PCMARK05 testing

so the coach finally put me in the game first time for the season, SNAPPED off my warm up pants, got the sweat bands ready to go,

GIGA X58-UD4P
GSKILL PS 17066 C8
GSKILL Falcon SSDs
ADAPTEC 5805 raid controller
1 GIGA iram
SAPPHIRE 4890

http://i200.photobucket.com/albums/a...eezy/33596.jpg

http://i200.photobucket.com/albums/a...y/IMG_0950.jpg

http://service.futuremark.com/compare?pcm05=2152753

u' so throwed dude :p::rofl:

she thinks your SOOOOO throwed man

Damm nice SSD tweaking!

what did you do to your "web page rendering" score, crazy pages/sec?!?! very nice :up:

thanks mate, just using a different browser for webpages/sec

Nice testing :up: This december I'll finally devote some time to this benchie. I've wondered about forcing a faster web browser for that test, you confirmed my suspicions :) Looking forward to the complete run.

Ah nice you found a better one than I did :D I gave that tip to dinos ages ago too

you did? i guess i dont really remember cause i dont enjoy this bench hahahah

welldone pro! you are the best ! (nacho libre voice)

Quote:

---

Originally Posted by dinos22

you did? i guess i dont really remember cause i dont enjoy this bench hahahah

---

Yeah was in PM when you were having crap results with 5x iram.

thanks karmakazi, you and chispy have always been very good at pcmark

Video encoding is HUGE.....How come?..... :eek:

P.S. Is it "legal" to "manipulate" the "default" benchmark?.....
I mean, this thing about the PCMark and CHANGING the DEFAULT browser of it with ANOTHER one, is it "within the default settings" of the benchmark?... :(

Cool score.

Futuremark has actually not implented the browser in the benchmark - PCmark05 is only looking for Internet Explorer and starting it. The have seen this loophole and closed it in PCMark Vantage - which comes with it's own browser included.

Ahhh.......I see........thanks... :)
So the only think someone has to do, is to uninstall the IE and install another browser right?...... :)

Quote:

---

Originally Posted by hipro5

Ahhh.......I see........thanks... :)
So the only think someone has to do, is to uninstall the IE and install another browser right?...... :)

---

Yeah and rename it iexplorer.exe. Same goes for media encoder. Find a better encoder, maybe a multi threaded one and let it do the crunching.

Is it only possible to tune at the level of application or also at the level of the file?

Please guys, speak ENGLISH.......It's only Greek to me.... :D
Explain I mean with simple words...... :D

Wow PCMark is such a lovely benchmark, maaaan I'd love knowing how to tweak it that much :yepp:
Grats pro :up:

Quote:

---

Originally Posted by hipro5

Please guys, speak ENGLISH.......It's only Greek to me.... :D
Explain I mean with simple words...... :D

---

Changing browser/decoder => tuning application used
Changing file to decode => tuning, ehr, file.

The first seems to be no problem, but I don't know about the second.

Quote:

---

Originally Posted by massman

Changing browser/decoder => tuning application used
Changing file to decode => tuning, ehr, file.

The first seems to be no problem, but I don't know about the second.

---

For browser test; changing the temporay html file wouldn't really do much. Disabling showing images in your browser settings helps a tiny bit - but something like that I feel that's a bit of an optimzation...

I had attempted to mess with the other temp files ( general usage , etc etc.) but never accomplished anything. Don't recall trying to mess with the encoding - but honestly doubt it. If I recall correctly they are actually encrypted. But been too long ago that I looked at those files.

I like your thinking!

they are some sort of archive file, if you do a check with a hex editor you can see all sorts of contents, webpages, images, etc... i have no been able to uncompress them to examine the content properly...

browser test, yeah different browser definatly does help, and browser optimization also helps...

using a different video encoder and playing with the encoding settings definatly scales this one up aswell..

and for the hdd tests you need three arrays and dynamically switch between them during the benchmark, each hdd test wants different optimization, so you need to treat them all seperately... one array for all three will almost certainly leave you missing something...

Quote:

---

Originally Posted by youngpro

thanks karmakazi, you and chispy have always been very good at pcmark

---

Thanks for the kind words :)

When you throw a 5870 in there... my god... 34.5k easily... and then max out xp startup it will be :eek:

with all of these 'optimizations' is that score even submittable on the hwbot? i kinda think this is going against the spirit of the benchmark. a benchmark is so that you can compare your system to other peoples, but if you alter the benchmark ITSELF, isnt that cheating?

i disagree with that, tweaks and optimization is what makes the difference.. i think if your willing to put in the time and effort to understand a benchmark you shouldnt be penalized for that....

i mean if clocks was all we were after, we could use cpuz and gpuz validations, but that wouldnt be much fun at all :)

Nice work and tips there, never even thought of that :p

Quote:

---

Originally Posted by youngpro

i disagree with that, tweaks and optimization is what makes the difference.. i think if your willing to put in the time and effort to understand a benchmark you shouldnt be penalized for that....

i mean if clocks was all we were after, we could use cpuz and gpuz validations, but that wouldnt be much fun at all :)

---

100% agree with you. :up: Same goes for your IGP 2k6 methodology, I loved it. :yepp:

'tweaks and optimizations'?? changing the very program itself does not classify as a tweak or optimization. you tweak or optimize the COMPUTER, not the benchmark. if you are writing your own benchmark, basically, then why do you bother posting it under PCMARK05. why not call it YOUNGPRO05? you cannot use this to compare to others, you are using different programs. i have a feeling these scores would not pass muster in the hwbot if they knew the methods used to obtain the scores.

Quote:

---

Originally Posted by Computurd

'tweaks and optimizations'?? changing the very program itself does not classify as a tweak or optimization. you tweak or optimize the COMPUTER, not the benchmark. if you are writing your own benchmark, basically, then why do you bother posting it under PCMARK05. why not call it YOUNGPRO05? you cannot use this to compare to others, you are using different programs. i have a feeling these scores would not pass muster in the hwbot if they knew the methods used to obtain the scores.

---

I couldnt disagree with you more:shrug:

Tweaking is where its at. And who do you think the person posting in here named Massman is? He is a hwbot admin...

massman himself has expressed some hesitation about these changes as well, if you read his post above

Quote:

---

@massman
Changing browser/decoder => tuning application used
Changing file to decode => tuning, ehr, file.

The first seems to be no problem, but I don't know about the second.

---

i can appreciate massmans opinion for sure but firstly im not submitting this to hwbot at this stage and secondly i dont think, with all respect to pj that he is the global oc authority

Great work James amazing result mate!!!

Quote:

---

Originally Posted by youngpro

i dont think, with all respect to pj that he is the global oc authority

---

TRUE story

i understand youngpro that pj isnt the global authority. however the spirit of overclocking used to be getting the fastest performance out of various pieces of your hardware in conjunction with each other. lately it seems that everyone is using different 'tweaks' to the benchmarks themselves that radically alter the scores. i think when you go so far as to use a hex editor the alter the benchmark itself you have gone too far. what is going to stop the next guy out there from modifying it to a point that he gets 60k? oh well i digress. just my two cents.:rolleyes:

Quote:

---

Originally Posted by Computurd

massman himself has expressed some hesitation about these changes as well, if you read his post above

---

To be honest, the words you quote are not hesitation to allow this kind of tweaking, but wondering whether it's possible to also tune at the level of the file.

For the record: no official hwbot policy or opinion has been expressed in this topic.

Quote:

---

Originally Posted by youngpro

secondly i dont think that he is the global oc authority

---

Couldn't agree more.

I don't like where this is going...

Quote:

---

Originally Posted by Computurd

i understand youngpro that pj isnt the global authority. however the spirit of overclocking used to be getting the fastest performance out of various pieces of your hardware in conjunction with each other. lately it seems that everyone is using different 'tweaks' to the benchmarks themselves that radically alter the scores. i think when you go so far as to use a hex editor the alter the benchmark itself you have gone too far. what is going to stop the next guy out there from modifying it to a point that he gets 60k? oh well i digress. just my two cents.:rolleyes:

---

Oh really? You think people JUST started using tweaks? LMFAO :rolleyes:

Do I need to dig up a thread about hardcore tweaking 3dmark 2001 from 7 years ago ? If you read what he said, he could only look at the files with a hex editor, not alter them. There is nothing wrong with understanding how the benchmark runs and what it is composed of.

Me either. Might be time to dump PCM05 and move to PCMVantage.

Quote:

---

Originally Posted by [XC] gomeler

Me either. Might be time to dump PCM05 and move to PCMVantage.

---

pcmvantage sucks.... its way too long. half hour to get a result? What happens if you lock up when bench is almost done, or if that happens a few times. You spent 2-3 hours to get one bench in? lame.

Not to mention you can only run it once unless you pony up some cash

you are calling it tweaking karmakazi. i call it cheating. i am sure they were doing it seven years ago. people were also murdering people seven years ago, but does that make murder okay? hell no it doesnt.
when you are actually modifying the very benchmark itself you are cheating. you are no longer running the benchmark. you are using diifferent browsers, encoders yada yada yada/.
that is my opinion, and one that many others share.
there are many others who share your opinion.
so lets keep it that way.

agree to disagree i say.

pcmark vantage is a total system benchmark. just because you actually have to have a stable system to run it does not mean it sucks. does not take 2-3 hours either.

Quote:

---

Originally Posted by Computurd

you are calling it tweaking karmakazi. i call it cheating. i am sure they were doing it seven years ago. people were also murdering people seven years ago, but does that make murder okay? hell no it doesnt.
when you are actually modifying the very benchmark itself you are cheating. you are no longer running the benchmark. you are using diifferent browsers, encoders yada yada yada/.
that is my opinion, and one that many others share.
there are many others who share your opinion.
so lets keep it that way.

agree to disagree i say.

pcmark vantage is a total system benchmark. just because you actually have to have a stable system to run it does not mean it sucks. does not take 2-3 hours either.

---

Your a troll and I wont bother feeding you anymore :p:

i will not post on this subject any more, either. but calling me a troll and moving away from the debate proves nothing on your behalf. i take it as a gracious exit however.

Sorry to any thread hijacking pro. This is an AWESOME fully tweaked and perfectly run legit result. Big ups to ya! :up:

Quote:

---

Originally Posted by TheKarmakazi

Oh really? You think people JUST started using tweaks? LMFAO :rolleyes:

---

Well if you change the way a program does it tests by effectively altering some integrated benchmarks of it, it's going a bit too far to my liking. There's a difference to change a run order in a 3dmark then there's to change an internet browser. This goes just that bit to deep. I love that people are always innovating and improving things, but try keep the test on a fair level. You can't do all this in a live event, as it will not be allowed.

Quote:

---

Originally Posted by Leeghoofd

Well if you change the way a program does it tests by effectively altering some integrated benchmarks of it, it's going a bit too far to my liking. There's a difference to change a run order in a 3dmark then there's to change an internet browser. This goes just that bit to deep. I love that people are always innovating and improving things, but try keep the test on a fair level. You can't do all this in a live event, as it will not be allowed.

---

What does a live event have to do with this topic? You dont have to change the internet browser from IE to tweak it either. I can get quite a nice boost by tweaking IE options

As there it is where you show your tweaking and use the given tools... not add your own. You are looking it from your perspective only, I'm looking at it more globally. (This from what I've learned the last years and what I know people do, to get a better score or win an event legally or illegally.)

James has shown a few tips lately :p and I must say I'm impressed by his devotion to stay up front, but some stuff is far out of my league ( too stupid or too lazy for it ) or offends some of my own principles.

You adapt or change what you like, that's your own perogative.

Quote:

---

Originally Posted by Computurd

i think when you go so far as to use a hex editor the alter the benchmark itself you have gone too far.

---

i just want to make it really clear the only benchmark that has been physically modified is the 3dmark01 team au edition and that is clearly modified for a fun and not competitive benching purpose, i certainly havnt and wouldnt touch anything with a hex editor that i would show benchmarks with, without specifically saying so

RE: BenchZowner,

Well i think there is these kind of tweaks in all benchmarks, i think LOD is a prime example of a tweak that is way more important than anything im doing and easily available, we all know it and use it.. i guess what i wonder is, does it matter how much a tweak alters a benchmark score, if it is a tweak, and not something specifically banned by the rules (such as altering the perception of a benchmark or mipping for example) then is it not just advancing benching? this was going to happen sooner or later with the advancing state of this sport we are playing... i remember reading what happened when ln2 cooling first came in, all the guys on air and ss didnt like it and were not interested, but look at us now? we all have dewars at home and spend our pay cheques on a scientific liquid to cool our pcs...

oh and just on the tweaking, to me i have alot more fun sifting through the registry or analyzing the way a benchmark works and trying to find ways to tweak its performance than i do benching with ln2.. thats just me though, everyone has to do what is enjoyable to them..

i know since i posted this ive found some more things to advance the score of pcmark further so i would say 45k would be totally possible with a fully setup rig...

In the old days, LN2 benching was "cheating". Then LOD was "cheating". And so on and so forth...

Quote:

---

Originally Posted by Gautam

In the old days, LN2 benching was "cheating". Then LOD was "cheating". And so on and so forth...

---

:rofl: true true

I think we'll should be thankful to youngpro for sharing his tweaking method and trick as we all are here to learn, right? :up:

Waiting for the 45k now.. :D

Quote:

---

Originally Posted by Marco André

I think we'll should be thankful to youngpro for sharing his tweaking method and trick as we all are here to learn, right? :up:

---

Couldn't agree more :up:

Nicely done youngpro alot to thing about the next time i bench pcmark05 thanks:up:

Looks like the score got pulled :shrug:

yeah lucky im not interested in hwNOT ;)

If i might interject my 2 cents here.......I saw this in person and the benchmark itself is in no way altered.........

I would say this is no further out of bounds than what people do to get SLI to work on AMD platforms with AMD chipsets.........

Major thread necro but for those interested in HWBot:
hwbot PCMark05 submission must comply to the following rules:

•use default PCMark settings
•have a valid screenshot (see example below): clearly show PCmark score, PCmark settings, processor in CPU-Z, GPU-Z, Subtest Scores (XP startup speed!), unless you have a valid futuremark ORB url
•have a valid futuremark orb link as verification for scores in the top 20

Forbidden Hardware:

•software ramdisk (iRam is allowed)
•MFT Software

Forbidden Tweaks:

•any software or human interaction altering the perceived speed of the benchmark program, tricking it to believe it ran faster

So no you cant use it on HWBot legally, If you want to use it for other purposes or for your own gain then so be it :)

and the reason you up this thread? we talked about and it was a closed book ;)

Damn....no kidding. Case was closed a long time ago :shakes:

some people can't let this go, so here, ill be a real nice guy and explain a little clearer..

results i put in the public domain, i never physically modified any binaries, i never replaced any codecs or replaced dlls.. i used traditional tweak methods with new ideas..

can i share one example? the pcmark 05 video encoding going through the roof... 1 simple registry patch.. this part of the benchmark is controlled by the windows media encoder, and this is why you must installed this windows media encoder and windows media player 10 if you want to play pcm05..

you might also want to check out wmv powertoy as you can manipulate all the settings of this powertoy directly via a gui

here are the settings i used, im sure you can find better ones if you so choose
http://i200.photobucket.com/albums/a...oysettings.jpg

Code:

---

Windows Registry Editor Version 5.00

[HKEY_CURRENT_USER\Software\Microsoft\Scrunch\WMVideo]
"Compression Optimization Type"=-
"DenoiseOption"=dword:00000000
"Dquant Option"=dword:00000000
"Dquant Strength"=dword:00000000
"Force B Frame Delta QP"=-
"Force Encoding Width"=dword:00000020
"Force Encoding Height"=dword:00000020
"Force LoopFilter"=dword:00000000
"Force Median"=dword:00000000
"Force NoiseEdgeRemoval"=dword:00000000
"Force NumThreads"=dword:00000004
"Force Overlap"=dword:00000000
"Force Video Scaling"=dword:00000000
"Lookahead"=-
"Macroblock Mode Cost Method"=-
"Motion Match Method"=-
"Motion Search Level"=-
"Motion Search Range"=-
"Motion Vector Coding Method"=-
"Motion Vector Cost Method"=-
"NumBFrames"=-
"Perceptual Option"=dword:00000000
"VideoType"=-
"Lookahead Ratecontrol"=-
"Encoder Thread Affinity"=-
"Force Range Reduction"=-

---

thats it right there, feel free to test it to compare if you want...

Well that ought to shut up the haters, well done pro :)

Don't give them ideas Young Pro :p

Darn, seems you got far too much time to explore all them settings...

Hardcore stuff there dude, well done! It's been so long since someone actually found a new tweak in the benchmarks we all use :up:

Quote:

---

Originally Posted by youngpro

some people can't let this go, so here, ill be a real nice guy and explain a little clearer..

results i put in the public domain, i never physically modified any binaries, i never replaced any codecs or replaced dlls.. i used traditional tweak methods with new ideas..

can i share one example? the pcmark 05 video encoding going through the roof... 1 simple registry patch.. this part of the benchmark is controlled by the windows media encoder, and this is why you must installed this windows media encoder and windows media player 10 if you want to play pcm05..

you might also want to check out wmv powertoy as you can manipulate all the settings of this powertoy directly via a gui

here are the settings i used, im sure you can find better ones if you so choose
http://i200.photobucket.com/albums/a...oysettings.jpg

Code:

---

Windows Registry Editor Version 5.00

[HKEY_CURRENT_USER\Software\Microsoft\Scrunch\WMVideo]
"Compression Optimization Type"=-
"DenoiseOption"=dword:00000000
"Dquant Option"=dword:00000000
"Dquant Strength"=dword:00000000
"Force B Frame Delta QP"=-
"Force Encoding Width"=dword:00000020
"Force Encoding Height"=dword:00000020
"Force LoopFilter"=dword:00000000
"Force Median"=dword:00000000
"Force NoiseEdgeRemoval"=dword:00000000
"Force NumThreads"=dword:00000004
"Force Overlap"=dword:00000000
"Force Video Scaling"=dword:00000000
"Lookahead"=-
"Macroblock Mode Cost Method"=-
"Motion Match Method"=-
"Motion Search Level"=-
"Motion Search Range"=-
"Motion Vector Coding Method"=-
"Motion Vector Cost Method"=-
"NumBFrames"=-
"Perceptual Option"=dword:00000000
"VideoType"=-
"Lookahead Ratecontrol"=-
"Encoder Thread Affinity"=-
"Force Range Reduction"=-

---

thats it right there, feel free to test it to compare if you want...

---

:clap:
Thanks for bring it out in the open so everybody can see. :yepp:

Since this isn't physically changing the app, just using a 3rd party app, does this make using this app legal? Seems it would be since we use many other 3rd party apps to help improve scores.

Love your work Pro, thanks for the heads up :up:

And this is so true dude....

Quote:

---

Originally Posted by pro

and remember, if everyone has it, its a tweak, if one person has it, its a cheat

---

Now the browser trick :p

hwbot may allow encoder tweaking, but it needs to be looked at/discussed first. Pretty sure FM would say no but hwbot on the other hand may because it has always been more open to tweaks.

Changing the browser however will never fly.

Quote:

---

Originally Posted by Leeghoofd

Now the browser trick :p

---

pro said before he's using another browser to bump the score *a bit* :D

Serious skill :bows:

well im not asking hwbot allow anything, i am simply putting something in the open that i hear regular untrue rumours about.. thats all guys, if you want to play with these sorts of things, please keep them off futuremark and hwbot unless those guys say they are okay

Pro you take the time to test and try stuff - that is how so many of the accepted tweaks we use now were found. People like you, true enthusiasts, who simply wanted to see just what could be done pioneered them all. So as I have said to you in the past you keep on rocking dude and thanks for sharing what you find. More power to the enquiring mind I say.