Thinking of benchmarking speed using set rules...interested?

[jnk76]

Hello,

With all the debates of AMD vs Intel I was wondering if anyone was interested in starting benchmark test which tested computer performance with set guidlines to follow?

What I have come up with so far is:

Movie needs to be ripped to HD first.

Program I was thinking of using is DVD2SVCD which automates everything for this test.

You will need CCE and I believe only version 2.5 works with DVD2SVCD.

I would make ini file which would just be loaded into program. You load saving dir and load IFO file and start....that's it.

The log file can be copied and posted.

As for movie.....everyone has different opinions which is most common.
I have never tried to back-up Matrix but that I would think is very popular in these forums. If someone has a better than we can choose.

If interested please let me know and we can get started.

If not....then the debates can go on.

This can also be as tools to see performance increases due to different chipsets on Mobo or ram mhz....etc. It will not only be AMD vs INTEL.

[gmatov]

OK, jnk,
Wasn't hard to find your post, and with the auto reminder, will keep up.
I don't have Matrix, but think my youngest has. I can make them a backup.
I'm in. I WOULD like to know if my machine is a piece of crap or not.. Actually, I do all my encoding on the 1800XP in the next room, but that's no problem.

[freak_in_cage_10k]

this is a god idea, one i had (and probably other had) a while ago, check this post out, made by myself, would answer many of your questions!

https://www.videohelp.com/forum/viewtopic.php?t=149184&highlight=encoders+benchmark

(if it doesn;t work, search for "we need an encoders benchmark!!", authorede by myself)

sum 1 even suggested using the matrix as a source as it was one of the most common dvds!

i dunno why the idea never started, i guess it was going to be hard to make (i wasn;t skilled enough to tweak around with the software to make it).

personally i think a large source file (say.700megs?) of a standard movie should be downloaded (from this website), a package containing all the software could be downloaded (trial & freeware etc) and the exact same parameters could be loaded up & applied

the prob is all this downloading would be a prob for 56Kers and ppl could easily alter the parameters so as to "cheat" (due to amd vs intel etc). if it, however, a good idea

[Faustus]

This is an excellent Idea... thought a little out of my league I'd love to see it happen. Getting good across the board results would answer once and for all alot of questions. Like the Intel AMD thing, and how much does Dual CPUs help. Of course it would be specific to a specific encoder but it can't be helped. Maybe like a demo version of either CCE or Tmpgenc, or both.

Cheating is always an issue of course, but there isn't much to be done about it, it would be pretty much an honor system.

[freak_in_cage_10k]

yeah dude, it wuld be gd if ppl with hds in raid etc could post their scores, which would allow us to see the major determining factors, and those ppl who know a little about pcs culd remove sticks of ram, put in old, slower cpus and post those scores also, that way ppl would get a very good idea of how they could speed up their encoding time!

i wuld happily spend a few mins putting a few of my older processors in, and also removing 256 megs of my ram to see what kind of a difference it makes!

i think the 2 main probs encoundered wuld be using trial software and also acquiring the source file! i expect putting a large section of a specific movie (used for the source) would void copyright laws, and so the only way i could see it would be to download a specific trailer (eg matrix 2 trailer, as this will be avaliable for quite some time and is probably quite large!) also, this would mean that vcdhelp wouldn;t have loads of ppl downloading large files from them, thematrixonline could do that for us!!!! i doubt they would mind as it is free advertising for them!!!

as i said when i made the thread a few months ago, i would happily ty to contribute although i dont know much about this software stuff

[jnk76]

clarification:

This test piece would be a movie you rip to your drive in file mode.
There you need to own a DVD rom and know how to do it.

You also need to know very basic knowledge of DVD2SVCD. All you would need to do is load settings. Choose output dir and choose IFO file of main movie. Press start. That's it.

I am trying to make some constants and leave only the computer as variable.

Some good points made as well.

I have some 333mhz ram and 266mhz ram. It would be a good test to try both.

Warning about this benchmark. It can take a long time. I have it set at 4 passes. If you have an older machine......go to sleep.

If anyone wants the ini file which is to be loaded into DVD2SVCD send me your emails so I can email to you.

[Faustus]

Which movie was decide upon? The matrix? If so I'll have to buy it... I've seen it like 100 times when my old roommate had it so never actually purchased it.

On a side note I do own matrix revisited. doh.

[jnk76]

I think Matrix is very abundant amoung the users. I have never actually tried it so I will try tonight to make sure it works.

Another note.

*****Only original copies are to be tested******

[freak_in_cage_10k]

yeah, the matrix would be a gd idea, but they would all have to be either pal or ntsc!

why not all download a large trailer, that way everyone has EXACTLY the same source file!

[jnk76]

$HIT.

Forgot about that part. Pal and NTSC.

I don't agree with downloading a file because there are people with slow connections and I wanted this test to be a longer one not a trailer or short clip.

Well. That will have to be stated and will remain a variable.
hopefully we get to machines very close in equipment and then we can work out a ratio for one being slower than the other.

I am in NA so my discs are all Region 1.

[i_am_dave]

why not all download a large trailer, that way everyone has EXACTLY the same source file!

That sounds like a good idea. Any suggestions? Should probably be less than 200 meg. I will have to download from work since I have dialup at home.

[jnk76]

DVD2SVCD does not work with mpegs as that is what it is supposed to output. It can work with avi's but it is very buggy. This needs to be done with a ripped DVD structure. As I stated....Pal and NTSC will have to remain a variable. if possible....DO NTSC.

Only a couple people have asked for ini file which is required to be loaded into DVD2SVCD.

This test needs to be done only with this specific file.

PM me with your email if you want it.

[jnk76]

Ok.

The movie will be Matrix for now. Region 1 if possible.
Use the ini file that I have made and saved from DVD2SVCD to load into DVD2SVCD and go. I will run when I get home and post for my machine.

Any issues people have or questions please post.

[freak_in_cage_10k]

i_am_dave- how about the trailer for the matrix reloaded, or even better the matrix 3.

jnk76- you cant have such a large variable between the nstc and pal! thats stupid!! you cant have that kind of difference, the whole point in a benchmark is that you all have exactly the same version!!! thats like comparing this:

person A has a 1GHz AMD and a GeForce2 card, He gets 5000 Marks on 3d mark 2001

person B has a 2500+ AMD and a Geforce 4 Ti4600, he gets 3200 marks on 3d mark 2003

according to you person A has the better system because his score was higher, but you CANT COMPARE THE TWO BECAUSE THEY ARE USING A DIFFERNT BENCHMARK!!!!!!!! So, in conclusion there is no way you can have different sources!!! its stupid!!!!

you do have a point about dvd2svcd not accepting mpeg as a source, so lets not use dvd2svcd!!! Why not use sumthing like DVDx. v simple to use, everyone uses the same version and it does accept mpeg as source file!!! only prob is it may not allow source & output files to be altered (so raid drives could not be taken into account) as it is pretty basic

[jnk76]

No need for bashing and if my ideas are stupid then I will let you run with it.

The difference in encoding speed may be very minimal between NTSC and PAL. I don't know....do you? The big difference would be that we (NTSC) would have to do a pulldown where PAL would not. There are times for when programs get launched so that time could be taken out. Actually thinking about it more.....there are probably setting so it does not do pulldown.

As I stated though. In a perfect world yes the source would be the same but downloading a small trailer file to do the testing IMO is not the way to go. The file is too small and will not show a long term result.

[gmatov]

jnk,
Why not just do it in 2 forms, those in PAL regions do PAL, and post, NTSC do the same. No argument about, well you're doing PAL, no fair. Comparisons would be output specific.
List all the results from PAL users, then all the NTSCs in their own column.
Unless the .ini file IS specific, in which case, how hard would it be to change?

[jnk76]

It would be pretty easy to change ini file but unfortenetly I can not test to make sure it works. Splitting the results for each Region is a good idea.
Once people start feeding back info then I can start entering into spredsheet type of form and then post.
I will make mod's to ini file tonight and if you require the file please specify if you require PAL or NTSC version.

[freak_in_cage_10k]

jnk76- i made a post v soon after you made your one starting with "No need for bashing etc", i dunno why it didn;t get posted correctly??? prob my pc!

ayway, it basically appologised for getting a bit tetchy with you. you would have to distinguish between pal and ntsc, and a bit ofother stuff, sorry m8