The King is Dead . . . Long Live the King!

[Hulk]

My P4 2.4B has been CRUSHED by an Athlon XP 2600+.

Take a look:

MS Pro Test

- Mark

[Dr Mordrid]

hehehe....could have told you that

Dr. Mordrid

[wkulecz]

How can Dave E's Athelon XP2000+ DDR266 Nforce2 results be so much better than Paul F's XP2400+ DDR266 Nforce2? Paul's 2400 seems barely better than Jerry's XP2000+ Nforce 420

What does the 7,3,3,2.5 refer to on Mike L's XP2600+ results?

Also Maurizio's XP2200+ seems no better than Terry's XP2000 both using SIS735 and DDR266. Can you explain this?

These inconsistancies bother me.

The best benchmark is the applications that matter. Best bang/buck on MSP6.5 is my only driver for ugrading at the moment, so I'm giving your page a careful look.

I'm waiting to see if MSP7 shows enough benefits from P4 before deciding which way to go.

--wally.

[Jerry Jones]

Well, we're probably going to have to start a new test category when MEDIASTUDIO PRO 7 ships.

I have a feeling that software is going to change rendering times even on the very same computers that ran version 6.5 of MEDIASTUDIO PRO.

Jerry Jones

http://www.jonesgroup.net

[davpen]

Dave E's machine is a bit of an anomaly. The other differences are fairly easily explained by memory bandwidth. Video compression seems to like lots of memory bandwidth. This is why a P4 2.53 with DDR266 can be beaten by a P4 2.4 with DDR333. It also explains why the XP2600+ is so far out in front. The 2600 comes with a new 333MHz front side bus in place of the 266MHz bus used in slower Athlon systems. When combined with DDR333 memory this makes for a noticeable improvement in memory performance and a corresponding decrease in compression time.

As for the 2200 vs. 2000 SIS735 systems the difference might be in the different versions of MSPro being used. Maurizo's system is using MPSPro 6.02 and Terry's system is running 6.5. I don't know how much of an effect this would have but this is probably the main problem wiith this benchmark page. It mixes quite a few different versions of MSPro so you have to be very careful when you read it that you are comparing like versions. In some cases the MPEG2 compressor is completely different (Ligos vs. MainConcept). It is also possible that different DV codecs are being used. I don't know how much performance difference there is between DV codecs.


Cheers,

David.

[Hulk]

Wally,

Please, don't shoot the messenger! I only report the scores.

First, remember that there are going to be inconsistancies in reported scores. Not everybody is adept at setting up a system for maximum performance. Without actually going through someone's system, it's hard to say. If the person seems to run the benchmark correctly I report the score.

Now, I will try to answer your questions.

Also Maurizio's XP2200+ seems no better than Terry's XP2000 both using SIS735 and DDR266. Can you explain this?

This one is easy. Maurizio is running 6.02, 6.5 is much faster for MPEG-2 encoding. The rest of the difference can be system setup.


What does the 7,3,3,2.5 refer to on Mike L's XP2600+ results?

RAM timings. Since this is the fastest system, and the owner submitted the timings, I thought I would include them for completeness.


How can Dave E's Athelon XP2000+ DDR266 Nforce2 results be so much better than Paul F's XP2400+ DDR266 Nforce2? Paul's 2400 seems barely better than Jerry's XP2000+ Nforce 420

This is tougher to explain. Dave E's system is performing much better than other similar systems, and Paul's much lower than it should be. I would suspect the faster run was on a system with a barebones installation of the OS and MS Pro with memory set at the fastest timings and a problem with Paul's setup.


There are always going to be inconsistancies in testing when you have the public submit scores. I often question people about abnormally high or low scores and they just said, "hey, I did it like you said and that's what I got." Then I have to decide to accept or reject the score.

The trick is to analyze the data and look for the trends. Yes, Dave's system is about 9% faster than it "should be" and Paul's about 15-20% slower than it "should be."

You can still see what the fastest AMD and Intel systems configuarations are.

Perhaps in the upcoming MS Pro 7 test I will be a little more stringent in accepting scores that fall far out of the norm for that particular system.

-Mark

[gt40]

Just tried it for the heck of it using mspro6.5

DV rendertime 4:27
Mpeg 4:04


Asus p4te, 2.4 Ghz, 512mb rdram, 400mhz fsb, intel 850 - untweaked

me thinks a little overclocking/tweaking would get me over the Athlon.

[Dr Mordrid]

Several things can affect the scores;

1. VIA chipsets have a reduced PCI bandwidth vs. SiS, Intel and nForce chipsets. Relative performance;

Intel
SiS
nForce
(big gap)
VIA

2. a given system may have more tasks running in the background than another, slowing its rendering performance.

3. on a tweaked system MSPro 6.5 screams & MSPro 7 screams even louder since it has a new multithreading core.

4. when comparing to Premiere we have to give the old man some sympathy. While Premiere supports dualies in a few functions MSPro7 supports them throughout. Those tests will be really interesting

I'm getting ready to release a new-from-scratch version of the test for MSPro 7.

Dr. Mordrid

[Hulk]

As soon as MS Pro 7 is officially released I'll start a new table of results. If a update is released in the future that significantly changes performance, I'll enter the scores for the new release in a separate table.

gtforty - Your scores are strange! All of the other P4 systems have dv and mpeg-2 scores that are about equal, or the mpeg-2 score is higher. Your mpeg-2 score is lower. Weird. Are you sure you're using "field a" and not "frame" in the option>general dialog. That would lower your mpeg-2 time.

-Mark

[gt40]

Mark,

I re-ran the tests, checked all my settings and got almost identical results (4:26/4:03)
changing to frame based lowered the mpeg2 times to 3:24.
I do believe Bob B. also had lower mpeg scores than dv.
Interesting to see the 2.4a(mine) a little quicker than the 2.4b's listed.

[Hulk]

Okay, I'll post the score in a bit.

Your system with a 400MHz bus is faster than the 2.4B with the 533MHz bus because RDRAM is still the fastest solution for the P4. I don't think that's going to change for quite a while, at least until we see dual DDR400 chipsets with the release of Prescott.

RDRAM is about 10% faster than DDR333, which is what I'm running.

One more thing to check. When you run the both tests did you make sure "smart render" was turned OFF?

-Mark

[Hulk]

Doc,

Is the new test going to be the same result as the old one, just using features in 7 that may expedite it? Or is the new test going to be completely new, thus making a more up to date and comprehensive test?

I'm thinking it's option number two. I don't think we need any backward compatibility.

-Mark

[gt40]

Mark,


Smart render? - couldn't find this option.

if you let me know how to check this option, I'll make sure it's turned off.

Thanks

Pat

[Hulk]

When you go to file>create>video

the save dialog pops up

then options>video editor tab

you'll see a "smart render" option, make sure it's unchecked.

[gt40]

Ok, found it,
unchecked and it didn't affect the time at all.

overclocking the m/b to 2640/440fsb gave the following : 4:04/3:42

You can post either result (incase you want to trounce the athlon with the o/c result), but I normally run my system unmodified.

Could the lower mpeg times be a result of some other h/w drivers (rtx10 drivers - keep in mind it has no mpeg accel hardware)


-Pat