flat assembler
Message board for the users of flat assembler.

Index > Heap > Google: Computer memory flakier than expected

Goto page Previous  1, 2
Author
Thread Post new topic Reply to topic
kohlrak



Joined: 21 Jul 2006
Posts: 1421
Location: Uncle Sam's Pad
kohlrak
I'm saying two things really.

First, google is probably complaining about problems with ram so they can later blame ram for the problems with their algorithms.

Secondly, the parts breaking down in ram are most likely due to melting or parts that weren't stable to begin with.
Post 20 Oct 2009, 00:31
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger Reply with quote
revolution
When all else fails, read the source


Joined: 24 Aug 2004
Posts: 17247
Location: In your JS exploiting you and your system
revolution
A bad workman (or workwoman) blames his (her) tools.

Is Google a bad workperson?

Why use ECC if one just wants to be lazy and blame memory errors for all the problems?

Using ECC corrects the errors. So that would eliminate memory errors from the possible sources of blame for the software crashes.
Post 20 Oct 2009, 01:07
View user's profile Send private message Visit poster's website Reply with quote
kohlrak



Joined: 21 Jul 2006
Posts: 1421
Location: Uncle Sam's Pad
kohlrak
Quote:
A bad workman (or workwoman) blames his (her) tools.


Because they do not get better tools.

Quote:
Is Google a bad workperson?


I'd argue so.

Quote:
Why use ECC if one just wants to be lazy and blame memory errors for all the problems?


Setup for later. If they immediately argue a conspiracy for slower computers to excuse their algorithms failing to meet their demands, people'll know what they're doing. I'm not saying Google is stupid, just not as bright as we'd like them to be. Let's give a similar example, american politics for instance. Obama blames Bush for all economic problems, so when he spends money we don't have, he didn't do anything wrong, we're just feeling the ripples of Bush's mistakes.

Quote:
Using ECC corrects the errors. So that would eliminate memory errors from the possible sources of blame for the software crashes.


Not software crashes, slowdowns. Make the RAM makers look evil, then they can claim conspiracy. Customers agree with google (as always except in the mystery install issue) and blame hardware for the problems. Not that hardware isn't often at fault, but a programmer shouldn't code their algorithms assuming that the hardware is responsible to make their program work not the programmer.
Post 20 Oct 2009, 22:47
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger Reply with quote
LocoDelAssembly
Your code has a bug


Joined: 06 May 2005
Posts: 4633
Location: Argentina
LocoDelAssembly
kohlrak, using ECC memory is actually putting them in evidence when a crash occur, because if one occurs and it is not an uncorrectable ECC error event, then the probability that it was a software error is much higher. Your argumentation would be valid for non-ECC memory, in that case they could say "Oh no, it is your low quality memory at fault, our software is bullet proof".

PS: Just an anecdotal case in which the memory is actually in fault but is it undetected as such because there is no ECC: http://board.flatassembler.net/topic.php?t=9952
Post 21 Oct 2009, 00:09
View user's profile Send private message Reply with quote
Borsuc



Joined: 29 Dec 2005
Posts: 2466
Location: Bucharest, Romania
Borsuc
Actually that happened to me too and I blamed a bug in .7z for giving me corrupt archives. Then after a week or so I tested the memory with memtest86+ and it was a RAM stick at fault (no ECC=no evidence).

However this was a "hard" error, meaning, it was a faulty component and errors that would be repeatable, not just happening once (ECC doesn't help here much, but it does at least tell you it's broken).

_________________
Previously known as The_Grey_Beast
Post 21 Oct 2009, 15:50
View user's profile Send private message Reply with quote
kohlrak



Joined: 21 Jul 2006
Posts: 1421
Location: Uncle Sam's Pad
kohlrak
Quote:
kohlrak, using ECC memory is actually putting them in evidence when a crash occur, because if one occurs and it is not an uncorrectable ECC error event, then the probability that it was a software error is much higher. Your argumentation would be valid for non-ECC memory, in that case they could say "Oh no, it is your low quality memory at fault, our software is bullet proof".


Truth is, the only ones they'll convince are the misinformed and the people who readily blame RAM because they really don't know what ecc is. We've seen some of that here already... The very title of this post would make one think that RAM is a big bad problem.
Post 21 Oct 2009, 18:46
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger Reply with quote
r22



Joined: 27 Dec 2004
Posts: 805
r22
@Any commenter who hasn't RTF[Research]
(it's not that long, in fact you can skip to the conclusions at the end)

Perhaps ACTUALLY reading Google's research
http://www.cs.toronto.edu/~bianca/papers/sigmetrics09.pdf
would significantly lower the "riddled with false assumptions" rate of your comments on the subject.

Here's a good quote from the research:
Quote:
The conclusion we draw is that error correcting codes are
crucial for reducing the large number of memory errors to
a manageable number of uncorrectable errors. In fact, we
found that platforms with more powerful error codes (chip-
kill versus SECDED) were able to reduce uncorrectable er-
ror rates by a factor of 4–10 over the less powerful codes.
Nonetheless, the remaining incidence of 0.22% per DIMM
per year makes a crash-tolerant application layer indispens-
able for large-scale server farms.
Post 22 Oct 2009, 14:17
View user's profile Send private message AIM Address Yahoo Messenger Reply with quote
Display posts from previous:
Post new topic Reply to topic

Jump to:  
Goto page Previous  1, 2

< Last Thread | Next Thread >
Forum Rules:
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You can attach files in this forum
You can download files in this forum


Copyright © 1999-2020, Tomasz Grysztar.

Powered by rwasa.