xbench.NET v0.5 - Quick XviD Encode Speed Test

[The Scientist]

Application to test your systems encode speed using the XviD codec. May be handy when overclocking or upgrading to compare speeds or compare against different peoples systems.

Download, unzip, and install xbench.NET program file.
Download, unzip, and install xbench.NET media file.
The media file is now a seperate download and only needs to be downloaded once, even when program gets updated.

The 'xbench.NET' process should only take a minute or two to complete. On completion, the installation folder will contain a both a screenshot and text file of your score.

Download xbench.NET Program File >> HERE (http://www.dvduk.pwp.blueyonder.co.uk/_files/xbench.NET.zip) <<
Download xbench.NET Media File >> HERE (http://www.dvduk.pwp.blueyonder.co.uk/_files/xbench.NET Media File.zip) <<

My system with Vista. (XviD v1.1.2)
http://www.dvduk.pwp.blueyonder.co.uk/_files/c2d259.png

Ouch!
http://www.dvduk.pwp.blueyonder.co.uk/_files/K6-500.png

[jeffy]

http://img454.imageshack.us/img454/6167/benchtz2.jpg
XP Pro

[BigDid]

Hi,

If this test is the same as in this post: http://forum.doom9.org/showthread.php?p=956303#post956303
could I suggest using Xvid 1.2 (cvs) for the duo/multi-core owners?

Available in Celtic_druid builds: http://ffdshow.faireal.net/mirror/XviD/
:thanks:

Did

[The Scientist]

Hi, could I suggest using Xvid 1.2 (cvs) for the duo/multi-core owners?

Available in Celtic_druid builds: http://ffdshow.faireal.net/mirror/XviD/
:thanks:

DidWhich file is that exactly? I can't see any v1.2

[BigDid]

Which file is that exactly? I can't see any v1.2
First line (last build) from cvs:
http://ffdshow.faireal.net/mirror/XviD/XviD.cvs.head.MTK.exe

Did

[The Scientist]

Thanks, will check it out.

[The Scientist]

could I suggest using Xvid 1.2 (cvs) for the duo/multi-core owners?Thanks for the 'heads up'. As you can see from the first post there is quite a significant increase in speed using the CVS, more so as the 'intensity' of the encode increases. cool.

Thanks again BigDid :goodpost:

[jeffy]

First line (last build) from cvs:
http://ffdshow.faireal.net/mirror/XviD/XviD.cvs.head.MTK.exe

Did

I'm sorry but ClamAV shows it has: Trojan.Downloader.Zlob-545
Any explanation, eg. false positive?

[jeffy]

http://img153.imageshack.us/img153/8715/suspiciousxvidmtkai5.th.jpg (http://img153.imageshack.us/my.php?image=suspiciousxvidmtkai5.jpg)

I have decided to submit the file to VirusTotal.com, 4 alerts were issued.

Note: it was renamed before submitting as (20122006) XviD.cvs.head.MTK.exe

[The Scientist]

Seems fine here, I'd suspect the clamAV. It did pick something up, but said it was Zlob-606, seems it can't make up its mind.

Anyway, that XviD file, author, and site is used by many here so I wouldn't expect anything dodgy to be in the files.

[Inventive Software]

Yeah, don't be surprised to find certain AV companies picking up UPX'd .exe's as viruses. Most likely they've been paid to detect certain coding practices so that another product gets flagged instead (the NSIS virus anyone?)

[jeffy]

Taking your word for UPX, I tried upx -d unpacking.
http://img144.imageshack.us/img144/6918/suspiciousxvidmtkunupxtm3.th.jpg (http://img144.imageshack.us/my.php?image=suspiciousxvidmtkunupxtm3.jpg)
However, there were still 2 (different) warnings being issued by VirusTotal.com, even after unpacking.

What about "Trojan-Downloader.Win32.Small.dnb"?

[jeffy]

Xvid Koepi's build as found here:
http://www.koepi.org/XviD-1.1.2-01112006.exe
http://img68.imageshack.us/img68/654/xvid11201112006dd8.th.jpg (http://img68.imageshack.us/my.php?image=xvid11201112006dd8.jpg)

x264 build as found here:
http://mirror01.x264.nl/x264/revision622/x264.exe
http://img100.imageshack.us/img100/2227/x264r622se9.th.jpg (http://img100.imageshack.us/my.php?image=x264r622se9.jpg)

[The Scientist]

Now, after unpacking, that ClamAV shows no virus/trojan doesn't this make you think that lots of these type of programs (trojan checkers e.t.c.) are just shite.

Further to this, I just checked the official XviD v1.1.2 file and guess what, the totalvirus scan just picked up four so-called virii/trojans. What a load of crap...... I'm made a note of the four scanner programs so I make sure I never buy/use any of their products.


Just a note:
This thread seems to have gone completely off track. Any chance an admin can split the thread, maybe at post #8, to its own thread about the virus/trojan stuff, if that's possible?

[BigDid]

http://img454.imageshack.us/img454/6167/benchtz2.jpg
XP Pro
Hi,

I'll try the new test later (encoding on actually).
I'm like jeffy with my E4300 o/c to 2400 I have no comparable results as the E6600/2400 :(
I'll investigate ...

@The scientist, could you developp on your x, y z parts? Is it 1/2 pass related or ...?
:thanks:

Did

[The Scientist]

@The scientist, could you develop on your x, y z parts? Is it 1/2 pass related or ...?They are 1 pass (quant) mode, but it doesn't really matter, it's not a test to determine how many FPS you can encode with using XviD as such, but the encode speed in comparison to someone else running the same program (xbench.NET). So I could get the FPS value and * 2 / 3.542 ^ 3 -1982 + 7, it doesn't matter, as long as it does the same everytime it's ran and on other systems and/or configurations.

lol, anyway, now that's out of the way, roughly, this is some of the settings (approx):

[X]
quant_type 0 (H263)
No Chromame
no qpel
no gmc
motion_search/vhq_mode 5/1
turbo ON

[Y]
quant_type 1 (MPEG)
Chromame ON
qpel ON
no gmc
motion_search/vhq_mode 6/1
turbo OFF

[Z]
quant_type 2 (MPEG Custom)
Chromame ON
qpel ON
gmc ON
motion_search/vhq_mode 6/4
turbo OFF
Custom Matrix

Like I said, it doesn't matter what the settings are, as it's a comparison test to other tests using the same program and settings.

[The Scientist]

I updated to v0.4. Added an automatic screenshot of the application window on completion of the test. PNG file will be saved to the application folder.

There's no excuse to not post screenshots now......... :)

[Mr_Odwin]

My first automatic screenshot (:)):

http://img116.imageshack.us/img116/3418/xbenchnet1qp7.png

My second, more useful shot:

http://img253.imageshack.us/img253/9544/xbenchnet2nm0.png

[jeffy]

Further to this, I just checked the official XviD v1.1.2 file and guess what, the totalvirus scan just picked up four so-called virii/trojans. What a load of crap...... I'm made a note of the four scanner programs so I make sure I never buy/use any of their products.


I'm sorry but I considered as an official 1.1.2 build Koepi's build, as the official Xvid platform download page leads there:
http://www.xvid.org/Downloads.15.0.html

As you can see above, 30 out 30 AV scanners gave it a CLEAN bill, without any trace of viruses/trojans.

Anyway, I'm sorry that I made this thread off-topic, but the results really surprised me. I will get back on the topic with another screenshots soon.

[jeffy]

Cropped a little
XP Pro, Xvid 1.2-127 Dec 8, 2006 as supplied with AutoGK 2.40
http://img405.imageshack.us/img405/9148/xbench04gs9.png

[The Scientist]

Anyway, I'm sorry that I made this thread off-topic, but the results really surprised me.No need to apologise, at least it gave us all something to chat about. :)

Thanks for posting the screenshots, nice to see the different numbers on the different systems.

@jeffy: That's some over-clocking you got going there, I had my old p4 (2.4GHz) clocked at 3GHz for years, it gave me a warm feeling knowing I was getting something for free :) But a 2.4 to 3.5, phew, I'd be scared, I wanna stick mine up to 2.8 or maybe 3.0 but untill I get hold of a decent modern mATX board that actually has some overclocking features I'm stuffed.

Your results are interesting though.......
Notice the first test [X] is almost the same between mine and yours, even though I'm running @ 2.4 and you at 3.5. The [X] encode is using low (fast) speed encode settings so I would assume here the CPU is not the bottleneck and it's the rest of the PC (board/ram/hard drive) that's struggling to keep up, hence the similar value. But as we progress, and the encode uses higher and slower settings (in comparison), the rest of the PC can handle the flow and it appears the encode then gets more dependant on the CPU which is why the gap in results then becomes more apparent, it becomes a much better test between the different processors and speeds.

[jeffy]

I don't know if you can incorporate them, but maybe if it could enable multithreading (eg. 2 threads) the results would have been even more different.

Overclocking....
Well, Core 2 Duo are rumored to have a really good overclocking potential....

I can only recommend to you reading these articles:

Cheap Thrills: Core 2 Duo E6400 Overclocked to 3.33 GHz
http://www.tomshardware.com/2006/10/10/cheap_thrills/

Quote:
"Using the stock cooler, we were able to reach 3.1 GHz at 387 MHz FSB and 1.45 V using the Intel stock cooler, where increasing the voltage only made the CPU overheat sooner. Adding the CNPS9700 LED brought down those temperatures phenomenally, but only got us to 3.15 GHz at 394 MHz FSB and 1.48 V before things once again got too hot. The big gun Gigabyte 3D Galaxy II liquid cooler brought far greater success:3.33 GHz at 417 MHz FSB and 1.52 V!"

http://www.tomshardware.com/2007/01/04/overclocking_guide_part_2/page3.html

500 MHz FSB? Core 2 Duo Overtakes Core 2 Extreme
http://www.tomshardware.com/2007/01/11/core-2-duo-overtakes-core-2-extreme/
Quote:
The test system was based on a Gigabyte GA-965P-DQ6, and reached an amazing 485 MHz FSB speed at stability sufficient for benchmarking. We recommend against running a system at these extreme settings in the long run, so as not to cut component lifetime down too much.

[The Scientist]

I don't know if you can incorporate them, but maybe if it could enable multithreading (eg. 2 threads) the results would have been even more different.It already did.

Well, Core 2 Duo are rumored to have a really good overclocking potential....I know, that's why I got one.

I can only recommend to you reading these articles:I already have :)

[BigDid]

Hi,

Finally got time to make the test.
E4300 @2400 with cool&quiet DDR1 Ram XPPro Sp2, still in test:

http://img01.picoodle.com/img/img01/7/2/21/f_xvid1204m_97748ae.png (http://www.picoodle.com/view.php?srv=img01&img=/7/2/21/f_xvid1204m_97748ae.png)

Also at stock speed: 1800Mhz

http://img136.imageshack.us/img136/4416/xvid121800av2.png (http://imageshack.us)

@The scientist, thanks for the details. I still don't understand the difference in the 1st test between a C2D same speed but the stock is fast, the overclocked one is slower :(

Did

[ricardo.santos]

Here´s my results, seems its the slowest of all the ones here :(
512mg Ram

http://img530.imageshack.us/img530/6249/xbenchwe4.jpg

[ditche]

Atlon 64 X2 4200+ s939 @ stock :

http://img170.imageshack.us/img170/2831/xbench01ph3.png

@ 2,53 GHz :

http://img170.imageshack.us/img170/2897/xbench02xw5.png

[The Scientist]

Thanks for all that took part with your screenshots. I've now updated to v0.5, made a proper installer, and pulled the media.dat file out into a seperate download. First post contains download links as usual.

xbench.NET now just does a low settings (fast) encode and a high settings (slow) encode as I think the X,Y,Z thing confused people. Each test is ran twice and the best score is the one displayed. After each test encode, both a screenshot and text file of the result is saved in the xbench.NET folder if needed.

If enough people test I'll put a table together in the first post like 'graysky' has with his MeGUI tests.

[jeffy]

Xvid 1.2-127, Dec 8 2006, as supplied with AutoGK 2.40
(slightly cropped)
http://img241.imageshack.us/img241/2473/croppedny7.png

DDR2-533@600

[hopstiii]

Hi

My test
Xvid-1.1.3-28062007 _Final Release_
Based on CVS from 28.06.2007 18:30h MET

CPU : AMD Duron(tm) Processor
SPEED: 796Mhz

Best Time: 48,0 Speed: 20,8 FPS
Best Time: 197,4 Speed: 5,1 FPS

Most of my movies have been encoded @ 1,5 to 2 fps
With: MPEG, Quarter Pixel, GMC, 6-Ultra High, 4-Wide Search, Chroma Motion, Trells Quantization.

:)