Low clock 3D Pedro's Game style...

[T_M]

LL can be higher yes, but it will sack more important tests if starting with LL, I experienced this recently when I tried GX2

Can you explain this one a bit more, i dont quite understand

[George_o/c]

Can you explain this one a bit more, i dont quite understand

Whichever is the FIRST test that you test (the first from the test sequence), no matter which is that, it's going to be higher than normal For example as M.Beier says if you first run LL, LL will be higher than when you test it in this order : N – CH – LH – DL – DH – LL – CL

But there are some tests than you need to test first, become they need more CPU and Memory Power than the others These two are CH and LH You can apply what I'm telling you, and you 'll figure out all these yourself Tim

Also one thing : When testing sequences, you 'll notice that when the High and Low tests are run in High-Low or Low-High order, you 'll get a boost in comparison with High-High, or Low-Low order You know why ?
HIGHs always require more cpu power (and memory, but cpu is what needed the most ) than LOWs And so, if your test sequence has two HIGHs in a row, that will put more load on your cpu and the score will be worse

Last but not least, you should never forget that :

1) Cars are in desperate need () of cpu power
2) Dragos like graphics memory bandwidth ... Hence, try increasing your VGA Memory Frequency a bit higher when you 're benching Dragos (My favorite test )
3) Lobbies like 2 things ... 1st high FSB and 2nd system memory bandwidth ... So open MemSet and start tweaking as much as you can your RAM timings and sub-timings, and you 'll notice a boost
4) Nature is all about pure gpu power (well most) ...

I hope I was of a little help m8 Keep pushing it (Sorry Oliver )

[fubarswe]

Tried it a lil..
E8400 Air
EVGA 8800GTS Sli aircooled
EVGA 790i
GSkill DDR3 1800



Another usefull way for F1..



.., fixed a couple of GTX´s.., so I tested them too...

[SIOUX]

Whichever is the FIRST test that you test (the first from the test sequence), no matter which is that, it's going to be higher than normal For example as M.Beier says if you first run LL, LL will be higher than when you test it in this order : N – CH – LH – DL – DH – LL – CL

But there are some tests than you need to test first, become they need more CPU and Memory Power than the others These two are CH and LH You can apply what I'm telling you, and you 'll figure out all these yourself Tim

Also one thing : When testing sequences, you 'll notice that when the High and Low tests are run in High-Low or Low-High order, you 'll get a boost in comparison with High-High, or Low-Low order You know why ?
HIGHs always require more cpu power (and memory, but cpu is what needed the most ) than LOWs And so, if your test sequence has two HIGHs in a row, that will put more load on your cpu and the score will be worse

Last but not least, you should never forget that :

1) Cars are in desperate need () of cpu power
2) Dragos like graphics memory bandwidth ... Hence, try increasing your VGA Memory Frequency a bit higher when you 're benching Dragos (My favorite test )
3) Lobbies like 2 things ... 1st high FSB and 2nd system memory bandwidth ... So open MemSet and start tweaking as much as you can your RAM timings and sub-timings, and you 'll notice a boost
4) Nature is all about pure gpu power (well most) ...

I hope I was of a little help m8 Keep pushing it (Sorry Oliver )

Not 100% true though. Some tests, like DL, gives the same boost, nommater of you run it as 1st 2nd or 3rd (maybe you loose 4fps). after that it will degrade drasticly......wich is very importent when you choose the running order. ....also, very offen, DL can be ran with 50mhz higher cpu speed.
Lobby low also dont mind to much being ran 2nd. the fps do not drop a lot before 3rd run.

In my experience CH and LH benefit mostly from being ran first......

[George_o/c]

1) Not 100% true though. Some tests, like DL, gives the same boost, nommater of you run it as 1st 2nd or 3rd (maybe you loose 4fps). after that it will degrade drasticly......wich is very importent when you choose the running order. ....also, very offen, DL can be ran with 50mhz higher cpu speed.
Lobby low also dont mind to much being ran 2nd. the fps do not drop a lot before 3rd run.

2) In my experience CH and LH benefit mostly from being ran first......

1) Hmmm thanks for the info, I appreciate it m8

2) I was right then

[massman]

Run the test with the lowest FPS first. I always run CH first as it's the hardest one, the lowest FPS and worth FPS x20 points

[George_o/c]

and worth FPS x20 points

Is this guaranteed ? If it's for real, then screw LH ... CH first then

[massman]

CH-DH-LH-N: FPS x 20 = points
CL-DL-LL: FPS x 10 = points

This is basic stuff, right? :p

[George_o/c]

CH-DH-LH-N: FPS x 20 = points
CL-DL-LL: FPS x 10 = points

This is basic stuff, right? :p

Mass thanks a lot Useful info

[BenchZowner]

Oh Vince, you just stole this idea from me
Well, you got to the lake first actually, as I'm sure you've been thinking about this thread for a while
I'd be a little more strict ( CPU & VGA freq's given ), but strict CPU & free VGAs is also fine.

Guys, 322FPS CH @ 4.58GHz is more than good actually.

Gonna join the "club" soon ( gotta get some of my cards back from a friend )

p.s. The "gun" has fired once Vince, and it's 9800GTX @ 1068MHz Core at UnKnOwN Voltage ( rushed everything, including insulation, mods, measurement points, etc ) @ all 3D benches @ -52C ( -53C is coldbug )

[M.Beier]

Not 100% true though. Some tests, like DL, gives the same boost, nommater of you run it as 1st 2nd or 3rd (maybe you loose 4fps). after that it will degrade drasticly......wich is very importent when you choose the running order. ....also, very offen, DL can be ran with 50mhz higher cpu speed.
Lobby low also dont mind to much being ran 2nd. the fps do not drop a lot before 3rd run.

In my experience CH and LH benefit mostly from being ran first......

Hmm...
Now I prefer N->LH, because its the two tests that are more likely to crash when pushing CPU/RAM...
BUT.... Running DL at first use to be beneficial.... - and after each run, deleting 01SE, rebooting, reinstalling, and then running again, we all remember that 5-9FPS gain now dont we? but thats 50-90marks, when around the 100K mark, other things are more beneficial to do....
Atleast, thats my oppinion.

[k|ngp|n]

Good on air Vince!

Interest thing 9800GTX about have same performance as 8800GTS 512MB, but 173.67 driver give very-very lot boost for 9800GTX in nature...I have almost same results as 9800GTX in all test, but nature is slower 100-150fps.

Me and gautum were talking same thing earlier. Did you try 169.21?

[SIOUX]

Hmm...
Now I prefer N->LH, because its the two tests that are more likely to crash when pushing CPU/RAM...
BUT.... Running DL at first use to be beneficial.... - and after each run, deleting 01SE, rebooting, reinstalling, and then running again, we all remember that 5-9FPS gain now dont we? but thats 50-90marks, when around the 100K mark, other things are more beneficial to do....
Atleast, thats my oppinion.

Nobody said anything about running DL first. Nature can in some cases give a boost by running it first. Imo with the 8800 series that was the thing to do. With gx2 i dont see that boost, and its therefor useless for me to start with...in my opinion....different hardware different settings and tweaking.....







BTW: Where is Pedro ??? it worrys me to see a thread like this, and he isnt here

[T_M]

Wow, this thread is turning gold very quickly. Thanks guys, such a simple question i asked and look at the knowledge flowing

[massman]

01 is all about finding the bottleneck of your system. Once you've found it, you can tweak the run order, tweak memory timings and so on.

[k|ngp|n]

Pedro will show up sooner or later I'm sure of it
Anyone lose any sleep last night running 01?? Anyone?
Gautum did prolly
I expect to see some killer 01's when I get back on friday....maybe 100k at same clocks or lower would be the own
600x7.5 anyone??

[Oliver]

Sweet work and thread Vince... really like the topic ...

I will try to catch up with gx2 on air and slightly overclocked and under 4.6g on cpu and see whats going on...

If i had 98gtx i would 4 sure give that a run or two ...

Keep pushing it...

[Viss]

600x7.5 is what i had in mind, too bad i have only have Ati card

[dpa]

nice post.. might need to test this out!

where can I download the modified .inf file?

[hey]

i will try with gts's when find a 790i

[Pedro Rocha]

BTW: Where is Pedro ??? it worrys me to see a thread like this, and he isnt here

I'm here my friend, sorry for not post earlier but May is really the worst month in the year in my profession ..

Thank you Vince for creating this thread, I totally agree with you: the "real original art" of benching still is 2K1 - our favorite game

Try to find highest effeciency is really the best part of the game for me..

Now my 790i / E8500 still are on cascade where I did recently did the "Best of The Rest" - i.e. the best 2K1 result in the ORB without counting Mr. k|ngp|n results

Did 114.737 with the E8500 on the cascade and 2 9800GTX's stock - see http://www.xtremesystems.org/forums/...0&postcount=36

115 is so so close, I may do it in the the next run...


My best "All Air" 9800GTX run was 99.760 I think, I have the screen cap somewhere in the bench disk - used E8500 @ 4.6 cooled by Tuniq / 9800GTX's 825-1240 stock..

I will try the 100K "All Air" as soon I dismount the setup from the cascade, but 1st I must do 115K with CPU on cascade


I think you all know my runnig order, but here is it again : - LH -> LL -> CH -> DL -> DH -> CL -> Nature

In Lobbies (my favorite tests of the game) I always use the best memory timmings I can and the best performance level the mobo can give ..hello ..hello Mr. Voltage

CH - It's CPU, memory and luck ...set priority to high may help your luck

Dragos nowdays is much more CPU/FSB than GPU Core/Mem ..believe me extra CPU/FSB can help you more than extra GPU speed..

CL is random as you all know, altought I have a "tweak" to force my luck - pray to OC goods for a good Car Low and pray for the dam thing dont crash

In Nature with SLI the CPU speed still count, but must important is of cousre GPU Core/Mem speed ..

The key for good runs also is Bios optmization, in 790i settings like HTT Multipliers and specilly SPP-MCP Ref Clock have great impact !

In short all counts and that's why 2K1 is :

[BenchZowner]

nice post.. might need to test this out!

where can I download the modified .inf file?

Attached here.
Rename to inf ( nv4_disp.inf ) and copy it to the temporary folder used for the GeForce drivers installation.

I'm here my friend, sorry for not post earlier but May is really the worst month in the year in my profession ..

Thank you Vince for creating this thread, I totally agree with you: the "real original art" of benching still is 2K1 - our favorite game

Try to find highest effeciency is really the best part of the game for me..

Now my 790i / E8500 still are on cascade where I did recently did the "Best of The Rest" - i.e. the best 2K1 result in the ORB without counting Mr. k|ngp|n results

Did 114.737 with the E8500 on the cascade and 2 9800GTX's stock - see http://www.xtremesystems.org/forums/...0&postcount=36

115 is so so close, I may do it in the the next run...


My best "All Air" 9800GTX run was 99.760 I think, I have the screen cap somewhere in the bench disk - used E8500 @ 4.6 cooled by Tuniq / 9800GTX's 825-1240 stock..

I will try the 100K "All Air" as soon I dismount the setup from the cascade, but 1st I must do 115K with CPU on cascade


I think you all know my runnig order, but here is it again : - LH -> LL -> CH -> DL -> DH -> CL -> Nature

In Lobbies (my favorite tests of the game) I always use the best memory timmings I can and the best performance level the mobo can give ..hello ..hello Mr. Voltage

CH - It's CPU, memory and luck ...set priority to high may help your luck

Dragos nowdays is much more CPU/FSB than GPU Core/Mem ..believe me extra CPU/FSB can help you more than extra GPU speed..

CL is random as you all know, altought I have a "tweak" to force my luck - pray to OC goods for a good Car Low and pray for the dam thing dont crash

In Nature with SLI the CPU speed still count, but must important is of cousre GPU Core/Mem speed ..

The key for good runs also is Bios optmization, in 790i settings like HTT Multipliers and specilly SPP-MCP Ref Clock have great impact !

In short all counts and that's why 2K1 is :

for sharing Pedro.
This is what overclocking and real overclocking legends are here for, playing, helping, sharing & exploring

[k|ngp|n]

I'm here my friend, sorry for not post earlier but May is really the worst month in the year in my profession ..

Thank you Vince for creating this thread, I totally agree with you: the "real original art" of benching still is 2K1 - our favorite game

Try to find highest effeciency is really the best part of the game for me..

Could we ask for more than to have THE man show up and drop some 2k1 knowledge
Thanx so much for the post 01 legend..

nice post.. might need to test this out!

where can I download the modified .inf file?

I used the one that Billy(benchzowner) linked to somewhere on XS Search for 173.67

[BrokenWall]

I am going to try to join you guys tonight and see what I can push as far as effenciency out of my setup.

It will have to be all air, and no more than 3.4Ghz on the AMD X2 and 800/2400 on the 98GTX just to push stuff a bit on the hardware. I don't have a current run at the moment for my 98GTX recorded on the orb to base off of. But I will definately trying to break past my 350HTT on my AM2 and see if I can hit 400HTT. You guys wish me luck, Ill take what I learned here and see how it goes.

[k|ngp|n]

I am going to try to join you guys tonight and see what I can push as far as effenciency out of my setup.

It will have to be all air, and no more than 3.4Ghz on the AMD X2 and 800/2400 on the 98GTX just to push stuff a bit on the hardware. I don't have a current run at the moment for my 98GTX recorded on the orb to base off of. But I will definately trying to break past my 350HTT on my AM2 and see if I can hit 400HTT. You guys wish me luck, Ill take what I learned here and see how it goes.

Good luck man been awhile since X2.