The Huge Disaster Within The Linux 2.6.35 Kernel

[kraftman]

LTP, sysbench, etc allow you to confirm which subsystem has changed and quantify the performance. You have to run a bucket load of tests (the entire LTP suite, etc) to see the system.

Generally, throughout the industry I have seen *lots* of people who do *really* well at the component level. The system level is a lot harder. So if you find a sensitive proxy, there is no problem in using it.

This makes sense of course. I mentioned sysbench and LTP, because some people started spreading FUD about kernel performance being worse due to entire kernel size, Phoronix article or because it's "bloated" etc. Those mentined tools shows the kernel performance, if it's getting worse, better or stays the same level and PTS measures things mentioned by you and which are not less important for end users.

Again, the tests like apachebench are *NOT* intended to test apache. Since it's running on the same system, it measures the systems ability to do higly concurrent IO and memory access.

If it tests this then it's ok.

Right... My name is kraftman, the avarage computer user reads Phoronix (and is somehow interested in Linux he.. he he...), reads only that something terrible happened in the Linux kernel and suddenly runs to a store to buy Microsoft Windows...

Haha, it wasn't the point! I accidenticaly cut some part: " All this bitching about Linux development model which scares some people is plain stupid and its development model is very succesfull." and while there's short edit window I didn't bother to post a new post. ;>

[V!NCENT]

Haha, it wasn't the point! I accidenticaly cut some part: " All this bitching about Linux development model which scares some people is plain stupid and its development model is very succesfull." and while there's short edit window I didn't bother to post a new post. ;>

*sigh*

"Right... My name is kraftman,[...]" Was supposed to indicate that I was sarcastic as my name crealy isn't kraftman; it is yours.

It was not about me crawling into you skin and saying what you are...

>.<

[squirrl]

closed source

If the system you are testing on is older than that could explain the regressions. The latest systems will most likely be what that kernel is optimized for. The scheduler guy is running an eight core machine. Many of us barely have a two core machine.

[Elyotna]

Maybe the word "disaster" was a bit too dramatic when the article was released, but see : the perfs are still bad, more than a week after the article release.. So I guess that this word is getting more and more meaningful.

[locovaca]

If the system you are testing on is older than that could explain the regressions. The latest systems will most likely be what that kernel is optimized for. The scheduler guy is running an eight core machine. Many of us barely have a two core machine.

That does not make it any less of a regression and makes this story all the more helpful, if it uses hardware the developers do not own nor have access to.

[energyman]

http://marc.info/?l=linux-kernel&m=127534862110443&w=2

and

http://marc.info/?l=linux-kernel&m=127535359014939&w=2

to quote:
> What I really don't get is why they didn't talk to people on
> linux-kernel before posting those claims on phoronix. :S

Well thats kinda obvious, its "journalism" in the days of google
adwords. They make revenue by making people click on their website,
they don't make money being useful or interacting with others.
Phoronix in all the years I've been dealing with them as the only
place doing any reporting on graphics, have never once confirmed a
source, asked for information directly or anything you'd expect from
real journalists, again because that doesn't drive page hits, whereas
sensationalist useless headlines are the main point of the site.

[Chad Page]

A more journalistic approach would've been to share the problem with linux-kernel, and then write up in detail how it got fixed - along the way describing select(), udev, and how some of the kernel mechanisms worked. Enough of those articles and interested regular readers would know much more about the kernel and a couple of new kernel hackers might be on their way up

[Michael]

A more journalistic approach would've been to share the problem with linux-kernel, and then write up in detail how it got fixed - along the way describing select(), udev, and how some of the kernel mechanisms worked. Enough of those articles and interested regular readers would know much more about the kernel and a couple of new kernel hackers might be on their way up

If more readers actually became Premium subscribers or did not use AdBlock, it might actually be possible, but until then it's not as there simply are not the needed resources.

[mirv]

If more readers actually became Premium subscribers or did not use AdBlock, it might actually be possible, but until then it's not as there simply are not the needed resources.

Ahh, the chicken & egg problem.

[cynyr]

If more readers actually became Premium subscribers or did not use AdBlock, it might actually be possible, but until then it's not as there simply are not the needed resources.

Do you get paid for views, or just click throughs? Also does the site have moving ads or text only?

If it's text only views I'll happily turn off ad block here. If it's flashing/moving flash ads, and only click throughs there isn't much difference between me with adblock and me with out is minimal, and more likely to drive me away from the site. Actually, even it was click throughs and it was text only I would make a point to open one in a new tab. Yes the moving flash ads are that annoying. To be fair it's not that I have even looked, I simply block all ads by default. In some respects it is advertisers that are to blame. Targeted text ads for linux/FOSS things would be at least relevant.

The other reason people block ads is that flash ads, are a security risk. There have been many cases of people getting infected by "drive-bys" because of a flash ad, or any other picture based ad. If I recall correctly there was a point in time where Firefox had a jpeg exploit.