Sign in with
Sign up | Sign in
Your question

What do you think of these 3DMark2001 results...

Tags:
Last response: in CPUs
Share
January 21, 2002 3:12:40 PM

I just downloaded and ran 3DMark2001 and got the results below. What do you think? Are they decent marks?


----------------------------------
<b>System configuration & settings:</b>
----------------------------------

Operating System --------------<b>Microsoft Windows XP</b>
DirectX Version ----------------<b>8.00</b>

Mobo Manufacturer -------------<b>Dell Computer Corporation</b>
Mobo Model ------------------<b>Dimension 8200</b>

CPU -------------------------<b>Intel Pentium 4 1795 Mhz</b>
FSB -------------------------<b>100 MHz</b>

3D Accelerator ----------------<b>NVIDIA GeForce3 Ti 200</b>
Graphics Chipset --------------<b>NVIDIA GeForce3 Ti 200</b>
Driver Version ----------------<b>5.13.01.2183</b>
Video Memory ----------------<b>64 MB</b>

Resolution --------------------<b>1024x768 32bit</b>
Texture Format ----------------<b>Compressed</b>
FSAA ------------------------<b>Disabled</b>
Z-Buffer --------------------- <b>Depth 32bit</b>
Frame Buffer ------------------<b>Double</b>
Rendering Pipeline -------------<b>D3D Pure Hardware T&L</b>



--------------------------
<b>Detailed Test Results:</b>
--------------------------

3DMark Result ------------<b><font color=red>6202</font color=red> 3D marks</b>

Game 1 Car Chase - Low Detail ---<b>96.8 FPS</b>
Game 1 Car Chase - High Detail --<b>36.3 FPS</b>
Game 2 Dragothic - Low Detail ---<b>86.4 FPS</b>
Game 2 Dragothic - High Detail --<b>50.6 FPS</b>
Game 3 Lobby - Low Detail ------<b>103.4 FPS</b>
Game 3 Lobby - High Detail -- ---<b>49.8 FPS</b>
Game 4 Nature ------------------<b>30.0 FPS</b>

Fill Rate (Single-Texturing) ------<b>566.4 MTexels/s</b>
Fill Rate (Multi-Texturing) -------<b>1160.7 MTexels/s</b>

High Polygon Count (1 light) -----<b>13.7 MTriangles/s</b>
High Polygon Count (8 lights) ----<b>2.7 MTriangles/s</b>

Environment Bump Mapping ------<b>93.4 FPS</b>
DOT3 Bump Mapping -----------<b>91.9 FPS</b>

Vertex Shader Speed ------------<b>41.8 FPS</b>
Pixel Shader Speed -------------<b>67.0 FPS</b>
Point Sprite Speed -------------<b>13.0 MSprites/s</b>

Published -----------------------<b>No</b>



Just so that all the Benchmarks are contained in one post, I posted the following benchmarks on this forum earlier, using SiSandra 2001 (version 2001.0.7.10):

<b>CPU Benchmark:</b>
* Dhrystone ALU -----------------<b>3691 MIPS</b>
* Whetstone FPU/SSE2 ----------<b>938/2195 MFLOPS</b>

<b>Multi-Media Benchmark:</b>
* Integer SSE2 -----------------<b>7131 it/s</b>
* Floating-Point SSE2 -----------<b>8819 it/s</b>

<b>Memory Benchmark:</b>
* Integer ALU/RAM Bandwidth ----<b>1574 MB/s</b>
* Float FPU/RAM Bandwidth ------<b>1609 MB/s</b>

* Performance Rating (PR) = -----<b><font color=red>2154</font color=red></b>


My spec is:

<i>Dell Dimension 8200
Pentium 4 1.8GHz
384 RDRAM
nVidia GeForce3 Ti200
100GB Hard Disk (7200 rpm)
19" Flat Screen Trinitron Monitor
16x CDRW Drive
16x DVD Drive
Win XP</i>


<i><font color=green>When the cup is full, no more can be poured-
When you're full of your opinions, no more wisdom can enter</font color=green></i>

More about : 3dmark2001 results

January 22, 2002 12:58:44 AM

Is anyone experienced and knowledgable enough to tell me if these results look normal? I have not overclocked the CPU or the GPU. How can I improve these marks?

Wen the cup is ful,no more can be poured-
wen you'r ful of your opinions,no more wisdom can enter
January 22, 2002 1:37:13 AM

Considering your hardware, those scores are pretty normal.

Of course, you don't have to just ask me...you can use the 3Dmark2K1 "search and compare" feature to seek out a system almost exactly like yours and see how it stacks up.

Have you ever actually logged in to the MadOnion Online Results Browser (ORB)? (Just asking, not drilling you or anything)

<i>If a server crashes in a server farm and no one pings it, does it still cost four figures to fix?
Related resources
Can't find your answer ? Ask !
January 22, 2002 7:07:10 AM

Those are normal scores. Since 3DMark2001 is very heavily weighted toward graphics card performance, the best way to improve your scores is to either overclock your graphics card or buy a better card. So overclock that Ti 200 if you really want to get better scores. There are little things you can do, like closing down background programs while running the benchmark, but those will only get you so far. You'll see a little gain by overclocking your system, but the biggest gains will be from overclocking your video card.

<i>There are two theories on arguing with women. Neither one works.</i>
January 22, 2002 5:26:05 PM

In bios:
Set write cache to UC
Set AGP apature size to 64
AGP fast writes enabled
turn off all unused devices like serial ports, lpt, usb, PM
turn off bios and video shadowing

In windows:
remove all icons from desktop
set refresh to 60
set desktop to 1024x768x32
disconnect your floppy, CDROM and extra IDE devices
defrag your HD
load "cacheman" set vcache 128k min 120000max
load "rivatuner" and inside RT load matched driver set.
Overclock your video card in small steps.
if the bench locks to a still frame up its your core, if the image start to get artifacts its your memory. play with both increasing in 5Mhz steps.

Once you find a decent setting for your video card, save your score and jump into bios and change write cache to USWC and bench again. also test the APG apature size and 32 and 128. 64 usually works best.

Post your score when you did this stuff.
Anonymous
a b à CPUs
January 22, 2002 7:47:35 PM

"turn off bios and video shadowing"
Hi,
Why did you suggest the above? I thought this would actually improve the perfoirmance? Is that not so?

Regards
AJ
January 22, 2002 9:19:50 PM

Fugger, I'm going to try all those things you mentioned and then I'll see what happens and let you know the outcome...

Wen the cup is ful,no more can be poured-
wen you'r ful of your opinions,no more wisdom can enter
January 22, 2002 9:44:38 PM

Quote:

In windows:
remove all icons from desktop
set refresh to 60
set desktop to 1024x768x32
disconnect your floppy, CDROM and extra IDE devices
defrag your HD
load "cacheman" set vcache 128k min 120000max
load "rivatuner" and inside RT load matched driver set.
Overclock your video card in small steps.
if the bench locks to a still frame up its your core, if the image start to get artifacts its your memory. play with both increasing in 5Mhz steps.


That's great for getting the highest possible 3DMark2001 but you'll render (pun intended) your computer nearly unusable. It adds new meaning to the words "theoretical test".

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 22, 2002 9:57:42 PM

I personally find the benchmark seriously flawed. It is so heavily weighted on the video card alone, it is by no means whatsoever a good gauge of full system preformance.

AMD Man, I agree with you completely in that arena. The only benching I can see doing is the way you have your computer running 24/7.

:tongue: Have you ever tried cooking an egg on your HSF? Tasty. :tongue:
January 22, 2002 11:56:28 PM

your machine is still useable, you can always reattach the IDE devices when done, how many people actually have serial port devices? refresh rate is easy to set back. most people run in 1024x768x32 anyway. cacheman is a good all around program like riva tuner.

As far as the rest. icons on the desktop make no diff on your machines usability, move them to the start menu bar.

I was assisting in getting better score without the ability of overclocking his CPU.

Im sure my tips helped more people too.

Also driver versions are important, check the ORB to see what video drivers are best for your OS.
January 23, 2002 1:10:16 AM

Quote:


your machine is still useable, you can always reattach the IDE devices when done, how many people actually have serial port devices? refresh rate is easy to set back. most people run in 1024x768x32 anyway. cacheman is a good all around program like riva tuner.

As far as the rest. icons on the desktop make no diff on your machines usability, move them to the start menu bar.

I was assisting in getting better score without the ability of overclocking his CPU.

Im sure my tips helped more people too.

True, but my point was it's no longer a test of your overall gaming performance but merely a number that you can show off.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 23, 2002 2:39:38 AM

True, but my point was it's no longer a test of your overall gaming performance but merely a number that you can show off.

yes like all benchmark , do i play quake with 16 bit texture or 32 bit.Benchmark are totally usless.

http://gamershq.madonion.com/products/orb/?publish_comp...
January 23, 2002 3:11:54 AM

AMD_MAN Okay - no need to split hairs here. There is a limit you can tune a system to to still have a workable day machine. Some people will build and configure machines for nothing but the pleasure of getting the highest possible benchmark out of them. This applies to many areas, cars, bikes, boats and all sorts of things.

Since the original question was an unqualified "how can I get better scores" I think FUGGER did a very good job of listing some basic and more comprehensive steps on how to achieve that. How many of them the guy chooses to adopt is up to him.

Your point is valid, doing all of those steps would leave a fairly limited machine, however it all depends on the goals of the individual. Maybe you want to compare against the best possible scores your hardware could produce, knowing day to day, you run with a compromise because your XYZ requirements prohibit you from having your system 'optimal'?

my .02c

-* <font color=red> !! S O L D !! </font color=red> *-
To the gentleman in the pink Tutu
January 23, 2002 6:14:43 AM

Or you could overclock your videocard till it fries, and then buy a new cpu to get a higher score.

;-)

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
January 23, 2002 7:26:39 AM

LOL ... gee, I wonder who you could be referring to? *cough* FUGGER *cough*

<i>There are two theories on arguing with women. Neither one works.</i>
January 23, 2002 8:55:56 AM

Quote:
"turn off bios and video shadowing"
Hi,
Why did you suggest the above? I thought this would actually improve the perfoirmance? Is that not so?

<b><A HREF="http://www.lostcircuits.com/advice/bios2/5.shtml" target="_new"><font color=red>BIOS Guide</A></b></font color=red>

>><font color=blue><b>Video BIOS Shadow (Disabled)</b></font color=blue>

<i>In old graphics cards (1 MB video memory), copying the video BIOS into the system memory could increase the graphics performance of the system. The only drawback is that the video BIOS will take up base memory addresses and, thus, possibly interfere with other devices because of resource overlapping.

For all modern graphics adapters, the setting should be disabled since there is no performance increase noted and the possibility to encounter errors because of occupying other necessary system resources increases.

<b></i><font color=blue>System BIOS Shadow (Disabled)</b><i></font color=blue>

Once the BIOS is loaded, there is no reason to shadow it in the main memory. Similar as with the Video BIOS, it will only occupy base memory addresses and, thus, make them unavailable for other system devices.</i><<

:smile: Good or Bad have no meaning at all, depends on what your point of view is.
January 23, 2002 5:32:37 PM

Still jealous of my score matisaro?? get over it already.

You will never be able to <A HREF="http://gamershq.madonion.com/compare2k1.shtml?2516811" target="_new">keep up =)</A>

You didn't know when were crying like a baby calling liar that I was stacked on hardware. I forgive you.
January 23, 2002 8:05:57 PM

I never cared about your score, if you still dont realize what I and everyone else who called you on that score were talking about, you truely are dense.

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
January 24, 2002 11:52:59 PM

Holy Smoke !!! :eek:  :eek: 

How did you get a score like that, Fugger? Those Northwood processors must be real fast. I can't wait to get my hands on one of those.

Anyway, I tried some of the options you listed in your post above, but I could not get a higher score than 6202, it averaged to around 6185.

Wen the cup is ful,no more can be poured-
wen you'r ful of your opinions,no more wisdom can enter
January 24, 2002 11:58:40 PM

Quote:

Holy Smoke !!!

How did you get a score like that, Fugger? Those Northwood processors must be real fast. I can't wait to get my hands on one of those.

Anyway, I tried some of the options you listed in your post above, but I could not get a higher score than 6202, it averaged to around 6185.

Fugger seems to be amazingly good at tweaking his systems to the absolute maximum. You're looking at a 2.5GHz Northwood with a highly overclocked GeForce3 (I'm guessing nearly 300/600). Even if you get a 2.2GHz Northwood, you won't even get near his numbers without overclocking both the CPU and graphics card.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 25, 2002 5:50:26 AM

Quote:
How did you get a score like that, Fugger? Those Northwood processors must be real fast. I can't wait to get my hands on one of those.



If you want scores that high you WILL have to buy a new processor.(just like fugger did)

And you will have to strongly overclock your videocard, be careful you dont fry it.(just like fugger did).

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
January 25, 2002 5:51:41 AM

PS: your score seems just right, my friend got 5700 on a duron900/gf3ti200 combo everything at stock.(well the duron was at 133fsb)

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
!