Hi,
I just read an article[1] from phoronix about performance comparison on several larger Linux distribution and I'm really surprise to see the result that Slackware performs "bad" compared to the other distro. What do you think guys?

[1] http://www.phoronix.com/vr.php?view=17688

deleted.

I think that if you want to truly understand the results you have to find out which kernel and what user land SW versions are used, which patches were applied (if any) and what the effect of those is on performance.

1 members found this post helpful.

I'd be interested to see benchmarks with standardized kernels. There was a regression present in 3.2.23 as seen here, but I'm not sure if it affected performance or just reporting. These tests also do nothing to determine the possible cause of lacking performance which is unfortunate. The table of information on page 1 is also inconsistent with testing on the same hardware, so I'm not sure what to make of the tests. It should be noted that it is very poor form, and certainly bad journalism, to compare a pre-release version of Slackware with fully released distros, though I'm not sure whether or not there would be a difference now (at least the kernel version has been bumped in Slackware). See this thread.

All I have to say about the benchmarks: they don't matter to me. I have tried multiple distros on my machines and I feel that Slack runs the fastest (and most stable). I know its subjective, but that's my test.

2 members found this post helpful.

Forget the benchmarking results and such comments like "oddly disappointing," because it is really splitting hairs for the most part. The take-away is that slackware is there, and it's a popular distro! Slackware didn't seem so popular a few years ago and would not get mentioned often, but it's popularity has appeared to pick up. More and more people have gone through the undecided phase of trying many distros, but come back to slackware later. The slackware "KISS" approach, keeping slackware-specific scripts and tools simple, allows the user to feel that slackware is their own distro that is easy to reconfigure, fix, and feel confident to use.

I'm suspicious of measuring technology on performance only. Often, gains in performance come by cutting corners on proper sanity/error checking and using other less-safe modes of operation. Maybe slackware compiles with less compiler optimization or assumes a more basic CPU feature set than some other distros? In the end, a little lower performance can sometimes mean better compatibility and stability. This is my opinion, not intended to start a big debate on this.

4 members found this post helpful.

I'm guessing the writer/tester did not read CHANGES_AND_HINTS.TXT and was probably using the huge kernel. Other little post-install actions may also have been left out.

There's a lot of variations in the DE/WMs that was not considered.

Additionally, I could care less how those tests run on an i7 Core proc w/ 8GB RAM. I'd like to see the results on a P4 w/ 1GB RAM because that's basically the ballpark of what I typically use currently. I guarantee Ubuntu, CentOS, and Fedora with Unity/GNOME would not fair quite as well!

1 members found this post helpful.

I think it is just another meaningless test. Not any huge differences that one would see during normal use. It's certainly not going to convince me to switch to a "faster" distribution. Slackware is fast in my book that's all that matters.

Benchmarks are bullshit compared to real usage scenarios.

Slackware may not be the fastest Linux distribution, but it's the most stable distribution you can find outside of Linux From Scratch.

Once you use Slackware, there are no other distributions... besides LFS.

Agreed. Our distro has legendary stability and security; Slackware earned its reputation. There is no substitute.
Slackware!

2 members found this post helpful.

Benchmarks are one test and stability is another test. It's not wrong to test performance.

(Note: I have no idea if these benchmarks were done properly.)

Indeed, I'm totally agree with you

I know this isn't normally done, but this deserves a fitting Meme...

Attachment 10788

3 members found this post helpful.

Wow, what a great idea to benchmark OS in development stage against stable competitors!

It's probable that the version of Slackware used there was hit by some of sort regression in the kernel, and let's not forget the "i486" thing (although it shouldn't make that much of a difference). Still, it doesn't imply anything about stability and those differences are not things that once can notice during normal desktop use anyway.