PDA

View Full Version : May Madness a ROG Challenge



Pages : [1] 2

HiVizMan
04-20-2013, 02:17 PM
Time for a bit of fun with our hardware. There are no prizes and this challenge is open to all who are members of the ROG forum. You only have to use a ASUS motherboard to qualify. Offical challenge back ground to be used for screenshots. HWBOT rules apply.

19810

The Benchmarks

3Dmark2001SE
Wprime1024
Aquamark

Hardware Limitation
ASUS Motherboard only
The same hardware has to be used for all 3 benchmarks
No Sub Zero cooling - photographic proof will be required
CPU frequency has to be the same for all three benchmarks - and if your result is to good to be true, then it most probably is.
Single Video card or Dual Video card
If you are in the top three a video will be required showing a successful completion of your Wprime run.


Benchmark Limitations
3DMark2001SE Nature Score is Capped at 1000
Same Operating System for all three.
Same CPU frequency for all three.


Start 1 May 2013
Close 31 May 2013

Points will be allocated 1 to 20 places.
1 - 20 pts
2 - 19 pts
3 - 18 pts and so on down to
20-1 pt

Bonus Points for when entries Submitted
Week 1 - 5 pts
Week 2 - 5 pts
Week 3 - 2 pts

http://freeallsoftwares.com/wp-content/uploads/2013/01/Aquamark-3.jpg

20247

Zka17
04-20-2013, 02:30 PM
OMG, we have a new challenge! :cool: Thank you, Sir!

Questions:
1. Even the GeForce Titans are allowed? I haven't done 3DMark2001SE or Aquamark with them, but they're a beast!
2. The same OS has to be used too? I know that the wPrime 1024M likes the WinXP, but the 3DMark2001SE and Aquamark may like the Win7 better (not sure as I haven't tried yet)

Chino
04-20-2013, 02:41 PM
Sounds good, my friend. Typo in the title though. Should be Challenge. ;)



2. The same OS has to be used too? I know that the wPrime 1024M likes the WinXP, but the 3DMark2001SE and Aquamark may like the Win7 better (not sure as I haven't tried yet)
3DMark2001SE and Aquamark favor Win XP too. ;)

HiVizMan
04-20-2013, 02:45 PM
If they have it then let them use it. The Nature cap will make the playing field a little more even. But there will never be parity or a perfect handicap. Such is benching.

Zka17
04-20-2013, 02:58 PM
OK, then! :cool:

Menthol
04-20-2013, 05:38 PM
HiVizMan,
Thanks for putting on these challenges, My weakest benching is 2d so this is just what I needed.
I just made a trial run of 2001 and scored 700+ in nature, and increasing gpu frequency only lowered the score, I will have to search the internet to get some ideas on what to do as 1200 looks a long way off.
Which background do we use?

HiVizMan
04-20-2013, 06:38 PM
Back ground will be uploaded on the last day of April :D

There are many many tricks to 2001 indeed. And I will send you a guide as you will compete here for sure.

Regards

Arne Saknussemm
04-20-2013, 07:23 PM
Dagnabbit...you know what....I just put up the result on the Bot today of the stuff I learnt in the last challenge.....like three and a half months too late LOL.....but I got a decent result in the end!!! So sure...bring it on....I'll take part... just expect my best result around Halloween ;)

Zka17
04-21-2013, 04:12 AM
He-he, I did install the 3DMark2001SE on my present system and did some readings too... - man, this is a very fun benchmark! Why nobody mentioned it to me before?! :D

I did manage to get a 823 nature score on Win7 64-bit - but that's way below the 1200 cap we have in this challenge... so, there is plenty of room to improvement! :cool:

Also, my AquaMark3 score needs to be improved...

BTW, I observed that there will be no prizes... - come on, HiVizMan, you know that the biggest prize for everyone will be the score improvements! :) Oh, and not talking about the fun we're gonna have! :cool:

Dartmaul
04-22-2013, 04:45 PM
I'm in!
It's very good that the challenge starts at 1st may since 1st week of may is a country wide holiday in Russia)
Gonna have some time to bench)

BTW Nice to see you guys)

Zka17
04-22-2013, 05:18 PM
Nice to see you too, Dart! :)

Is it still cold in Russia these times?

HiVizMan
04-23-2013, 03:23 PM
11'C In Moscow this afternoon.

kkn
04-23-2013, 04:52 PM
6+C where i live......

Zka17
04-23-2013, 05:34 PM
OK then, let's ROG'N'ROLL! :cool:

You too, kkn! :)

I already pushed the Nature score to 999.0... :D

Chino
04-23-2013, 07:15 PM
Since everyone is showing off their city's temperature, I figured I might as well come in and show off my 35C too. LOL :p

Menthol
04-24-2013, 04:25 AM
Dart,
Good to see you around. Wouldn't be a competition without you buddy

Dartmaul
04-24-2013, 08:09 AM
Currently around +3 at night.
Anyway I gain some benefit from ambient temp OCing cards, and these tests are mostly about CPU. ( I guess )

Also, like always, 3820 cannot into wprime((


But whatever happens, I won't let it harm to the spirit of competition!

Luumi
04-24-2013, 05:47 PM
Hopefully I get my other 7970 back so I can also do some dual card results :p. 7970 isn't good for 3d01 though.

twisted1
04-24-2013, 07:59 PM
I'm in, this stuff is addictive :) And the tweaking part adds a whole new dimension to it, I have lot's of learning to do!

Menthol
04-25-2013, 02:55 AM
twisted1 you came to the right place to learn, nicest most helpful people here

twisted1
04-25-2013, 09:53 AM
So I have noticed! :)

I've always liked tweaking/benching and it's great to have something like this, gives more meaning to it than just randomly running benchmarks by your self. And it's like a hundred times more fun on a team, and not just any team :D

HiVizMan
04-25-2013, 05:10 PM
Nice to see a couple of you guys are up for the tweak, sorry I mean challenge. Now if 32Mil Super PI is the Blue Ribbon event of 2D benching then 2001SE is the 3D version. This benchmark is unique in that each and every game test can be individually tweaked. And the amount of time needed to find the optimal system set up is way more than for any other bench. The trick here is finding the settings for each game test, and the memory settings that provide the best results over all. Do not think that all you need to do is have the tightest memory timings that you can manage. Run some of the tests and you will be surprised at what provides the best outcome.

On one additional thing to ponder. With two exceptions you get your best result from the first run of the game test. All subsequent running of that game test will be worse than your first run. You need to reboot between each change of settings. :)

twisted1
04-25-2013, 05:15 PM
Oh no, rebooting. Is there a reasonable way to get xp to run off an SSD? Or is it better to just go with those old disks that makes noises and suck? ;)

Zka17
04-25-2013, 05:20 PM
And the amount of time needed to find the optimal system set up is way more than for any other bench.

LOL, I can already confirm that! :) I've been for couple days now in 3DMark2001SE, but still trying to optimize... :D

One question, though... it does not seems that the GTX Titans are scaling well in this benchmark... - I mean, even using the restarting after each run, does not seems to matter how the Titans are set... absolutely no benefit to raise the GPU and/or VRAM frequencies... (there is very good scaling with CPU and system RAM tweaking)

Maybe the Titans are to new cards for this benchmark?

HiVizMan
04-25-2013, 05:26 PM
LOL no you understand why I did not care what graphic card you used, or how new it was.

HiVizMan
04-25-2013, 05:27 PM
Oh no, rebooting. Is there a reasonable way to get xp to run off an SSD? Or is it better to just go with those old disks that makes noises and suck? ;)

Oh you can install XP on to a SSD that is no problem at all.

HiVizMan
04-25-2013, 05:27 PM
Will send out the guide to all folks who post up results in the first week.

twisted1
04-25-2013, 05:44 PM
Hmm, kepp getting BSOD like 20sec into the installer. No prob installing to a regular disk tho.

Hmm I'll try another xp install and perhaps not a sandforce based SSD.

Will be posting results, in bad need of guide lol

HiVizMan
04-25-2013, 05:59 PM
LOL yip you will get that. What BIOS version are you running?

HiVizMan
04-25-2013, 06:07 PM
Here you go, you are not alone.

http://rog.asus.com/forum/showthread.php?24938-Hivizman-needs-some-help...

HiVizMan
04-25-2013, 06:07 PM
Here you go, you are not alone.

http://rog.asus.com/forum/showthread.php?24938-Hivizman-needs-some-help...

Chino
04-25-2013, 06:20 PM
Will send out the guide to all folks who post up results in the first week.
Slip my name in there, please. LOL

Zka17
04-25-2013, 06:33 PM
That 0x7B error seems to be related to the SATA being configured as AHCI... - if I set it to IDE, then it's all fine! (when I'm switching from Win7 64-bit to XP 32-bit, I always forgot to adjust the SATA settings and get that error)

Just note I'm one a regular HDD, not SSD... not sure how the SSD would behave on the IDE settings...

So, HiVizMan, are you saying that this 3DMark2001SE is more about the CPU and system RAM then the graphics card? Does it matter at all which VGA I would use? (haven't tried yet to use an older card)

HiVizMan
04-25-2013, 06:52 PM
Me, I am not saying anything. All I am doing is suggesting to folks that testing is key with this benchmark. Try stuff. And by trying stuff you will find stuff out. :D

Graphic cards make a huge difference, but not in the way you would think each and every time.

Zka17
04-25-2013, 06:56 PM
Me, I am not saying anything. All I am doing is suggesting to folks that testing is key with this benchmark. Try stuff. And by trying stuff you will find stuff out. :D

Graphic cards make a huge difference, but not in the way you would think each and every time.

OK, I got it! :cool: At least, I think so... - will see how it goes in the Challenge... :D

Dartmaul
04-25-2013, 08:14 PM
That 0x7B error seems to be related to the SATA being configured as AHCI... - if I set it to IDE, then it's all fine!
Just faced this issue today. Switch to IDE was actually my 1st suggestion)

Bot clearly shows that GTX580 is the best for 3D2001SE, but we have some restrictions so I guess VGA wont hold any of us back.

HiVizMan
04-25-2013, 08:19 PM
Easy fix for the install on some boards is to use an early BIOS.

HiVizMan
04-25-2013, 08:20 PM
Just faced this issue today. Switch to IDE was actually my 1st suggestion)

Bot clearly shows that GTX580 is the best for 3D2001SE, but we have some restrictions so I guess VGA wont hold any of us back.

No VGA power or strength will not be the determining factor here. And AM3 will balance out what ever advantage one make has over the other.

Menthol
04-26-2013, 04:25 PM
I believe you install in IDE mode, then make that change in the registry to AHCI reboot to bios and change to AHCI, have had to do that many times. (I don't believe this works for XP)

Nice clues Viz, (try stuff and you will find stuff out), finally logic I understand:)

Dartmaul
04-26-2013, 05:39 PM
(try stuff and you will find stuff out)
That's how my life goes on)

HiVizMan
04-26-2013, 06:23 PM
Did not mean it to sound so trite but it is true. The more you try things out yourself, rather than have tweaks or tricks just fed the more you learn about your system and fully understand the benchmark.

Menthol
04-26-2013, 09:09 PM
Did not take it that way, I am used to trial and error, mostly error

twisted1
04-26-2013, 10:24 PM
Tried a few bioses(on the latest 17 something one now, usually use 1408), will try the release bios. Very informational thread in general tho ;)

I the same iso's on different systems and they worked great on all of them, was almost about to break out the p8z68 pro board and test with that.

Gonna flash bios and try again, and stay of the intel 6G ports :D

Henkenator68NL
04-28-2013, 06:52 PM
Hey Guys, good this new challenge. I am definitely going to give them a try. Good to see how my O2 is going to perform. Hope to learn as much as with the other challenges (hope not to blow my psu this time though). Good luck guys

Arne Saknussemm
04-28-2013, 07:59 PM
mostly error

Well if it comes to a trial let us know ...I know some good lawyers :)

HiVizMan
04-29-2013, 11:12 AM
I will be posting up the background tomorrow and then you guys can start posting on Wednesday.

HiVizMan
04-30-2013, 03:18 PM
Background image is available.

http://www.mediafire.com/view/?zyw7at4jkkvuo0n

20248

Zka17
04-30-2013, 04:17 PM
Cool background! :cool:

HiVizMan
04-30-2013, 05:45 PM
You know who made it for us. Thanks Chris.

Arne Saknussemm
04-30-2013, 06:07 PM
Suspected as much! Thanks chrs!!

Zka17
04-30-2013, 07:30 PM
Thank you, Chris!

Dartmaul
04-30-2013, 08:27 PM
Nice wallpaper thanks! (Would be awesome to have a bigger version, I could use it for 24/7)

Since it's 1st of May in Moscow already, let me start.
------------------------------------------------------------------------------------
HW used:
RIVE - H2O cooled 3602 BIOS version
i7 3820 - H2O cooled, 5174MHz
Gskill RipjawsZ - Stock air cooling, 2354MHz 9-11-10-28
1 x HD7970 (refference) - H2O cooled, 1350/1950MHz

Windows XP x32 SP3
+17 degrees Celsius ambient temp
-------------------------------------------------------------------------------------

3DMark2001SE - 84497
20256
Aquamark3 - 333291
20257
wPrime1024m - 147.780s
20258


I'm very happy to open this contest! (would be even more happier if my HDD RAID0 wont die because of my BIOS tweaks. Again)

PS seems that I didn't done any good submissions, didn't get these tests yet, but who cares where we start, the thing is where we finish)

chrsplmr
04-30-2013, 08:29 PM
Your Welcome.
Always my pleasure.

My favorite part is when Uguyz cover 'them' up with those great numbers.
Don't deny me my joy.
Have at it !!! Best of Luck to One and All.c.

Luumi
04-30-2013, 08:31 PM
Wow the background looks awesome, let's see what kind of scores people are going to get!

Menthol
05-01-2013, 01:50 AM
It is an awesome wallpaper, thanks chris.
Dart thanks for starting, I need to get win xp installed on a drive, having trouble trying to install in ahci mode, I have done it before, I didn't want to roll back bios but if that's what I need to do, how far back will I need to go?

Dartmaul
05-01-2013, 05:47 AM
I'm using IDE mode for XP.
Can't use RAID since there's no proper RST driver for XP, so I'm using a single HDD.

Henkenator68NL
05-01-2013, 06:09 AM
Nice starting scores Dart!

I wish/hope to even come close to your starting scores ... Well its all about the learning curve!

@Chris
Nice background again !

Luumi
05-01-2013, 06:47 AM
Alright lovelies I would like to submit my starting results. As dartmaul said it is not about where we start, these are still quite mild overclocks since this is a lot more about tweaking rather than pushing every last mhz about of our components. I'm really looking forward to some tips what folks have discoverd and hivizman's guide. I couldn't tweak aquamark in anyway except getting more cpu speed, so I would like some tips on that one.

Only single card results, since my better 7970 is still out of business and I don't have other decent cards at the moment.

Hardware used:
Asus Rampage IV Extreme
Intel Core i7 3930K @ 5016
Radeon HD 7970 @ 1250/1920
Corsair Dominator GT 6gb 2000 CL8 @ 2400 7-11-7-28 1T
Seagate Barracuda 340gb HDD
Enermax MaxRevo 1200W PSU

But anyways here are my first opening results:

3Dmark01SE - 84964
20271
Aquamark - 321976
20272
Wprime - 101.765s
20273
And finally the system:
20274

I wish all you folks good luck in this competition and hopefully we all learn much when benching!

-Luumi

Henkenator68NL
05-01-2013, 07:17 AM
Like Luumi I am also anticipating to learn much more in this challenge.

I must say, non off the benchmarks are familiar to me -yet- so it will be a complete learning curve from the start.

@HiVizMan, thanks for organizing this, I have learned a lot in previous challenges! Can't wait to see your guide!

I am starting today to get some first results.

Hasta la vista,

I'll be back :cool:

HiVizMan
05-01-2013, 08:11 AM
Dart please have a look at Luumi 3DMark2001SE screen shot, that is how I want it displayed please. Your current screen covers some of the key information on the actual bench application and the details need to show the game tests. But since that is the first submission it is not a big deal.

Zka17
05-01-2013, 09:38 AM
OK, here I come... :cool:

Hardware:
Mobo - Asus Rampage IV Extreme
CPU - Intel Core i7-3970X
RAM - GSkill Pi 2x2GB
VGA - Asus GTX560Ti
Cooling - water/chiller (set at 15.5C) for the CPU and mobo; stock for the VGA

3DMark2001SE: 91850

20285

AquaMark3: 167213

20286

wPrime1024: 97.937sec

20287

System

20288

chrsplmr
05-01-2013, 11:40 AM
Link to 1920x1080 wllppr.
http://www.mediafire.com/convkey/2c24/9fndrkcorsspcaafg.jpg

Edit to Below:
ROG.c.

HiVizMan
05-01-2013, 12:07 PM
Nice ones lads and big thank you to Chris for the wallpaper back ground.

Zka17
05-01-2013, 12:19 PM
Just to clarify to myself, the "Dual Video Card" limitation actually means dual GPU - right? So two GTX295s would be already considered quad GPU and not fit in the dual card category...

HiVizMan
05-01-2013, 12:27 PM
The limitation is based on the frame rate for nature. Please feel free to use any card or combination of cards that you like.

Zka17
05-01-2013, 12:31 PM
Thank you, Sir! :) Then actually may change my hardware (graphics card) along this challenge... - as you may guess, not into Titans... :D

But first, let's see where I can get with this GTX560 Ti...

Dartmaul
05-01-2013, 06:02 PM
Dart please have a look at Luumi 3DMark2001SE screen shot, that is how I want it displayed please
My fault, I just forgot to scroll detail tab (

But you can be sure that I haven't crossed 1200fps in nature;)

Henkenator68NL
05-01-2013, 10:39 PM
Oke here are the results of my first attempt.

CPU: i7 3770K on O2
Mobo: Maximus V Extreme
RAM: 8 GB Corsair Dominator 2133MHz
GPU: Sapphire 7970 GHz on O2

3DMARK01: 91621 3Dmarks

20315


wPrime 1024M: 149.938 seconds

20317


Aquamark 332567 (GFX:76018 CPU:29546)

20316


And here is a picture of the system

20314


See you all later !

Zka17
05-01-2013, 10:42 PM
I like your Aquamark score, Henk! :)

HiVizMan
05-01-2013, 11:25 PM
My fault, I just forgot to scroll detail tab (

But you can be sure that I haven't crossed 1200fps in nature;)

Yip for sure no 1200 FPS in nature. The big thing is that you have blocked off the settings for the benchmark. The main page thing.

HiVizMan
05-01-2013, 11:28 PM
Some very nice results guys. I will prepare the guide and get it uploaded by the weekend.

Zka17
05-01-2013, 11:37 PM
Thank you, Sir! :cool:

Menthol
05-02-2013, 01:26 AM
I better jump on board even though I am not quite prepared yet. Hope to prepare a win xp disk soon

Menthol May Challenge

M/B - RIVE WC
CPU - 3970X WC @ 5100
RAM – G.Skill PI @ 2176
GPU – Titan X2 @ 836 to 1037 - Titan no benefit that I see on these benches yet

Wprime 1024 = 101.243

20328

3D2001 = 68969

20327

AM3 = 366174

20322

System as it stands today

20324

Thanks HiVizMan for hosting this challenge for us

Luumi
05-02-2013, 01:46 AM
Menthol I think you have run wprime with only 6 cores, you have to put 12 threads in the advanced options. Also you should check your pictures and swap aquamark and 3dmark01 names :P .

Menthol
05-02-2013, 04:42 AM
Menthol I think you have run wprime with only 6 cores, you have to put 12 threads in the advanced options. Also you should check your pictures and swap aquamark and 3dmark01 names :P .

Yes Sir you are 100% correct thank you for pointing that out. Tried to correct everything and can't get rid of the thumbnail.
I don't remember saying I was very smart, in case I did this proves me wrong :)

HiVizMan
05-02-2013, 10:44 AM
Nope the Titan will not make too much different for these. Unless we develop some new tweak ourself that is. You never know.

Zka17
05-02-2013, 12:27 PM
Yeap, when running the wprime is good to select all the threads... also observed that less things are running in the background is better... - and the XP will make a difference too...

Myk SilentShadow
05-02-2013, 12:34 PM
One Century; when i've finally managed to afford to build my Testbench and new Gaming rig....i'll join you guys in the challenges again lol :p :cool:

HiVizMan
05-02-2013, 12:42 PM
You can use what ever system you have currently Mike, it makes not one jot of difference. The point being is that you will learn more about how you tweak your system for these benchmarks.

They were selected because each benchmark has a very specific array of memory settings that work best. So in other words there is no one set and forget memory timings or frequency that will work best for all three benchmarks. Each one will need special testing and focus.

Myk SilentShadow
05-02-2013, 12:49 PM
I know mate, but when I build my test bench and gaming rig, my test bench is also going to be my OC bench as well....so if I kill the CPU or RAM in it, it's no big deal...because it's not my main rig :p

HiVizMan
05-02-2013, 12:53 PM
I have benched competitively for a number of years (have stopped now) and killing stuff on water or air is very difficult to do. LN2 however is quite hard on motherboards and stuff like that.

Myk SilentShadow
05-02-2013, 01:00 PM
Yeah, but i'm sure i'd slip up and kill something...hence wanting to scratch build my own test bench, so I dont have to worry about it mate :)

Arne Saknussemm
05-02-2013, 06:13 PM
Hmmm I've shot myself in the foot a bit...here. Ordered a Titan thinking it would get here before I sold my 680s but they have flown out the door in the space of a few hours so I'm sat here with a crummy laptop and a Radeon HD3450 256MBddr2! My trusty 570 I had to lend to my brother as my nephews got into the PC case and it ended in tears....

But I promise the first benchmark I run on the Titan when it gets here next week will be one of the competition benches.......well, being serious about our benchmarks, I mean after Windows Experience Index of course ;)

Zka17
05-02-2013, 06:47 PM
You still have time, Arne! We just started this Challenge... :D

HiVizMan
05-03-2013, 12:27 AM
Post up your scores with what ever hardware you have.

I can assure you that Titan will not help you one iota. :D

Not wanting to be a kill joy or anything Ollie.

Menthol
05-03-2013, 01:05 AM
I am sure this will prove to be true. I hope no one let's what hardware they have on hand stop them from participating remember this more about having fun then anything else. And I use the same rig for my everyday use, if you don't disable any hardware protection it would be very hard to hurt your system, if over current protection doesn't protect you, thermal protection most likely will

Henkenator68NL
05-03-2013, 07:11 AM
I am sure this will prove to be true. I hope no one let's what hardware they have on hand stop them from participating remember this more about having fun then anything else. And I use the same rig for my everyday use, if you don't disable any hardware protection it would be very hard to hurt your system, if over current protection doesn't protect you, thermal protection most likely will

Thats a good idea Menthol,

Ill give my AMD system a go for this Challenge to, double the hardware = double the fun... or is it double the trouble LOL

Arne Saknussemm
05-03-2013, 10:33 AM
I can assure you that Titan will not help you one iota. :D

Don΄t worry I had gathered that....the problem is the old card doesn't seem to want to work so I am displayless....bench this laptop?:rolleyes:

My joy at finally getting GK110 is not related to Aquamark so no way to kill it!:D

Menthol
05-03-2013, 11:32 AM
No display does make it hard to bench Arne

Arne Saknussemm
05-03-2013, 03:07 PM
It was a challenge yes! LOL but I managed to beat the HD3450 256MBDDR2 into shape so watch out....

Hardware

RIVE
3930k @4600MHz
16GB Dominator Platinum RAM @ 2133
HD3450 256MBDDR2

OS w7 which has grat problems with WPrime.......:rolleyes:

WPrime 856.623
20361
2001 12869
20362
Aqua 38476
20363

System....with Titan challenging GPU that needs no power connectors!

20364

Zka17
05-03-2013, 07:31 PM
Arne, have you selected 12 threads on that wprime run? It doesn't seems so...

Dartmaul
05-03-2013, 08:56 PM
Ahaha Arne rocks :D

Few weeks ago I've spend 50$ for a 3870x2, gotta use it for these tests :p

Arne Saknussemm
05-03-2013, 09:06 PM
Arne, have you selected 12 threads on that wprime run? It doesn't seems so...

Ummm nope....thanks for the heads up...totally green with these benchmarks!



Ahaha Arne rocks :D

Few weeks ago I've spend 50$ for a 3870x2, gotta use it for these tests :p

LOL....do it Dart....this powerhouse has got me 2 gold cups one silver and 1 bronze on the BOT today :D

Zka17
05-03-2013, 10:26 PM
Ummm nope....thanks for the heads up...totally green with these benchmarks!

Just try selecting 12 threads and you may got surprised... :cool:

Menthol
05-04-2013, 12:25 AM
Arne,

You finally got some new ram also, I don't know how to use these old benchmarks either, waiting for VizMan's tips, I may have to pull my 580 out of my R3E and stick it in to use, I haven't got the Titan to use more than about 30% gpu usage in these benchmarks, I think they went to sleep during the run

Arne Saknussemm
05-04-2013, 07:36 AM
Yep Zka17 that did the trick!:o

Menthol....yeah, finally got some 2133 and my CPU seems happy to run it so I have to play about with OCing and or reducing timing....lots to learn!

HiVizMan
05-04-2013, 12:03 PM
I will send off the guides this weekend, will start with the 3D benchmarks first. Wprime is about memory strap and back ground processes.

Arne Saknussemm
05-04-2013, 08:23 PM
OK so some tweaking *ahem* like turning on all cores....and improvements (how could there not be, lets face it :rolleyes:) already...

Wprime 112.414
20390
2001: 15317
20391
Aquamark: 45936
20392

This with "aquaman" tweak ? and a RAM OC......gold cup at the BOT Yay! LOL :D

Zka17
05-05-2013, 12:20 AM
The wprime is better now, Arne... now turn off the power saving features... :)

Dartmaul
05-05-2013, 09:03 AM
And try to run GPU clock faster, at least then stock freq :D

Arne Saknussemm
05-05-2013, 09:54 AM
And try to run GPU clock faster, at least then stock freq :D

LOL....I would love to but it is completely fixed at 600! I don't know how to change it....no experience with ATI cards and bios flashes....I will get round to it I guess...have to read up

And yeah Zka17 I have to set up XP and go the whole 9 yards... all the services shutdowns etc....

nikosa43
05-05-2013, 10:08 AM
Hey guys, I watch the thread from the beginning. Congratulations, seems to me that you have made a nice playground, (a kids one for gentlemen :) ). I would love to join you, but with real kids around and free time only for a few hours after midnight, it's more than difficult to cope, but I will follow your tweaking adventure.

HiVizMan
05-05-2013, 12:14 PM
Here is a basic guide

Part 1

3Dmark2001SE Benching Guide

There is a whole heap of different stuff that goes into a good 3D benchmark with hardware being the most obvious single factor for achieving higher scores. But just as hardware has evolved so too have the utilities that drive that hardware. We have at our disposal applications that can clock the CPU and GPU clocks higher while within the various Operating systems (OS), we can manipulate voltages and ram settings at will it seems, plus we can modify and set the drivers that control how the hardware function. It is the synthesis of all these different factors and one not yet mentioned, the user, that result in a top class benchmark score that can be competitive in an international area.

The hardware modifications are not the subject of this little guide, this thread is focused on those system tweaks and settings that can increase the score of the 3Dmark2001 benchmark that do not involve actual over clocking or modification of the hardware.

Background and general history

The 3Dmark2001SE benchmark is one of the most addictive and fun 3D benchmarks to do. It is not actually a true test of current graphic cards but it is certainly a good overall computer system test. The nature test being the one definite exception to that rule as it stresses the graphic card to the maximum and is used by many benchers to demonstrate the merits of a particular graphics card. Benchmarks are also used to demonstrate or test an overclocked systems stability. A good example of that is lobby high detail as it is a very good overall system test, if you system can run Lobby high detail it will run most games with no stability issues.

This is the official blurb from Futuremark (http://www.futuremark.com/download/3dmark2001)





With that out of the way let us proceed. This guide is aimed at the first time 3D bencher who like me did not have a clue of how or what to do when starting out. If other forum members have something to please jump in and tell us.

The benchmark is available in two flavors, free-user and Pro user. The difference between the two is unimportant for the purposes of this guide. The tests that are needed to return a result are the first four however the fifth test, nature, must be included to complete a valid benchmark and without that test score your grand total will look rather sad. The most straight forward method to run this benchmark is just to press the benchmark button and sit back while all the tests run their course. That is however not going to return the best score possible. Sequence of the tests run plays a part as do a number of other factors in returning a high score as we will see.

Basic stuff

There are some basic things which are often taken for granted by experienced 3D benchers and often never communicated to the novice or n00b bencher for that very reason.

1. Benchmark scores are higher in diagnostic mode. Er how do I get to diagnostic mode? That is the bit they do not tell you – most will know how to do this but not all so here is what you have to do. Press RUN > msconfig > press enter and select diagnostic mode. Reboot and there you are.

2. Set the priority of the application. Ctl Alt Del once you have opened the benchmark application, select the benchmark and select real-time. You can ignore the warning as nothing else will be open or running while you do the benchmark.

3. Close all but the benchmark application before you start your run. Have no other applications open while running the benchmark. Not even cpu-z. Open and close each tweaking application as and when you need to use them. More about those tweaking applications later.

4. The first bench that you do after a reboot will be your best score. What that means is you cannot repeat or rerun your 3Dmark2001SE benchmark over and over again in the hope your scores will be better. The best time to run a 3D mark bench is immediately after a reboot. I suggest this sequence. Run benchmark – Reboot – Run benchmark.

5. A clean install of your operating system will provide a higher score.

6. If you have a CRT monitor - not likely any more but you never know. Reset your CRT monitor to the same default resolution as that which the benchmark uses – this is done in your windows control panel and not the benchmark application. 1024 x 786, 32 bit colour even if your CRT monitor is not a widescreen monitor.

7. Run the benchmark using the default settings in the Display and CPU Settings section, top right hand panel. In other words do not change the resolution of the default test settings.

Application specific settings

1. Do not run all the tests one after another. In other words don't just press the benchmark button and let them all go one after the other.

2. Run one test at a time and in any order that works best for your system. There are a number of reasons for this practice, one being the stress the different game tests put on your system. The Nett result of stress (load) is normally heat. This is true for both the CPU and the graphics card. So by running one game test at a time you prevent any stability issues due to heat by giving your system an opportunity to cool during the selection and application of specific game test tweak. This is most important when doing a benchmark run at near to or at the limit of your systems capability. Method of selecting Game Test Using the change option found in the selected tests part of the main panel open the test selection and toggle custom. There you will be able to select any one of the game tests, such as Lobby - Low Detail or Car Chase – High Detail for example. Do not forget to clear all the tests you do not want to run. There are other reasons such as test sequence has an effect on the scores returned.

3. Sequence of the tests is important for a high score. Each of the top benchers has their own preference of sequence. However there is one that can be done last generally – that is Car Chase – Low Detail. The reason is that this game test is the one that can be done repeatedly in the hope of a “golden” run. I do not know the actual reason why this is but I suspect it is an anomaly of the software whereby there is one run that has less physics to do that all the others. I normally bench the game test that is the most difficult (stressful) to my system first. In my case that is often Nature. But I would suggest that Car Chase – High Detail be run relatively early on in the sequence as for some reason best scores are achieved if run early and not near the end. This is the sequence that is most commonly used.

• Lobby – Low Detail
• Gragothic – Low Detail
• Car Chase – High Detail
• Dragothic – High Detail
• Nature
• Lobby – High Detail
• Car Chase – Low Detail

HiVizMan
05-05-2013, 12:14 PM
Part 2

LOD graphics driver Tweaks

Level of detail = LOD. The higher the positive number the less work the graphics card has to do. That is the simplistic reason why LOD works. However like most things in this world it is not that simple. Some of the game tests befit more than others to a high LOD setting. Nature being the most obvious as it is the test that relies on the Graphics card the most. But if the LOD setting is too high performance is adversely affected, so just the right setting is needed for optimal performance (that is another way of saying highest score).

There are a number of methods of setting the LOD but the most common is to use an application like Rivatuner or AtiTOOL and manually set the LOD for each test. Yes you heard right, each of the 9 game tests has a particular LOD setting that works best for that test. And to make it even more complex each graphics card has its own ideal LOD settings for each test. Actually every time you add or remove a component of your system, you will need to redefine what settings will be optimal for your system all over again. So unfortunately there is no universal or wonder method with the LOD tweak. Like all the other things that contribute to a top 3D benchmark testing is again the order of the day to see which settings work the best. And with this particular benchmark that means having to reboot after each individual test. So while not all might like the use of LOD tweaks they are for sure not a free lunch at all. But do not despair at the notion of spending the next few months of your spare time having to test each and every LOD setting as there are general guidelines that narrow down the testing. The top 3D benchers in the world do spend hours and hours testing settings. Please understand that each graphics card will be different and that individual cards of the same brand will differ slightly too.

• Lobby – Low Detail LOD 5
• Gragothic – Low Detail LOD 3.5
• Car Chase – High Detail LOD 5
• Dragothic – High Detail LOD 5
• Nature LOD 4
• Lobby – High Detail LOD 3.5
• Car Chase – Low Detail LOD 3

Please note that each graphic card has a unique set of optimal LOD settings. So you have to work out what is best for your card. These are what was best for my card at that time.

Nature LOD 0

http://img118.imageshack.us/img118/5140/lod0rz3.jpg

Nature LOD +3


20398

So for example I had a HD2900XT graphics card and my testing has shown me that the first game test has the best result with a LOD setting of 4.8. I simple tested Lobby – Low Detail ten times at each LOD setting between the range LOD setting 4 and 6. But that was specific to a particular motherboard and a particular CPU and ram. I suspect that I will have to retest when I replace my current system. That meant nearly 200 reboots to work out one test LOD setting.


Next up - How to change the LOD setting using the different tools the images are pretty self explanatory so words are just a wast.



http://img516.imageshack.us/img516/1428/rtuner2ml7.jpg

http://img528.imageshack.us/img528/2793/rtuner3gv1.jpg

http://img528.imageshack.us/img528/1408/atitool1xv8.jpg

http://img516.imageshack.us/img516/1081/atitool2vb6.jpg




***** WARNING ******

Midmapping using ntune or some other similar application is a cheat and will be picked up by HWBOT members or staff. Do not use it please, the world has enough cheaters and scum running countries lest not add to that number.

Remember benching is and can be fun - when the points you get for a bench become more important than the activity rethink why you are doing what you do.

Dartmaul
05-05-2013, 03:34 PM
Little update

------------------------------------------------------------------------------------
HW used:
RIVE - H2O cooled 3602 BIOS version
i7 3820 - H2O cooled, 5166MHz
Gskill RipjawsZ - Stock air cooling, 2352MHz 9-11-10-28
1 x HD7970 (refference) - H2O cooled, 1350/1950MHz

Windows XP x32 SP3
+17 degrees Celsius ambient temp
-------------------------------------------------------------------------------------
3DMark2001SE - 85906
20399

Aquamark - 333376
20400

wPrime - 148.484
20401


Used the tips from Sir HiVizMan, got 1400 more points in 3Dmark. It's an improvement but not as much as I expected it would be.
Currently I see no way for me to improve my scores (expect of squeeze some more mhz from the CPU). That's sad.
Top1 3DMark2001SE score on HWBot is like twice more then I get, (I know, better mem, LN2 cooled 3770k, magic GTX580 etc) but I newer was so far from the best results.
If there's some features that can greatly improve my score - I would try to find some. If not - I just need some other HW.

PS I'm gonna fix my 9800GX2 and bring from workplace 3870x2. Hope one of these (Probably 9800GX2) would buff me.

Zka17
05-05-2013, 04:02 PM
Dart, have you run each part of the 3dmark01 separately? Have you found the optimal settings for each part?

Just a hint - the highest clocks are not always the best for each part...

Dartmaul
05-05-2013, 05:14 PM
Dart, have you run each part of the 3dmark01 separately? Have you found the optimal settings for each part?
Yes. Off course I can adjust some settings more accurate but seems that it wouldn't give me my good improvement.
7970 card gives more VGA power then the test can take. That excludes GPU tweaking from the list of significant factors. I don't like graphic tests where GPU doesn't makes any noticeable difference.

Gonna use some older GPU to make tweaking more fun.

HiVizMan
05-05-2013, 05:50 PM
I will put up the first week score board later tonight.

Arne Saknussemm
05-05-2013, 07:36 PM
Nice one HiVizMan....if that lot doesn't help us nothing will! :)

Henkenator68NL
05-05-2013, 10:24 PM
Thanks for the Guide HiViz!

On to new round ! Luckily I found my 2 old 3870 cards .. back in the days they where my first crossfire set... sighhh

nikosa43
05-07-2013, 08:46 AM
Hi guys, as I said before to me is almost impossible to find time to to join properly but I couldn't resist to run some of the benchmarks with my computer, as it is (Win 7 64, with everything running as normal). So I discovered something that can boost 3dmark more than 10%. HiVizman is the one can give permission to share or not :).

Zka17
05-07-2013, 09:07 AM
nikosa, PM your finding to HiVizMan - he can decide if it's a legit tweak or not...

Also, I would be interested in such tweaks for Aquamark... :D

nikosa43
05-07-2013, 10:45 AM
nikosa, PM your finding to HiVizMan - he can decide if it's a legit tweak or not...

Also, I would be interested in such tweaks for Aquamark... :D

OMG, I was so excited for my finding that I forgot an important factor. I will send HiVizman a PM but my finding is totally wrong. On the second run I forgot that I changed two things, not one to have comparable results.:mad:

flexnl
05-07-2013, 11:01 AM
hi all
i screwd up my 3770k with delid...and it overheating now but wanted to send my scores aNYWAY :p
its with an Gigabyte gtx 680oc on Water
windows 8 64

PRIME 161.64
20434

AQUAMARK 370052
20435

3dMARK2001 77532
20436

my rigpic isnt up to date but new ones are coming :D

20438

nikosa43
05-07-2013, 11:14 AM
Hey flexnl, what you screwed with delidding? May I can help.:)

flexnl
05-07-2013, 11:34 AM
Hey flexnl, what you screwed with delidding? May I can help.:)

well...i delidded it and applied coollaboratory pro on die and al was good.. temps also good
but then i got the ideea that i had aplied to much so i did it again and in that process i apperantly did some damage
it isnt visible but my temps are 45x 1.2v full load 100c. :(
cooler is seated good and lid also..
i think its a lost cause :(

nikosa43
05-07-2013, 01:18 PM
well...i delidded it and applied coollaboratory pro on die and al was good.. temps also good
but then i got the ideea that i had aplied to much so i did it again and in that process i apperantly did some damage
it isnt visible but my temps are 45x 1.2v full load 100c. :(
cooler is seated good and lid also..
i think its a lost cause :(

I think you haven't damage anything. Coollabs is conductive, as a liquid metal, so if you spilled it somewhere you could have a shortcircuit. This is not the case. From your temps I think that something small as a piece of dust or hair is sitting between Heatspreader and the chip. Have you tried to completely remove Coollab and make a good inspection of the surface (both chip and underneath Heatspreader? You can PM me if it is not right to make the discussion here.

HiVizMan
05-07-2013, 01:27 PM
OMG, I was so excited for my finding that I forgot an important factor. I will send HiVizman a PM but my finding is totally wrong. On the second run I forgot that I changed two things, not one to have comparable results.:mad:

You do not need to worry you are not the first person to have done that and you will not be the last. And for your information if it had worked it would have been quite legal.


Regards

flexnl
05-07-2013, 03:54 PM
I think you haven't damage anything. Coollabs is conductive, as a liquid metal, so if you spilled it somewhere you could have a shortcircuit. This is not the case. From your temps I think that something small as a piece of dust or hair is sitting between Heatspreader and the chip. Have you tried to completely remove Coollab and make a good inspection of the surface (both chip and underneath Heatspreader? You can PM me if it is not right to make the discussion here.

yes i cleaned it and the die is absolutly spotles mirror surface...and again and again and then i ran outta coollab so i got some other paste on it.
also.....all 4 cores are hotter and need more juice then before ...but 1 core is stickin out with 105 when the others are at 85-90
20440

Dartmaul
05-07-2013, 04:32 PM
Well, I brought my 3870x2 back from job, but since I've left it there the fans bearing has cracked and fan started vibrating.
So I bought another 3870x2 that was presented as broken, just as a donor fan. But it actually wasn't broken.
Then I suddenly realized that I have stock air cooling from my 7970's that has exact same fan as on 3870x2 (same but better).

So here are my babies, clean and with improved cooling :D
20441

I was always very very bad at painting, couldn't event drew a circle, so I'm very proud of this hand drawn ROG logo!

HiVizMan
05-07-2013, 04:34 PM
Could one of you go through the results for 3Dmark2001 and tell me if anyone has a higher nature than 999 at present and if yes link me to the post please.

Arne Saknussemm
05-07-2013, 04:42 PM
No, no one higher than 999 :) I was close:rolleyes:.......not! LOL



Dart! Nice find with the cards!

HiVizMan
05-07-2013, 04:49 PM
OK in that case I would like to make one change please, and that is to cap the nature score at 1000MHz.

I will only do so once three members reply positively and there is not one negative vote.

Dartmaul
05-07-2013, 04:54 PM
nature score at 1000MHz.
HiViz is so addicted to overclocking so he measures benchmark score in MHz!

Arne Saknussemm
05-07-2013, 05:14 PM
I don't mind....I guess the aim is to limit my beast Radeon3450 from dominating outright...but I can be sportsmanlike and give the others a fighting chance....

HiVizMan
05-07-2013, 05:23 PM
HiViz is so addicted to overclocking so he measures benchmark score in MHz!

Darn that is so true, my bad, FPS

flexnl
05-07-2013, 05:44 PM
OK in that case I would like to make one change please, and that is to cap the nature score at 1000MHz.

I will only do so once three members reply positively and there is not one negative vote.

Yess :) great idea

HiVizMan
05-07-2013, 05:55 PM
Dart you agree or not?

Zka17
05-07-2013, 07:17 PM
I have reached 999 points in Nature once in my life - with a single Titan... and if I remember correctly, that was with a lot of tweaking...

So, I don't think that would change much if the Nature score limit is reduced...

Dartmaul
05-07-2013, 07:43 PM
Dart you agree or not?
I'm so far from even 1k so this change wouldn't affect me in any way)
Also I haven't learned the test good enough yet so I can't understand how exactly would this change affect the whole challenge.

So yes I agree.

HiVizMan
05-07-2013, 08:35 PM
Thank you guys. The rule is now amended and the nature test is capped at 1000 FPS.

Now that we have agreed to that take a look at this.

20452

This CPU will pass Wprime at 5.724 on air. So look at what hardware you have and be creative.

http://img.hwbot.org/u31829/image_id_582728.png

Dartmaul
05-07-2013, 09:07 PM
I guess we all wait for some educating part of the contest...

PS If you have any guide how to run a CPU at 5.7GHz without subzero cooling - I would like to take a look.

Zka17
05-07-2013, 11:27 PM
Yeah, those good old SB CPUs... I have a 2700k... - too bad I don't have a mobo for it... :(

But that Radeon HD 4800 series GPU is interesting! :cool: 400k in Aquamark, wow!

Will we get a guide for that Aquamark too? Please, pleae, please...

HiVizMan
05-08-2013, 11:58 AM
I guess we all wait for some educating part of the contest...

PS If you have any guide how to run a CPU at 5.7GHz without subzero cooling - I would like to take a look.


Dart the SB processors will run at their max on air. They will not last all that long but I first boot to 57 multi on a Pro Mag air cooler. This CPU is one of the very first batches that came out. It was bought retail on the first day they were available.

All day every day at 5.7 on water. Never gets above 68'C


Yeah, those good old SB CPUs... I have a 2700k... - too bad I don't have a mobo for it... :(

But that Radeon HD 4800 series GPU is interesting! :cool: 400k in Aquamark, wow!

Will we get a guide for that Aquamark too? Please, pleae, please...

Yes indeed there are many tweaks for Aqua and additional tweaks for 3DMark2001 too. All in good time.

Dartmaul
05-08-2013, 12:16 PM
All day every day at 5.7 on water. Never gets above 68'C
Holy f***!
I've chosen wrong platform(
I can barely handle my 3820 at 5.2 with 1.6 vcore on water with +17 ambient((

HiVizMan
05-08-2013, 12:40 PM
I do think that you have to trade off the wprime result with the 3D result. You see all three benchmarks are about CPU not graphic cards. That is the lesson contained in this challenge for the forum. Two might be located in the 3D section but they are actually CPU benchmarks. Oh and ram.

Use LH as your bench for your optimal ram settings and that will help you get a higher score. Work out which sequence works best for your system.

Zka17
05-08-2013, 12:47 PM
Use LH as your bench for your optimal ram settings and that will help you get a higher score. Work out which sequence works best for your system.

What that "LH" is?

HiVizMan
05-08-2013, 12:53 PM
Lobby High. Is the most useful of the game tests for memory. Your best result memory wise with LH is normally the most effective over all.

One other hint, look at the screenshot title for Aquamark.

Zka17
05-08-2013, 12:58 PM
Oh, I see - thanks! :cool:

BTW, you change memory settings too during the 3dm01se runs? Or that is set in BIOS and you just adjust graphics settings among different games?

Luumi
05-08-2013, 04:03 PM
Hivizman running 3dmark01 with the the computers at the Heathrow airport :D http://upload.wikimedia.org/wikipedia/commons/9/93/Blue_screen_dfw.jpg

I have to make some more runs today. I've been quite busy the last few days part from testing the memory.

flexnl
05-08-2013, 05:02 PM
shavedd off a bit...

still with ****ty hot 3770k @ 4600mhz and 105 c :P


prime 160.88
20468

aquamark 373798
20469

3dmark2001 78606
20470

HiVizMan
05-08-2013, 05:43 PM
Oh, I see - thanks! :cool:

BTW, you change memory settings too during the 3dm01se runs? Or that is set in BIOS and you just adjust graphics settings among different games?

You loose FPS if you change the memory while the benchmark is open. So you have to find what is the best overall memory settings for all the game tests of 2001SE, and then what is the best settings for Aqua which are very different. And of course Wprime has its own unique settings too.

Zka17
05-08-2013, 11:26 PM
OK, that's what I thought... thank you, Sir!

flexnl
05-09-2013, 05:07 PM
I have a question about the LOD setting...
i installed rivatuner but i think it is such an old program that my driver 314.22 isnt supported
is there any other way to adjust these settings?
i allready tried to set forcedriverversion to 31422 in power user tab...
still nothing :(

HiVizMan
05-09-2013, 05:33 PM
nvidia inspector is the application you are now looking for or setLOD for AMD

http://www.techspot.com/downloads.php?action=download_now&id=5077&evp=8f5198313ef42ef50e0842a19955b609&file=1

This is a older version but has a greater range of LOD useful for other benchmarks possibly.

http://www.mediafire.com/download.php?0zagdsscb407cib

SetLOD6

http://www.mediafire.com/download.php?ik7hlhyrom11abk

flexnl
05-09-2013, 05:49 PM
okey tnx, you rock! :P

HiVizMan
05-09-2013, 05:50 PM
I have a question about the LOD setting...
i installed rivatuner but i think it is such an old program that my driver 314.22 isnt supported
is there any other way to adjust these settings?
i allready tried to set forcedriverversion to 31422 in power user tab...
still nothing :(


That guide was written in 2007 and is meant to be a pointer of where and how you are meant to look.

I have posted up the latest software that you can use in place of Riva

flexnl
05-09-2013, 06:18 PM
another question
when you run a single game only to determin best LOD setting...you still need to reboot after every run?

Zka17
05-09-2013, 06:19 PM
Hm... NvidiaInspector is working with the AMD/ATI cards too? I had never tried that...

Right now I'm pushing a 3870 X2 and using ATI Tray Tool for adjusting the LOD... (for the newer AMD cards I have, the Matrix ones, I never even tried to set LOD)

Zka17
05-09-2013, 06:21 PM
another question
when you run a single game only to determin best LOD setting...you still need to reboot after every run?

I do restart the system... but with WinXP is pretty fast...

HiVizMan
05-09-2013, 06:37 PM
Nope only with nVid that is why there is SetLOD6

HiVizMan
05-09-2013, 06:46 PM
another question
when you run a single game only to determin best LOD setting...you still need to reboot after every run?

Scores are lower when you run the game test second time. You have one shot at it each boot.

HiVizMan
05-09-2013, 06:46 PM
another question
when you run a single game only to determin best LOD setting...you still need to reboot after every run?

Scores are lower when you run the game test second time. You have one shot at it each boot.

Zka17
05-09-2013, 06:55 PM
Nope only with nVid that is why there is SetLOD6

LOL, thanks!

At my first interpretation I thought that you want me to set LOD to 6... - but got it now! :cool:

flexnl
05-09-2013, 08:47 PM
tnx guys ...now i go figure out how to set different lod valeus for the different games |:P
or do you guys change it during benching manualy?
also.... i understand you can adapt the sequence of the parts of 2001se?
cant find how to do that eather :)

btw dagmar... you have found the time to look for that chip by any chaNce? :D:D

Zka17
05-09-2013, 08:55 PM
Well, simply open 2001se, then select your first game to run... then CTRL+ALT+TAB and set real-time priority...

Then you open your VGA tweaking application and set the clocks - and close it! Then set your LOD and close that application too...

Then run your first game... - and repeat from the VGA tweaking + LOD setting for each game... Do not close the 2001se during this, just select the different games... when you run all the games you will get the final score...

But, to get to this part, I would run each game individually until I find the optimal settings... restarting the whole system between running the same game again...

twisted1
05-09-2013, 09:02 PM
Is it ok to post results from W7, just so I can get some results in until I get Aquamark working in XP?

I think I need another xp install, I installed 64bit and long story short, no good :)

Zka17
05-09-2013, 09:04 PM
I think, it's OK... W7 just not as good as XP for these benchmarks... - but it's up to you which one you choose... :)

flexnl
05-09-2013, 09:07 PM
Well, simply open 2001se, then select your first game to run... then CTRL+ALT+TAB and set real-time priority...

Then you open your VGA tweaking application and set the clocks - and close it! Then set your LOD and close that application too...

Then run your first game... - and repeat from the VGA tweaking + LOD setting for each game... Do not close the 2001se during this, just select the different games... when you run all the games you will get the final score...

But, to get to this part, I would run each game individually until I find the optimal settings... restarting the whole system between running the same game again...

cool ...tnx m8
and yess im rebooting like crazy now gnegne
but windows 8 isnt a bad booter also :)

HiVizMan
05-09-2013, 09:11 PM
You do not change the sequence of game test.

You select one game test at a time.

You set LOD for each game test.

I tend to let my CPU cool down for about a minute of two before I run the next game test for those on air.

And yes I am nearly done. You would not be happy with me if I sent you just any old CPU now would you?

HiVizMan
05-09-2013, 09:12 PM
cool ...tnx m8
and yess im rebooting like crazy now gnegne
but windows 8 isnt a bad booter also :)


XP is king.

Luumi
05-09-2013, 09:28 PM
He wants your 5.7ghz sandy hiviz :P .

flexnl
05-10-2013, 12:56 AM
He wants your 5.7ghz sandy hiviz :P .

HELL YEAHHHH ! :d

flexnl
05-10-2013, 01:00 AM
You do not change the sequence of game test.

You select one game test at a time.

You set LOD for each game test.

I tend to let my CPU cool down for about a minute of two before I run the next game test for those on air.

And yes I am nearly done. You would not be happy with me if I sent you just any old CPU now would you?

HAHA didnt want to rush you boss :D
some things i got alot of, time is one of them :)

Henkenator68NL
05-10-2013, 08:36 AM
Hey guys, I seem to have hit some strange bumb. 1 can run my 3770k cpu @5000 and @5100. It nicely passes aquamark and 3dmark01. But with wprime something is not working. But I can't find the faulty setting/cause. what I try again and again I start by trying to get 32M d several runs. Mostly there already comes the problem that it rounds off 4-6 threada with 0 errors, but rest does not finish .... Cant from find the reason. lTemps max till about 80-85 degrees. tried dossens off voltage setting. System always boots. It runs prime95. But wprime it doesn't. I have not even began to OC or tweak my Ram yet. Any help would be mre than welcome. Thanks !!!

Zka17
05-10-2013, 08:46 AM
Henk, wprime 1024M is stressing the CPU and RAM much harder than 32M... so, if you get 80-85C with 32M, I would expect way higher temps for 1024M... - you may experience throttling because of high temps there...

twisted1
05-10-2013, 12:00 PM
I think, it's OK... W7 just not as good as XP for these benchmarks... - but it's up to you which one you choose... :)

Probs. with Xp, Aquamark wont work. I have time to fix it today tho, new xp install and all that good stuff.


Henkenator68NL: Like Z. said, wprime really puts intense stress on cpu/mem. My 3770k seems pretty stable at the same clocks as it is stable for the physics tests in 3dmark11.

HiVizMan
05-10-2013, 12:25 PM
Aquamark works just fine with XP I can promise you.

Please do use the HWBOT wrapper.

twisted1
05-10-2013, 01:06 PM
Yes, I know it does. Not on my bugged xp 64bit install though.

Changing to regular 32bit sp3. Yes always the HWBOT wrapper! Yes Sir! :D

HiVizMan
05-10-2013, 02:03 PM
Ah 64bit - now I understand.

twisted1
05-10-2013, 02:31 PM
Should have read the readme lol :)

Luumi
05-11-2013, 02:53 AM
Would anyone with information be willing to give me some tips so I can run aquamark properly on xp. For some reason my scores in xp are very bad compared to what I get in windows 7. Using my everyday overclock settings I get only 290-295k in Windows XP, whereas I get 340k or more in Windows 7. Am I missing something important on the OS itself, since this is pretty frustrating.

Chino
05-11-2013, 03:18 AM
These are some of the basic tweaks that I use.
1. Rename the executable. I use AM3.
2. Set the priority to Real Time.
3. Set the affinity to all cores.
4. Kill explorer before you run it pretty much like SuperPi.

HiViMan can probably reveal some of his super duper secret tricks that only elite benchers would know. ;)

Menthol
05-11-2013, 06:32 AM
HiVizMan,
I am having trouble with 2001, I can't get Nature under 1050 points no mater what I do, I tried under clocking my GPU, enabling all graphics settings, lowered priority, nothing works. Could you please raise the Nature score a couple hundred points please, or tell me what I need to do to lower that score. Also LOD does nothing for me, I think it has to do with the Titan driver and Win XP32. No sli setting available either, not that I need 2 cards, just asking if I am missing something.

Menthol
05-11-2013, 06:45 AM
RIVE
3970X @ 5.2
G.Skillpi@2400
Titan
GPU@1103
Mem@1677

AM3 = 389746

20541

wprime 1024 = 98.266

20542

3D2001 = 95248

20602

Dartmaul
05-11-2013, 08:16 AM
Damn, it's massive!

PS
I guess we can restore nature restriction back to 1200fps)

HiVizMan
05-11-2013, 09:56 AM
I have a question, what would be the best system for these benchmarks, z77 with a sandy bridge or ivey bridge processor. Is either one a better choice than sandy bridge e. It seems any modern gpu is ok.

This is a very common question and I am surprised that no-one has asked it yet. :cool:

I am not going to answer directly because I have not tested, instead I am going to tell you how to find out.

Look at results elsewhere - same frequency different CPU's you will notice large differences between Game Test scores. Why is that? Simple, one graphic card produces much higher Nature results, another lower Nature results. One graphic card produces higher scores in all the other game tests except Nature, but since Nature is worth 20x the FPS as a final score you have to balance out what card to use with a particular CPU.

I would think that IVY will be more efficient than SB or LGA2011 but I have not tested this. Why is SB my first choice simply because you can clock it much higher than Ivy on water. But as I said I have not tested so it is only a guess.

HiVizMan
05-11-2013, 09:59 AM
HiVizMan,
I am having trouble with 2001, I can't get Nature under 1050 points no mater what I do, I tried under clocking my GPU, enabling all graphics settings, lowered priority, nothing works. Could you please raise the Nature score a couple hundred points please, or tell me what I need to do to lower that score. Also LOD does nothing for me, I think it has to do with the Titan driver and Win XP32. No sli setting available either, not that I need 2 cards, just asking if I am missing something.


LOL Don't make your problem our problem. :D


Increase your LOD to positive numbers or have some other application running in the back ground when you do the nature test. This is all part of the tweaking and understanding how to get the best of your system. If you know what will slow it down, then you will also know how to make it more efficient.

And which Inspector are you using to set LOD?

Menthol
05-11-2013, 12:16 PM
Sorry I was going to change that post but didn't get here before you. I will figure it out. I had just finally got win xp installed and first time I had a wprime score under 100 and 2001 was my last bench for the night. I have been using ver. 1.9.6.6 that's been modified for very high LOD adjustments for Vantage. Rules are rules, I will abide.
Thanks

Luumi
05-11-2013, 12:20 PM
Nope, still the same chino. Thanks for your tips anyways. The cpu score just stays horribly low in xp no matter what I do. I used my identical 24/7 bios profile with just sata mode turned to IDE and ran only XP drive. THe total score just stays at 290k-295k. I've tried both 1.0 and 2.25 versions of hwbot aquamark and tried the aquamark patch aswell, but it didn't make any difference.

Am I missing something important on the OS itself or is XP just bad for aquamark compared to Windows 7? It feels very frustrating that 5.2ghz 3930k and 2500 cl8 memory in xp is as fast if not slower than a 4.6ghz 3930k with easy memory settings in 7.

HiVizMan
05-11-2013, 12:22 PM
Sorry I was going to change that post but didn't get here before you. I will figure it out. I had just finally got win xp installed and first time I had a wprime score under 100 and 2001 was my last bench for the night. I have been using ver. 1.9.6.6 that's been modified for very high LOD adjustments for Vantage. Rules are rules, I will abide.
Thanks


Just having a bit of fun with you Menthol it is quite easy to slow down nature by the way.

HiVizMan
05-11-2013, 12:25 PM
Nope, still the same chino. Thanks for your tips anyways. The cpu score just stays horribly low in xp no matter what I do. I used my identical 24/7 bios profile with just sata mode turned to IDE and ran only XP drive. THe total score just stays at 290k-295k. I've tried both 1.0 and 2.25 versions of hwbot aquamark and tried the aquamark patch aswell, but it didn't make any difference.

Am I missing something important on the OS itself or is XP just bad for aquamark compared to Windows 7? It feels very frustrating that 5.2ghz 3930k and 2500 cl8 memory in xp is as fast if not slower than a 4.6ghz 3930k with easy memory settings in 7.

If Win7 works for you then use that. See what 5.2GHz in Win7 will give you.

Remember that in XP many settings need to be set for you to get the best out of your system. In Win7 some of them are already enabled. This is a very testing benchmark by the way and not as simple as most folks seem to think.

HiVizMan
05-11-2013, 12:25 PM
Ram is key in Aquamark and the settings are not the same as say 32Mil or even 2001SE

HiVizMan
05-11-2013, 01:43 PM
Some more Aquamark tweaks.

First rename your aquamark .exe to crysis.exe in the programe folder.
When you select the benchmark from the wrapper you will have to direct the wrapper to find the renamed .exe


Now open the d3d overrider that comes with rivatuner, open it & leave it leave it running in systray while benching, with global highlighted set detection high, triple buffering on, & vsync off.

Make sure you have your affinity correct and the applicable amount of LOD and you will see the difference.

There are a few more too. :p

Fermi cards will need to change nvidiainspector setting to this. Look in hidden settings.

nikosa43
05-11-2013, 09:33 PM
Well guys, I couldn't resist. I dig up my old trusty Raptor and I installed WinXp 32 Sp3. I don't say it was the easiest thing but finally it worked. No matter the results it's an addictive thing and as Arne says, I will go as a sportsman :) .So now ...What I have to do to join the challenge? Later this night maybe I can post first results.

nikosa43
05-11-2013, 11:15 PM
As I said before here is the first results. I have to say, I am impressed from Winxp. That's why HiVizman says this OS is king. I am sorry I didn't use the wallpaper. I 'll download it for the next ones. The scores are with my 24/7 setup as I use it in win 7 64. The only thing changed is the OS.

3dmark 2001 95281

20567

Aquamark 348948

20568

Wprime 165,531

20569

Zka17
05-11-2013, 11:24 PM
Cool, nikosa! :cool:

Also, pay attention to the Nature score limit we have in this challenge...

nikosa43
05-11-2013, 11:57 PM
Cool, nikosa! :cool:Also, pay attention to the Nature score limit we have in this challenge... I have to reduce it... Wow. In Win7 I got a score of 680 in nature and with XP I have to lower the fps. So here comes all the fun :)

Menthol
05-13-2013, 01:07 AM
HiVizMan,
Have you posted any scores yet and I am just to blind to see them?

HiVizMan
05-13-2013, 04:26 PM
HiVizMan,
Have you posted any scores yet and I am just to blind to see them?

Nope sorry I have been a touch busy running around with family stuff. I will get to it ASAP.

Menthol
05-14-2013, 01:47 AM
No hurry, take your time, just curios as to what scoring will look like

nikosa43
05-14-2013, 02:05 AM
Here are some scores. Of course in the attempt to raise the clock I had to sucrifice wPrime which is multithreaded. I am sorry for the appearence of wPrime but I don't know why, when it finishes it stays there and doesn't show the normal screen.

3dmark 2001 : 110375

20623


Aquamark : 409388


20624


wPrime 1024 : 190,578


20625

HiVizMan
05-14-2013, 07:32 AM
Hello nikosa43 - did you download the latest version of wprime from the HWBOT site? And rememberer to change the number of cores to reflect your total number or possible threads and in your current OC set up that is 4.

HiVizMan
05-14-2013, 07:34 AM
No hurry, take your time, just curios as to what scoring will look like

The key is the bonus points. So those folks who have put up entries in both the first two weeks each have 10 bonus points right now. Add that to the totals of each benchmark leader board and you have your total.

Actually Dart or Menthol could you help me out here and put together the leader board for each bench please.

Cheers

Dartmaul
05-14-2013, 07:50 AM
Not this time man((
I'm about to face the upcoming exams session so I have no time even to update my own submission((

I'm sorry(

Zka17
05-14-2013, 12:25 PM
In couple hours I may be able to make an Excel file with the results submitted... - just let me know where to send it...

Menthol
05-14-2013, 12:33 PM
After work tonight I'll enter all the scores into an excel spreadsheet
Zka,
If you have the time I will b glad to let you do this, if there is anything I can do please let me know

HiVizMan
05-14-2013, 01:50 PM
Cheers thanks guys. Much appreciated.

nikosa43
05-14-2013, 03:06 PM
Hello nikosa43 - did you download the latest version of wprime from the HWBOT site? And rememberer to change the number of cores to reflect your total number or possible threads and in your current OC set up that is 4. Hi HiVizman. Yes the version is downloaded from HWBOT. For some weird reason is doing this. I was very tired last night to find why. The pathetic score is with 4 cores selected. Unfortunately for me I don't think I can go that high on air with 8 threads. I will search a bit the RAM settings but I don't believe it will have a huge impact. The way for prime is lots of threads. Besides I hate it. It has the less fun :) .

nikosa43
05-14-2013, 04:11 PM
Hey guys, I renamed Aquamark to Am3 but I think that somehow reduced the score, so used the original one. Is anyone knows if that works and why?

HiVizMan
05-14-2013, 04:13 PM
rename to crysis.exe

nikosa43
05-14-2013, 04:57 PM
rename to crysis.exe Thanks, I 'll try this. But how it works? It takes any settings from NVIDIA inspector? Because in Aquamark I didn't use it for any other settings, other than rev up the clock of the GPU a bit. Also I found out Priffinity, it helps to set affinity and priority in batch to the running services. Maybe most people know but until now never used or needed such an application.

Zka17
05-14-2013, 07:55 PM
nikosa43, when the wprime can not display the results in the original window, it may mean that your system is over-stretched... see in others submission how it should look like... - try to reduce the clock or increase the Vcore...

Zka17
05-14-2013, 07:59 PM
After work tonight I'll enter all the scores into an excel spreadsheet
Zka,
If you have the time I will b glad to let you do this, if there is anything I can do please let me know

Thanks, Menthol, so far I did it and sent it... :)

Does somebody know why I'm not getting email notifications about new posts in subscribed threads? Not even the notification for the PM came through... - before I use to get everything, but these last days I'm always amazed how many posts are in my favorite threads whenever I sign in...

HiVizMan
05-14-2013, 09:29 PM
Thanks, I 'll try this. But how it works? It takes any settings from NVIDIA inspector? Because in Aquamark I didn't use it for any other settings, other than rev up the clock of the GPU a bit. Also I found out Priffinity, it helps to set affinity and priority in batch to the running services. Maybe most people know but until now never used or needed such an application.


How it works is simple, Aquamark gives the best scores with affinity set to multiple processors. XP pro does not do this automatically. Win7 64 bit does. However Win7 is not the best option most of the time for Aquamark. So when you change the exe file to Crysis what you do is force your graphic driver and system to think that the application is a multicore application. Hence the increased scores.

Another tip - work out what resolution is best suited for each benchmark. And what quality of picture is best suited.

nikosa43
05-15-2013, 12:37 AM
nikosa43, when the wprime can not display the results in the original window, it may mean that your system is over-stretched... see in others submission how it should look like... - try to reduce the clock or increase the Vcore... Zka17, thanks for the answer I will check as soon as possible. Last night I had to do all these after midnight, after work...after lots of anything else. Imagine, find and test some clocks and test how they work, if they work and post some results. I wonder who is more stretched, me or my CPU? :D

nikosa43
05-15-2013, 01:09 AM
How it works is simple, Aquamark gives the best scores with affinity set to multiple processors. XP pro does not do this automatically. Win7 64 bit does. However Win7 is not the best option most of the time for Aquamark. So when you change the exe file to Crysis what you do is force your graphic driver and system to think that the application is a multicore application. Hence the increased scores. Another tip - work out what resolution is best suited for each benchmark. And what quality of picture is best suited. HiVizman, this makes sense in my mind, because I had to think that magic was another variable I had to add to the equation of May Madness. But you made a mistake. May Madness had to be Winter Madness Challenge. It's so many things and so many combinations I want to try, that unfortunately is impossible to do in one month. Especially for the RAM timmings, to find stable ones and test with each one is nightmare. And don't forget my kindergarden knowledge about RAM timmings, first I have to learn the terminology :). Last time I remember to try to OC with RAM it was with an AMD Athlon XP in the caves era. Anyway thanks for this, it's the best motivation to explore some things, other way I wouldn't bother if make no sense to every day use. Thank you, also, for all the other guys, for being here. If it was just an aggressive competition, I wouldn't be in the group, :).

flexnl
05-15-2013, 09:45 AM
what week is the challange in now? 3th?

nikosa43
05-15-2013, 10:22 AM
what week is the challange in now? 3th? I think is third. This month spreads over 5 weeks. The first one was half. It started from Wednesday.

Arne Saknussemm
05-15-2013, 10:28 AM
It's so many things and so many combinations I want to try, that unfortunately is impossible to do in one month

I am in the same position....not enough time with one thing and another....I'm swapping GPUs and rebuilding watercooling here at the mo. etc. and the amount of tips in this thread! Finding time to try and use and test them all is going to take me ages. As I said earlier the Christmas competition gave me my best results in April 4months later LOL. It will be the same with this comp.

Maybe after HiViz closes the thread and comp he could re open it a week later for the slowcoaches to post up progress:o

8 Pack
05-15-2013, 10:58 AM
Hi guys I am not participating but thought I would submit a Wprime 1024 time. This score is fifth in the world for 4 CPU and I completed it this weekend. I am hoping to try make top three in the next week or so. Time is 111.594s

http://img15.imageshack.us/img15/8941/wprime1024j.jpg (http://imageshack.us/photo/my-images/15/wprime1024j.jpg/)

Here is a pic of the rig.

http://img824.imageshack.us/img824/6162/img00371201305121220.jpg (http://imageshack.us/photo/my-images/824/img00371201305121220.jpg/)

and a top 10 Aquamark, I am to retry this also as a better clocking 580 will improve my rank for sure. This 7970 sucked!!!

http://img4.imageshack.us/img4/5771/aquamp.jpg (http://imageshack.us/photo/my-images/4/aquamp.jpg/)

and again a pic of the rig.

http://imageshack.us/a/img15/7492/gpuln2.jpg

Motherboard is MVF which is my preferred board for clocking RAM.

Luumi
05-15-2013, 11:17 AM
Amazing 8 pack! I was about to write the same on your submission at the bot :P .

Arne Saknussemm
05-15-2013, 11:19 AM
Very nice 8 Pack! as always....nice timings and frequency on that RAM! Ln2 looks good fun :D

HiVizMan
05-15-2013, 11:28 AM
what week is the challange in now? 3th?

Yes third week started yesterday. Each Wed is the next week as it were.


I am in the same position....not enough time with one thing and another....I'm swapping GPUs and rebuilding watercooling here at the mo. etc. and the amount of tips in this thread! Finding time to try and use and test them all is going to take me ages. As I said earlier the Christmas competition gave me my best results in April 4months later LOL. It will be the same with this comp.

Maybe after HiViz closes the thread and comp he could re open it a week later for the slowcoaches to post up progress:o

Ollie that is spot on, you need to take time for the stuff you learn in such a condensed burst during a contest to bed in and make sense. So ultimately it is always months down the line that you will show real gains as a bencher. And for that reason I do not have unlimited cooling as an option. That allows all who compete to experiment with little or no additional cost and then later if they want they can go to a more extreme cooling method. When you are going sub zero you have very little time to test, you have to bench. So little is learnt in the real sense.


HiVizman, .... Anyway thanks for this, it's the best motivation to explore some things, other way I wouldn't bother if make no sense to every day use. Thank you, also, for all the other guys, for being here. If it was just an aggressive competition, I wouldn't be in the group, :).

That aggressive on-line attitude and way of behaving does not have a place here on ROG at all. And most certainly not in any challenge or contest that I organise. Now I am not one of those who do not like competition, or think there should not be a winner or a looser. Nope, I am all for it. What matters to me beside the competing is the fun side of things. And of course the learning side of contests. Thank you for your words and the fact you are here with us, part of us now, means to me that this challenge has been a success.

Zka17
05-15-2013, 11:41 AM
Nice job, 8 Pack! :cool:

nikosa43
05-15-2013, 11:14 PM
That aggressive on-line attitude and way of behaving does not have a place here on ROG at all. And most certainly not in any challenge or contest that I organise. Now I am not one of those who do not like competition, or think there should not be a winner or a looser. Nope, I am all for it. What matters to me beside the competing is the fun side of things. And of course the learning side of contests. Thank you for your words and the fact you are here with us, part of us now, means to me that this challenge has been a success.

That Vizman and other guys' spirit, made me participate in this Forum and in this challenge, which means your words are very meaningful to me.
As for the competition, this is always for the win. Without this, there is no progress. The great thing is to enjoy your win, but when you loose not to have insecurity to honour your opponents.

Menthol
05-17-2013, 01:36 AM
Been very busy, I didn't make a spreadsheet since Zka said he would, If still needed I will make one this weekend. Didn't want to break my water loop again but it's time to throw an older video card in and re-bench, The Titan's don't seem to respond to inspector settings with these older benchmarks, maybe the bios I have flashed with not sure

nikosa43
05-17-2013, 05:51 AM
Hey Menthol, sweet problems you have there. We can trade the Titans with an ASUS GTS 8800 512 I have here :)

nikosa43
05-17-2013, 07:23 AM
nikosa43, when the wprime can not display the results in the original window, it may mean that your system is over-stretched... see in others submission how it should look like... - try to reduce the clock or increase the Vcore... Zka17, you were right. Wprime didn't gave me that last screen, even after had successfully finish the benchmark because CPUs were under throttling. Last night I rerun with the same settings and task manager open in the background and I could see interesting things. The peaks of the cores, instead being on 100%, remind me ice cream cones under the sun. So here is the latest data from last night learning session. Until I join the challenge never tried a multiplier over 50 on air. So my last IBT validated clock was with a multi of 49, 4 cores running plus HT. I could run 50, to other benchmarks, even games with no problems, BSODs or heating problems but because I use IBT maxed out using all memory, as a reference for overstress the CPU and MC and only if passes a few cycles I consider it stable. This time I decided to test the waters of higher clocks. So 51 and 52 was not a problem after a while and I could run Aquamark and 3dmark error free. Wprime is another story. Needs all possible cores and HT to get a decent result, going to 100% all cores and takes ages to finish. So, even my temps are away from Tjmax of 105, it throttles when in the 50% of the bench any core reach 80 C. I tried lots of combinations (HT ON, HT OFF, 2 cores, 3cores, 4 cores). For the given ambient temp of 23 C and a max Vcore of 1.55 (where the BIOS warning comes out for overvoltage and I don't want to override it with the given cooling solution), there is a thermal power package that cause the throttling. Although the temps are not close to Tjmax, total power drawing has reached. So the way to go is to find the most efficient way to use that upper limit of power drawing. For the moment 51 multi with 4 cores and HT is the most promising, which gives all 3 benchmarks a valid result. 52 multi with 2 cores and no HT is another option. The good thing is that MC never caused problems and I can have 2400 with a VCCIO of 1.15Volts up to 52 multi, (timmings are not tight, but I haven't tried anything yet to that domain). It's amazing how many things you can learn when you have set a target :D.

Menthol
05-17-2013, 02:36 PM
Hey Menthol, sweet problems you have there. We can trade the Titans with an ASUS GTS 8800 512 I have here :)
lol it's not a real problem, just stating for reference, maybe someone else using a keepler card has a different experience with inspector on these benchmarks and would like to share ,
there is a bios for Titans that eliminates gpu boost that works well with benchmarks but the nvidia inspector settings do not seem to take effect with these 2 benchmarks like they do in newer benchmarks, not going to hamper the fun of the challenge for me in any way

Menthol
05-17-2013, 04:33 PM
The key is the bonus points. So those folks who have put up entries in both the first two weeks each have 10 bonus points right now. Add that to the totals of each benchmark leader board and you have your total.

Actually Dart or Menthol could you help me out here and put together the leader board for each bench please.

Cheers

Viz,
Please explain exactly how to calculate the score since lower prime is better or are you scoring each benchmark separately?

nikosa43
05-18-2013, 12:14 AM
lol it's not a real problem, just stating for reference, maybe someone else using a keepler card has a different experience with inspector on these benchmarks and would like to share , there is a bios for Titans that eliminates gpu boost that works well with benchmarks but the nvidia inspector settings do not seem to take effect with these 2 benchmarks like they do in newer benchmarks, not going to hamper the fun of the challenge for me in any way lol I was joking :). Btw, do you have original name of Aquamark renamed to crysis or something else? If so, setting LOD from Nvidia inspector, settings not applied because target name is different. This trick I think works for win 7. In win XP leave original name, and set LOD -3 for Aquamark 3. Give it a run and then set +3. In my system it gives a difference if 4000 moreor less. Haven't tried other settings yet.

Luumi
05-18-2013, 06:56 AM
Time for me to put some newer scores here.

Hardware:
Asus Rampage IV Extreme 0006 bios
Intel Core i7 3930K @ 5105
Patriot Viper II Sector 5 @ 2500 9-11-9-27 1T
AMD Radeon HD 7970 @ 1250/1910

Aquamark 330939
20717

3DMark2001 SE 86178
20718

Wprime 1024M 99.968s
20719

HiVizMan
05-18-2013, 11:13 AM
Viz,
Please explain exactly how to calculate the score since lower prime is better or are you scoring each benchmark separately?

More points is better in the two 3D benchmarks. Lower is better in the 2D benchmark. Add the bonus points to the mix and then use the point totals to arrange 1st to 10th place.


lol I was joking :). Btw, do you have original name of Aquamark renamed to crysis or something else? If so, setting LOD from Nvidia inspector, settings not applied because target name is different. This trick I think works for win 7. In win XP leave original name, and set LOD -3 for Aquamark 3. Give it a run and then set +3. In my system it gives a difference if 4000 moreor less. Haven't tried other settings yet.

Er no.

If you change your Aquamark.exe to crysis.exe in XP you will get a huge gain. And you simple use the LOD - to set your LOD. I would suggest folks use my version of LOD not the latest version. It has more range shall we say. Do not pick profiles for LOD just use the actual setting. I will post up some screens later when I get a chance. Just really busy right now so sorry folks. Plus could not log in as I forgot what email addy I had registered with.

flexnl
05-18-2013, 11:23 AM
yah me 2

Asus MVG
3770k @4700mhz
gigabyte gtx 680 OC

AquaMark 385247
20720


3dmark2001 80455
20721


Wprime 157.679 :(
20722

HiVizMan
05-18-2013, 11:43 AM
Looking good Alex. And no I have not forgotten, just busy.

flexnl
05-18-2013, 12:58 PM
Haah okej...well in holland everybody gets vakation money ( or watherver thats called in english :P) bonus or something :P from there boss so any time now would be great as i got mine in da pocket this morning :D:D....

flexnl
05-18-2013, 12:59 PM
but no hurries :) me@ patience :)

HiVizMan
05-18-2013, 01:26 PM
LOL I hear you.

Down to two CPU's. @ -40'C the one has about 150MHz more than the other which has a better IMC.

Will pm you later tonight.

flexnl
05-18-2013, 02:14 PM
aight m8 :D me @ very grateful :D

HiVizMan
05-18-2013, 04:44 PM
20724

Correct as they stand,will do the bonus points for week 3 on Thursday morning. So get your results in there guys.

Dartmaul
05-18-2013, 05:14 PM
2 questions
1) Would you balance wprime score according the amount of cores used (exists)?
2) So if I'll update my submission I would loose bonus points?)

nikosa43
05-18-2013, 05:45 PM
Er no. If you change your Aquamark.exe to crysis.exe in XP you will get a huge gain. And you simple use the LOD - to set your LOD. I would suggest folks use my version of LOD not the latest version. It has more range shall we say. Do not pick profiles for LOD just use the actual setting. I will post up some screens later when I get a chance. Just really busy right now so sorry folks. Plus could not log in as I forgot what email addy I had registered with. I am sure that you are 1000% right when you say that, but if I change the name to Crysis, LOD seems not responding. I 'll definitely try again. One question more. How the points rating works? In Aquamark I got 20 points with a score of more 409000 and in 3dmark 13 points for a score of more 110000. Is it because in Nature I have 1004 fps? Do I have to reduce it under 1000 to the point?

Menthol
05-18-2013, 07:44 PM
Oh you can install XP on to a SSD that is no problem at all.
I slipstreamed the custom x79 ahcii drivers from this site into my copy of xp32 and instal xp in ahci mode went without a hitch

http://www.win-raid.com/t22f23-Integration-of-Intels-AHCI-RAID-drivers-into-a-Windows-XP-CD.html

HiVizMan
05-18-2013, 07:55 PM
2 questions
1) Would you balance wprime score according the amount of cores used (exists)?
2) So if I'll update my submission I would loose bonus points?)

1) Nope it is who is quickest with the same frequency CPU used for all three benchmarks. :p
2) No idea what do you want to update your score with? What CPU?


I am sure that you are 1000% right when you say that, but if I change the name to Crysis, LOD seems not responding. I 'll definitely try again. One question more. How the points rating works? In Aquamark I got 20 points with a score of more 409000 and in 3dmark 13 points for a score of more 110000. Is it because in Nature I have 1004 fps? Do I have to reduce it under 1000 to the point?

LOD really works well if you change the name for XP. I wonder if you are maybe doing something else that is causing you to not get top scores with your system.

I gave you last place score in 2001SE yes, because you had gone over the FPS limit for Nature. :) Get below 1000FPS and you would have 20point. :D

Last thing - nice benching. :p


I slipstreamed the custom x79 ahcii drivers from this site into my copy of xp32 and instal xp in ahci mode went without a hitch

http://www.win-raid.com/t22f23-Integration-of-Intels-AHCI-RAID-drivers-into-a-Windows-XP-CD.html

Any chance you could upload that OS as a IOS for me to test some time. Pretty please :o

Menthol
05-18-2013, 09:04 PM
Of course, I have never uploaded any files before, just tell me where

flexnl
05-18-2013, 10:42 PM
I slipstreamed the custom x79 ahcii drivers from this site into my copy of xp32 and instal xp in ahci mode went without a hitch

http://www.win-raid.com/t22f23-Integration-of-Intels-AHCI-RAID-drivers-into-a-Windows-XP-CD.html


tnx man... i was/am struggling with that pff...
i have asus MVG..and my dvddrive gives bsod @ install xp (both ahci and ide)
the link you gave is talking about setting the sataslot of your dvddrive to IDE
do you or anyone else know how to set a sata slot to IDE when the others are AHCI? it isnt possible is it? :)

anyway great link

Menthol
05-19-2013, 12:12 AM
You need to put you DVD/CD drive on the Asmedia port, or whichever ports that are not Intel, set that controller to IDE in the bios under on-board devices where the on-board sound is located, at least on the RIVE, set the Intel SATA controller to AHCI, hard drive on the Intel controller,

flexnl
05-19-2013, 12:23 AM
aaah so I C FINALY! TNX
you dont wanna know how many times i saw a bleuscreen in lpast few days :D
or not recognized hdd"s.
1000x tnks m8 :)
now.... bback 2 benchin :P
;)

Zka17
05-19-2013, 02:43 AM
Wow, we have a leader board! :cool:

Have to hurry up and do some more runs... undergoing major upgrades to my ROG corner/man cave, so all my systems are down right now... - but obviously will be back as soon as possible! :)

chrsplmr
05-19-2013, 04:13 AM
.. Excuse me Please, for interrupting .. Simply Amazed by what you All, are not only doing, but teaching.
I can not Thank you enough.
It sometimes makes me feel like a kid being led by the hand to FunHouse. Very Exciting.
Soon I won't be able to stand it anymore .. and then my wife will kill me. hahhahaa

Sorry. Had to say it. Carry on.
Congrats to the All for moving me forward.c.

HiVizMan
05-19-2013, 09:08 AM
Of course, I have never uploaded any files before, just tell me where

Depending on the size of the ISO I would suggest uploading to Mediafire. If you prefer a more secure option I will send you a user name and password for a private file hosting site.

HiVizMan
05-19-2013, 09:53 AM
Right time for some more tips and one important update first.

********Update**********

With most of the newer hardware I am seeing that Win7 64bit is working best for AM3 and the new HWBOT Wrapper I am talking to the designer of the wrapper about why that would be the case. I too have learned something this challenge.

************************************************** ********************************************

Tips

8 Bits Colour Tweak

This tweak will work in all the OS's I have tested. Start your benchmark (3D) and let it idle. Change the desktop resolution to the corresponding resolution of the benchmark, apply. Now in the same screen you scroll down and select Colour settings to 8 Bit. Apply and close. This setting will remain even after you close your benchmark. It does make your screenshot look like poo, so I suggest you change your picture quality before you take your screen for submission. And yes it is legal to do this tweak.


Aquamark3 Tweaks

The wrapper has changed some things about how this benchmark runs for some reason. And W7 64bit seems to be the most effiecent, however I am not 100% convinced by that as yet so I am going to include all OS tweaks here.

The starting place is ensuring that your AM3 is set up right, so use the built in AM3 64bit patch that is included in the wrapper, I like to delete the Direcpll.dll and I still rename the .exe file even with the wrapper.

Next right click on the application icon in programes set the compatibility settings to the following 256 colours and the screen resolution to 640 x 480.

Now I rename the Aquamark3.exe into what ever name I favour, I use 3Dmark03 or Crysis as a rule as they give me the greater gains. This name change forces more than a single core of your CPU to work with the benchmark.

Now here are more advanced tweaks and I expect you to find where you set them yourself.

AA: Disabled
AD: Disabled
Tripple buffer: on
Pre Render FPS: you need to find what works best for your hard ware I found that 3 worked just peachy with the 4890 card I used in the screenshot I posted
Display: Single Performance Display
Sli Mode: as needed
Texture Aniso Option: On
Texture Neg. Lod bias: Clamp (Forces Lod to be 0 if you use LOD negitive then there is no point in using this option)
Texture Trilinear Option: On
Texturefiltering Quality: High Performance or Performance
Threaded Option: Force On
Vsync: Disabled


Now those of you that are first time Rivatuner users or not sure how to increase the amount of LOD range in Rivatuner here is how.

Navigate to Power User -> Nvidia Global -> LODBiasRange : Set to 15 ( this means you can selecet -15 to +15 LOD)

That will do for now. There are still heaps more tweaks that can be used by the way. But the key to AM3 is finding the best ram settings for your system. You want bandwidth. Test and tune in OS and then do a run with the settings locked in via the BIOS.

Arne Saknussemm
05-19-2013, 10:29 AM
Thanks a lot HiVizMan....that just set back my final entry to this comp another month! :D

HiVizMan
05-19-2013, 11:03 AM
Thanks a lot HiVizMan....that just set back my final entry to this comp another month! :D

Two entries. You want one before wed for two bonus points and then your winning entry to take the crown from ???

Arne Saknussemm
05-19-2013, 11:09 AM
LOL ...yeah right.....but I'll give it a go.....off home now to do a bit of benching......dig out a HDD and install XP etc... :) I've got a WD caviar blue that's empty and a caviar black that's full of stuff....both 7200 rpm and 32 mb cache.....any real difference?

HiVizMan
05-19-2013, 11:40 AM
Nope makes no difference to these benchmarks at all.

Arne Saknussemm
05-19-2013, 01:59 PM
Thanks HiVizMan! I'll go with the easy option then.....

Menthol
05-20-2013, 03:04 AM
Depending on the size of the ISO I would suggest uploading to Mediafire. If you prefer a more secure option I will send you a user name and password for a private file hosting site.

sent you an email with link, let me know if there's a problem

HiVizMan
05-20-2013, 07:41 AM
Host you uploaded to expects me to pay for premium status to download. Sorry no can do. Mediafire if possible Menthol. :)

Thanks for trying.