hot it starts making computational errors, am I right? It still > > makes me nervous. I was hoping to be able to fix the issue just by > > recompiling my kernel, but no such luck. I'll mess with it some more and see > > what I can do. Can you give me any advice as to what I should to to a) not > > violate my warrantee and b) avoid killing my computer as much as possible? > > Could it just be something with my Gentoo install? I guess that's a stupid > > question; I've had this problem on an older computer, but it was a Desktop > > and it was much easier to swap components without messing up my warrantee. > > So if it were a hardware problem, wouldn't you think that suse 10.2 would > > have run into it as well? I used to run 10.2 (used to as in 3 days ago) for > > hours on end without any problems at all. I agree that Gentoo can run the > > computer harder, but that doesn't quite click. > > > > -Peter > > > > > > > > You're being silly. Software torture tests are not going to kill your > > hardware. Just run them and see what you get. Memtest will give you the > > address where the error occured, and I've always been able to determine > > which stick was bad from that, using a little deductive reasoning (I usually > > verify by testing the sticks alone, but so far I've not been wrong). > > > > As for voiding your warranty, memory and the hard drive are typically > > considered user-servicable parts. In fact, most of the time if either of > > those are the problem they'll just send you the parts and you'll have to > > replace them yourself anyway. > > > > More on torturing hardware: really, the only component that's at all > > vulnerable to this is the hard drive, simply because it's a mechanical > > device, but it will take an absurdly long time to do any actual damage. I > > used to test hard drives for video servers (think Tivo, but starting at > > $100k). We tried a wide variety of drive testing suites, but it turned out > > none of them ran the drives harder than our normal application. A surprising > > number of the oldest version of our product are still running, on the > > original drives, after over 10 years, in situations that are very demanding > > (like serving multiple channels for DirecTV). So, really, stop being so > > paranoid about software torture tests. It is a complete myth that you can > > ruin your hardware by running them. > > > > > > > -- > [EMAIL PROTECTED] mailing list > > -- [EMAIL PROTECTED] mailing list
-- [EMAIL PROTECTED] mailing list
