Wow I broke 10k check it

[hoodedrat]

10k broken on this system check it out

xp1800 @ (1.73ghz 150mhz fsb)
msi kt3 ultra aru
384mb crucial pc2100
radeon8500LE (299/299)

click here

[grimfandango]

Originally posted by hoodedrat
10k broken on this system check it out

xp1800 @ (1.73ghz 150mhz fsb)
msi kt3 ultra aru
384mb crucial pc2100
radeon8500LE (299/299)

click here

says
The project id is not valid - please check the URL.

[user0209]

did you also break the url?

[hoodedrat]

this should fix it

» http://service.madonion.com/compare?2k1=4401509

[JuNacy]

Cool

[grimfandango]

Originally posted by hoodedrat
this should fix it

» http://service.madonion.com/compare?2k1=4401509

you just made it!!!




But very good for an
RADEON 8500/RADEON 8500LE!!!

[OpK Chowdy]

wow good job. im trying to break 10k myself with my geforce4ti4200 128mb. im waiting to install ramsinks before i take the ram passed 550mhz.

[gibsinep]

Yes but this was just the "game test" none of the other.

[salandar]

ya know, i'm really concerned about my system and overall
3DMark2001SE Score

I have the following, yet I can barely keep up with this one:

AMDXP 2200+
MSI-KT333 Ultra ARU Mobo MS-6380E
Corsair DDR2700 cas 2 cl 2 RAM
Maxtor IDE 133 40gb HDD
MSI GeForce4 4400 graphics card.

When I look at his, and compare it to my scores, his NATURE GAME, (game 4), blows mine out of the water. (We have the same exact mobo also).

Game 4 - Nature
(FPS)
My Score: 39.5
His Score: 55.4

What can I do to improve upon this terrible FPS score? I know this is really hurting my overall score which is a measly 10295 or so

[omega31]

You could change the mipmap detail level to best performance.

That would be cheating, though.

[foxwolfen]

Originally posted by salandar
ya know, i'm really concerned about my system and overall
3DMark2001SE Score


Game 4 - Nature
(FPS)
My Score: 39.5
His Score: 55.4

What can I do to improve upon this terrible FPS score? I know this is really hurting my overall score which is a measly 10295 or so

Please note...
pc2100 ram on a 150mhz fsb = overclocked...

Frankly, these numbers are quite meaningless other than demonstrating to the machines owner that his efforts at squeezing speed were sucessfull.

If you want to add a thousand points to your score do the following...

Short the L1 bridges and unlock the multiplier. Short the the voltage bridge to produce an internal voltage of 1.85 and puch your machine to 2600 or better.

Do force your fsb above 166mhz as you have 333mhz ram

Then depending on your video card do the same for the memory chips (heatsinks on the memorychips on todays vid cards is pointless, they dont get hot..they need voltage).

Then off Anistropic filtering and use trilinier mipmaping, double buffering, with vsync off /10 frames render ahead, no AA and only run the games part of 3dmark.

Then you should be able to get some marks that relate to nothing in the realworld but look cool in forums.

FoxW

PS: If you really want some impressive numbers, run openGL benchmarks too... for reasons known only to god, it seems everybody here only knows about 3dmark which is a very pretty to look at benchmark but pretty lousy at predicting a machines perfomance in games. Here is a list of benchmarks I use that, when combined, provide a somewhat more realistic predictor of actual in-game performance.

[System]
CPU - ZDNets 3D Winbench 2000
Memory - Sandra 2002 or better yet, vans Memory Latency Plus
Bus - Vans Bandwidth Burn
2d bus speed - Final Reality 1.1
3d bus speed - Final Reality 1.1
AGP Bus Speed - Final Reality 1.1

[video]
Fill rate - Final Reality 1.1
Pixel Shaders - 3dmark2001SE
Hardware T&L - 3D Winbench quality
Synthetic GL - GL Excess
Real World/Synthetic GL - Vulpine GLmark
Synthetic D3D - 3dmark2001
Realworld GL - Serious sam/quake 3
Realworld D3D - ????? I know of nothing that qualifies... this is likely due to MS not wanting D3D realworld numbers compared to OpenGL... OpenGL is considerably more advanced and far far faster.

Again, all these numbers do is show you what effect changes you make to your system, not how your system compares to others. For expamle, My friend has a machine that is speced higher than mine... it beats me on synthetic benchmarks, but when we go realworld, I kick his but all over the place... He will bog out in a heavy firefight in Serious Sam when I am running smooth. Why... because there is more to computing than Synthetic Videocard Benchmarks.

FoxW

[salandar]

hoodedrat, I don't mean to hijack your post, but it's a good place and looked at topic

foxwolven, thanks for the ideas, i'll try them, except for the part ""Short the L1 bridges and unlock the multiplier. Short the the voltage bridge to produce an internal voltage of 1.85 and puch your machine to 2600 or better." I wouldn't know where to begin with that and i'm sure everything would explode in my face if I was to try that

I was messing with some settings in my BIOS last night and found that they really don't do much with my overall benchmark scores.

CorsairXMS 2700DDR 333mhz cas 2 RAM
MSI KT333-Ultra Mobo

Below is what options I have under the Advanced Chipset Features and DRAM Timing Control:

1) SDRam Frequency ~ left at SPD (serial presence detect) vs. HOST CLK, HOST CLK+33, HOST CLK+66

2) SDRAM CAS# Latency ~ Chose (2) vs. (3) since my corsair ram is able to handle 2 clock cycles.

3) Row Precharge Time ~ I tried changing this a couple of times, but each time I changed the number, it wouldn't let me play games. They wouldn't load, or they would give me a fatal memory read error. The available settings were 2T and 3T, and I had to chose 3T.

4) RAS Pulse Width ~ I was able to change this one without any effects to my games, so I chose 5T vs. 6T.

5) RAS to CAS delay ~ I tried to change this one, but it would mess up games, so I had to stay at 3T vs. 2T.

6) Bank Interleave ~ This is one that I haven't messed with too much. The booklet tells me to leave disabled if 16MB of SDRAM is installed, but I have DDRam, so I don't think that applies to me.
Options for setting is: Disabled, 2-Way, and 4-Way

7) SDRAM Burst Length ~ This one seems to mess up my computer a bit if I change the settings. The options for this are, 4QW and 8QW. My BIOS is set on 4QW ~~ What does QW stand for?

8) SDRAM 1T Command ~ Either enable, so it does a command rate of 1T or disable at 2T. I have it disabled at the moment, but not sure how much of an improvement 1T would give.

9) Fast Command ~ Ultra

10) AGP Aperture Size ~ have it set for half (64mb) of my video card memory of 128. I believe I read somewhere on these forums, that doing this would give a performance increase ~ Is this true?


Well, if you made it this far, maybe there is something in this mess of information that you could tell me...

[foxwolfen]

Originally posted by salandar
hoodedrat, I don't mean to hijack your post, but it's a good place and looked at topic

foxwolven, thanks for the ideas, i'll try them, except for the part ""Short the L1 bridges and unlock the multiplier. Short the the voltage bridge to produce an internal voltage of 1.85 and puch your machine to 2600 or better." I wouldn't know where to begin with that and i'm sure everything would explode in my face if I was to try that

ROFL, that was my point, you sometimes have to go to rediculous lengths to get comparable numbers.

<SNIP>

6) Bank Interleave ~ This is one that I haven't messed with too much. The booklet tells me to leave disabled if 16MB of SDRAM is installed, but I have DDRam, so I don't think that applies to me.
Options for setting is: Disabled, 2-Way, and 4-Way

Bank interleave is when you have two or more Dimms installed and want to fine tune the memory to a sort of parrallel configuration... what that means is the system will access memory from both banks at the same time effectivly speeding up the whole system... Usually you need ecc memory at minimum which is not yet available in 333 i believe. (this is usually something you would do if you were planning on turning your machine in to a hi-transaction e-commerce server. The rest of would not likely see any performance increase by interleaving and in many circumstances we would acctually see a decline.)

<SNIP>

10) AGP Aperture Size ~ have it set for half (64mb) of my video card memory of 128. I believe I read somewhere on these forums, that doing this would give a performance increase ~ Is this true?

A good rule of thumb is half of your installed *system* memory, up to 128mb for games. The aperture allows the card to access up to that much memory for texture swapping. This will not likely make much difference with your video card as 128 mb of texture memory is not likely to ever be used up in the next year or so.. leaving it at 64 is a safe bet.

FoxW

[nramsey]

Nice job! I just finally broke 10K yesterday too. My system:

http://service.madonion.com/compare?2k1=4455062

Card is Gainward Geforce 4 Ti 4200 64 mb Golden Sample at stock "performance" clock settings using their tool (think it's 270 gpu/550 mem or so).

Previous score with above system but with ATI Radeon 7500 was in the 5500 3D marks range. Not a bad upgrade (5500 -> 10300) for $150

[salandar]

the thing that has me really frustrated is this:

Game 4 - Nature

My SCORE: (FPS) 39.5

COMPARED SCORE: (FPS) 68.9

Why in da hell do I get a FPS of 39.5, when I have a geforce4 4400 128mb gpu, while the compared has a geforce4 4200 64mb GPU?????????