All the statements below are taken from the Slashdot dicussion at http://www.slashdot.org/articles/00/10/25/1448233.shtml. The postings are just an outline, but - I think - they summerize things up quite good.
Sounds like a good plan..... (Score:1) by soulsteal (...) on Wednesday October
25, @11:46AM EST (#2) (User #104635 Info) http://ocean.otr.usm.edu/~estisdal
Windows kills the memory, Linux nurses it back to working health.... "
Drag queens are only guys who like to eat, drink and be Mary."
Re:Oh, sure, Linux users are this desperate (Score:2) by Ed Avis (...) on
Friday October 27, @05:47AM EST (#319) (User #5917 Info) http://membleded.com
I wouldn't use RAM with intermittent faults. But if it had a handful of known
bad bits, with a guarantee that all the others were solid, I wouldn't have a
problem with mapping out the bad 0.001% (with say 0.1% wasted space) and using
the rest. [...]
Actually, this has real-world applications... (Score:1) by Troy Baer on
Wednesday October 25, @01:05PM EST (#151) (User #1395 Info) http://home.columbus.rr.com/tbaer/
This is very useful for production systems with very large amounts of memory.
For instance, Cray systems have a capability where bad bits in memory can be
"flawed out" on the fly. Extending Linux to support the same kind of thing (especially
in combination with ECC memory!) would very useful for shops that have big memory
requirements and need as many 9s of uptime as they can get. [...]
Nice... now where can I find faulty 64/128MB DIMMS (Score:1) by ndnet on
Wednesday October 25, @12:06PM EST (#50) (User #3243 Info) http://www.ndnet.org/
This is a cool idea to make use of bad hardware, and while it shouldn't be used
to make new systems, it will. It's really nice for poor geeks like me who would
be happy to have more than 64MB. My system is a workstation from Dell, and I
imagine it can handle 128 to 192 MB RAM total. [...]
Better hurry... (Score:5, Funny) by darial on Wednesday October 25, @12:28PM
EST (#90) (User #177051 Info)
I beat feet to my local purveyor of crappy used hardware as soon as I saw this,
and all I have to say is:
handfull of busted 256m DIMMS: $10.71 with tax 6
reboots, a little math, and a partial kernel compile: 21min
The look on my roommate's face when I typed "top": priceless!
Absolutely good. (Score:1) by jabber01 on Wednesday October 25, @11:58AM
EST (#28) (User #225154 Info) http://slashdot.org/comments.pl?sid=jabber
It SURE is... Now Linux is SO GOOD, it can even run well on defective hardware.
Let's see M$ do that one! Now who's the one chasing tail-lights? The only way
for M$ to one-better Linux is to make an OS that runs on NO RAM, and we know
where they stand on bloat.. [...]
Great, deliberate instability :-/ (Score:2) by Nick Driver on Wednesday
October 25, @11:48AM EST (#6) (User #238034 Info)
Now we have a way to deliberately make Linux instable.... if you subscribe to
the theory that if a DIMM has bad areas then that increases the probability
that more of its areas will fail in the future.
Every day I learn something new that NT cannot do and something new that Unix
CAN do!
Predictable faults? (Score:2) by ackthpt (...) on Wednesday October 25,
@02:35PM EST (#209) (User #218170 Info) http://www.dragonswest.com
So tell me, how, without a memory tester, do you know what's predictable vs.
unpredictable? As far as I'm concerned with DIMMs, 5% broken is 100% broken.
Nice... now where can I find faulty 64/128MB DIMMS (Score:1) by ndnet on
Wednesday October 25, @12:06PM EST (#50) (User #3243 Info) http://www.ndnet.org/
[...] My worry is that some local computer makers will use it to screw other
linux geeks. Luckily, most Linux users will notice. I'm mostly worried about
those buying a preinstalled box as their first Linux PC. Here's hoping that
they're safe. [...]
Re:sounds good for high-performance computing shop (Score:1) by dbretton
(...) on Wednesday October 25, @12:42PM EST (#120) (User #242493 Info)
No! I high performance computing shop would never use substandard parts, nevermind
parts that are KNOWN to be BAD. Bad parts, regardless of whether it is somewhat
usable, can corrupt data or cause incorrect results. Therefore, those parts
cannot be trusted. -Dennis Don't read this.
Re:sounds good for high-performance computing shop (Score:1) by ackthpt
(...) on Wednesday October 25, @12:52PM EST (#135) (User #218170 Info) http://www.dragonswest.com
Nobody (except the brainless) leaves defective parts in mission critical hardware.
If I were experimenting with something, I might consider this patch, but for
someone who leaves a server running, and expects stability for months at a time,
it's not within the realm of consideration.
Re:sounds good for high-performance computing shop (Score:1) by ibpooks
(...) on Wednesday October 25, @02:40PM EST (#211) (User #127372 Info)
In a situation like that, it costs far more for a company to pay an IT worker
to swap RAM chips than it does to buy quality, top-of-the-line components the
first time around. Not to mention, mission-critical systems should NEVER use
crappy hardware.
A better solution... (Score:2, Insightful) by Anonymous Coward on Wednesday
October 25, @12:34PM EST (#101)
Buy a real computer. Pay a lot of money for it. Get a warranty/service contract
(they're often free for a certain level of support). Bad RAM? Send it back.
Bad mainboard? Send it back. Bad anything for the first three years? You guessed
it...send it back. Get the replacement by next-day air. Free. [...]
Bad Ram! Bad! (Score:1) by bperkins on Wednesday October 25, @12:34PM EST
(#103) (User #12056 Info) http://www.netspace.org/~bperkins
Hmm. I think the price of bad RAM just went up. I'm not sure I'd use this myself
though. I find myself resurecting machines that won't boot from lilo anymore
quite a bit. It'd be awfully annoying not to be able to use an off the shelf
rescue disk. It's bad enough when you have to get some weird scsi driver working.
OTOH, it'd be a lifesaver if you have some wounded machine you need to get back
up ASAP.