Jump to content
Gowator

Benchmarking

Recommended Posts

OK I been meaning to do this for a while;

If you patched your kernel did it make it any faster

is 64 bit that much faster or slower with 32 bit code

xfs vs ext3 or resiser 4 ?

 

ATI or NVIDIA

nv or nvidia etc....

 

Suggestions for repatable benchmarks

Im interested in codecs ... how long to rip a movie/CD track ...

 

all serious benchmarks considered

 

proposal : settings are collected (I guess a standard tool ) and sets of config files downloaded so test is repeatable ...

 

I guess first thing is a list of tests ....and benchmarks...

Share this post


Link to post
Share on other sites

If you can pick up the Linux Journal Dec 2003, there is a bunch of articles on the Ultimate Linux Box 2003 that describes an attempt to build a 64 bit machine plus other articles on DVD Player Shoot out and Authoring DVD's under Linux.. Includes a whole table of benchmark readings too..

Share this post


Link to post
Share on other sites

it's all hardware and 'your box' dependent anyway. What is good for one could be bad for another. Have fun though! :D

Share this post


Link to post
Share on other sites

Yep I actually have thgat mag I was in England that month its also got MDK 10 on the DVD :D

 

Im really interested doing things like recompiling lame or transcode for athlon/P4 and AMD 64 and compare this with benchmarks on plain 586 and even 386 ...

 

if gcc is good the differences should be huge!

 

but for example Id do that and try it on a 2.4 kernel then a 2.4mm patched kernel etc. and repeat it with 2.6 then repeat again with all kernels optimised for the cpu's

 

we need a lot of folks I guess but much as i know the theory of the x86 mm instruction code Ive not seen benchmarks on what the differences are.

 

The benchmarks in the mag test amd64 2.4 against 586 2.6.... whiuch is partly what got me thinking about this... hey what about amd64 2.6 too.

 

be interesting to test opensource drivers vs closed source (like nforce) as well

Share this post


Link to post
Share on other sites

hmmm,

I think Ill abandon this as a workbench topic since a mod has to aproove all replies its simply too much effort posting to it.

 

Kinda like moving the reasons to use linux to the FAQ's it just kills the thread when normal members cant post or reply.

 

So can someone please delete this ...

?

Share this post


Link to post
Share on other sites

For what it's worth, we are all getting into the habit of handling this forum. Please give it a chance. As far as the hassle, this is a different sort of forum that we have not had before. If your subject has no interest, then I advise waiting a bit. Some people don't even know the forum is here! B)

Share this post


Link to post
Share on other sites
it's all hardware and 'your box' dependent anyway. What is good for one could be bad for another. Have fun though! :D

it seems the locks was a toothing problem :D so lets carry on

bvc,

 

Thats exactly the point :D

 

What I though is is it worth recompiling say LAME if you have an athlon ...does it make 2% or 20% difference ?

 

So far I see very little about this, Gentoo folks all say yeah but what testing I saw it made little difference overall, possibly negative in the one test I saw from here.

 

Then again would you really wanna use a 386 optimised RPM on AMD 64....

Even just common tasks, Ive found free memory influences ripping MP3's massivly if they can read and convert the whole track without having to write to disk.

 

The same for disk... how much tuning can you do. Is xfs really faster for media and large files... and is resierfs4 so fast for small ones.

 

The idea is just to get a set of tests individuals can do and post bvack while hopefully learning something too.

Share this post


Link to post
Share on other sites

i would be interesed in benchmarking RAID.

Basic question is : is it usefull to go for a RAID 0 setup for improved performance ?

What differences in perf between a software raid and a hardware one ?

 

Also gowator, i would like you to make a default install of mdk10.0 32 bits and 64 bits (in the same partition on the same disk) and to benchmark the difference (boot time, kernel compilation, glx gears, openoffice launch...)

Share this post


Link to post
Share on other sites

the last one is underway...

except I couldnt get the nvidia 64 bit drivers working...

 

and the RC for 10 amd 64 is pretty poor as a RC....

its not fair to use a 2.4 64 vs a 2.6 32 without knowing the patches and other comparisons and indeed that is what the artivcle cannonfodder pointed out did.

 

Id like to compare like with like kernel by kernel....

 

the RAID is a good idea... for RAID 0 HW /SW shouldnt be much different unless its a cached controller (mylex do some nice SCSI ones but theyre thousands each) used in netapps and auspex NAS boxes...

 

see my piece on winraid controllers for an explantion...

 

for a quick speed boost try hdparm ... it can also be used to make drives quieter if they support it!!!like in a multimedia boxen :D

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

×
×
  • Create New...