Slashdot Mirror
Ubuntu 10.04 Alpha 2 vs. Early Fedora 13 Benchmarks
Given that early benchmarks of the Lucid Lynx were less than encouraging, Phoronix decided to take the latest alpha out for a spin and has set it side-by-side with an early look at Fedora 13. "Overall, there are both positive and negative performance changes for Ubuntu 10.04 LTS Alpha 2 in relation to Ubuntu 9.10. Most of the negative regressions are attributed to the EXT4 file-system losing some of its performance charm. With using a pre-alpha snapshot of Fedora 13 and the benchmark results just being provided for reference purposes, we will hold off on looking into greater detail at this next Red Hat Linux update until it matures."
beta vs beta! Is anyone expecting valid results.?
Catering to niche users at the expense of the majority.
Removing functionality from X. Deleting the ability to restore a feature.
Making it damn near impossible to troubleshoot X crashes.
Ppppppp-p p p ulseaudio
I'm not much enthused by Ubuntu anymore.
I'd go on a Vegan diet but the delivery time from Vega is too long. --brownkitty
or flash objects turn grey and i have to restart firefox
You'll have to ask Adobe about that one. Ubuntu developers cannot trace into software for which they do not have the source code. Or is this happening to you in Gnash?
What often really matters are the upstream apps. Often, other than reporting an upstream bug in an application to the developer, there is not much one can really do about bugs in upstream applications like KDE. I am seeing that now with KDE and X.org. Currently, there is a bug in evdev and dga in X that prevents X from working right with a Wiimote. It can't really be fixed by the distributor. Only X.org can fix it.
So far I have:
Broken Sound effects on Stratagus. (Mandriva 2010.0)
Broken GLX Support on QuakeForge. (Mandriva 2010.0) But DarkPlaces Quake still works.
Broken Wiimote Support in the evdev driver.
These are just a few examples of applications that don't work becaues of a problem upstream.
Seeing as Google is switching to it, I certainly hope so.
By pulling a computer from a dumpster, outfitting it with a $100 hard disk, and installing Linux, I get a giant file server, saving me $200 on an easy backup solution (vs. Apple's Time Capsule). That makes me $200 richer than I would be otherwise, meaning I can use that money elsewhere. With the money I've saved over the years thanks to Linux and other open-source packages, I will soon be taking a Caribbean cruise. Has your "real" Mac ever paid for your vacation?
You do not have a moral or legal right to do absolutely anything you want.
As I understand it, the issue is that the default time between cache dumps to disk is 4 seconds. This is much longer than ext2/3. So, if you yank the power cable during this time, on the next reboot ext4 will have no record of the event ever having occured and will use the previously journaled data instead. If this is actually the case, then I don't really consider this a bug. It's just a larger cacheing window. Someone please correct me if I'm wrong.
There's no 'compiling of the kernel, no conf files to edit' in Debian or any other mainstream distro either. Hell, even Arch Linux does well without custom kernel compilations, and most editing of config files as well, IIRC (depends on usage of course; not that I'd recommend it to any newbie anyway).
The 'it just works' factor isn't something unique for Ubuntu: almost all the others have it as well (LFS an exception). The only thing Ubuntu gives you is a package that will mostly fit the average desktop user in the default install. Pretty much like Mandriva and others. Kernel compilation is not and has not been necessary for more than ten years for any of the mainstream distros.
What the hell is a bad performance improvement?
unless of course it is a windows alpha or beta, then it is blazing fast and feature rich. Sigh....
A computer once beat me at chess, but it was no match for me at kick boxing. Emo Philips
Yes, the inductors and transformers for Macs are hand-rolled on the thighs of virgins.
Shut up! Don't you know Ubuntu is the ONLY linux distribution that does not require to write kernel modules in assembly by yourself ?
I'm glad they finally switched to assembly, I've misplaced my paper card puncher and I've been afraid to reboot ever since !
May contain traces of nut.
Made from the freshest electrons.
If it breaks user's expectations or destroys user's data, no matter how much anyone tries to convince me otherwise, it is a bug.
First off, let me say that I use Ubuntu 9.10 on my system at work. I am also running CentOS on servers, various Ubuntu on servers and a couple of Fedora systems. As you can see, I have experience with all of them.
So why is this review useless? Because they are testing development systems, which are not optimized, have loads of debugging flags set, and essentially are not ready for prime time. Of course it may be running slower!
IMHO, you should ignore benchmarks until the release candidates, at least. I generally ignore benchmarks on unreleased systems. I do, however, like to read and learn about new features which may be present in early releases.
They resolved an issue which lead to file being overwritten being left empty on a crash. The problem was that they were optimising the write order to make performance better. This lead to metadata being updated too early in some cases so you would get a corrupted file. Now the issue has been resolved which lowered the performance although I think there may be an option you could turn on. So if an application is updating the file you will get the old version or the new version (assuming they have written the program in a half decent way) of the file which is good enough. If you want anything better than that you should be running a UPS which should be correctly configured to safely shut the system down (unlike one system I experienced that had a UPS but then everything crashed when the UPS ran out of battery because the sysadmins were appalling).
Consider when a typical application exits. It does something like:
* Write new config file to a temporary file
* Rename the temporary file over the top of the old config file
This way if the computer crashes, applications expect the config file to always be valid. i.e. they expect the data to have been written to disk, completely, before the rename happens.
This worked in ext3 and other filesystems, but originally not in ext4. The result was that the config file could end up being empty.