Complete Offtopic: CSC

[The_Tonx]

Well. I made a little program (to calculate computer's power). And since it's so new, I'd like for you guys try it.
(I might see what computers are you using too (for DOOM III):p)

Here are some examples:

100 MHz 486 DX4 - 139
300 MHz Celeron A - 330
300 MHz Celeron A - 333 (little more RAM)
667 MHz Celeron - 397
800 MHz Duron - 473 (changed BIOS settings)
800 MHz Duron - 475
1300 MHz Athlon - 563
1400 MHz Athlon - 603

I had a report that one 1.2 GHz ThunderBird had ONLY 384 units.. :( So, what do you say?

If anyone is interested then I can send you it (CSC) :)

[Lord FlatHead]

What are these numbers based on ?

[Eraserhead2416]

Is this a program we'd download and run? What # would be good for Doom3? You have AMD processors, but what about Pentiums?

[The_Tonx]

Well. I'm not that good @ english, but I try..

The program shows you 2 lines on text (made from 2D graphical font) and calculates haw many times you'r system can show it. It is done 10 x 1 second and then you'll get your #.

CSC (public test) is downloadable from:
http://www.dako.edu.ee/~tonx/CSC.exe
(win32 self-extracting archive)

If you have downloaded it then run CSC.EXE. Hit "Test" and you'll get the #. If you wan't to get better result (or test you'r computer on same condition that the other computers (I've meantoned)) please make a CSC Boot Disk using MAKEBOOT.EXE. It takes 1 Floppy, but don't insert a Floppy where's important stuff on. CSC Boot Utility will (Quick)FORMAT you'r disk so.. :/
Anyway, if the disk is done then Boot you'r system from the Floppy and CSC should automatically start. Then hit the "Test" button and you should get better # :)

Hope it's not too hard to understand me :/ Anyway, it's more easier to do it than say it (atleast for me) ;)

[The_Tonx]

Eraserhead2416 said:

Is this a program we'd download and run? What # would be good for Doom3? You have AMD processors, but what about Pentiums?

Well. I can't tell you what # is good for DOOM 3, but if most of us download it and test their systems then we should see HOW FAST PCs they have.

Anyway, DOOM 3 and CSC are 2 completelly different things. CSC will not test you'r 3D graphics, but you can check you'r system OVERALL speed :/

About Pentium CPUs: I had no Pentium CPU to test the CSC, but there are some Celerons. Celeron 667 should be almost equal to 500 MHz Pentium III. But ask someone (who own's Pentium CPU) to test his system and you'll get the # for Pentium ;)

[The_Tonx]

I think that there is no major difference between Intel or AMD CPU # because Intel CPU's FBS > AMD's, but AMD manages it's FSB better than Intel.

As we have seen from other tests and benchmarks (did I wrote it right?) that Athlon XP 1900+ (1.6 GHz) can be faster or almost same speedy as 2 GHz Pentium 4.

[The_Tonx]

You may also check results from:
http://www.dako.edu.ee/~tonx/PC

and:
http://www.dako.edu.ee/~tonx/PC/reports.txt

[fraggle]

Thats something more likely to be affected by your video card than your processor. It certainly shouldnt be used to measure processor power in any way, shape or form.

[The_Tonx]

fraggle said:

Thats something more likely to be affected by your video card than your processor. It certainly shouldnt be used to measure processor power in any way, shape or form.

Why not? For example:

By setting by RAM latency from 3 to 2 my rates increased.
Also I experienced MAJOR differences by changing the L1 and L2 work..
By changing it (L1 and L2 cache) I got about 2 times less units..

Please note that the 2D graphics there is done WITH SOFTWARE RENDERING. That means that it is NOT ONLY affected by you'r GPU.

[fraggle]

The_Tonx said:

Why not? For example:

By setting by RAM latency from 3 to 2 my rates increased.
Also I experienced MAJOR differences by changing the L1 and L2 work..
By changing it (L1 and L2 cache) I got about 2 times less units..

Please note that the 2D graphics there is done WITH SOFTWARE RENDERING. That means that it is NOT ONLY affected by you'r GPU.

I'm not saying it isnt affected by the processor, but the ultimate bottlenecks here are the graphics card and the bus because you are drawing to the screen. On comparatively slow processors it may provide a quite good measure of CPU performance. However, on fast modern processors the speed will eventually be limited by the graphics card you are using.

In any case, drawing things to the screen doesnt provide a very comprehensive test of all the functions of the processor. You're only testing a few specific features of the processor. A real game would use many more.

[The_Tonx]

fraggle said:

I'm not saying it isnt affected by the processor, but the ultimate bottlenecks here are the graphics card and the bus because you are drawing to the screen. On comparatively slow processors it may provide a quite good measure of CPU performance. However, on fast modern processors the speed will eventually be limited by the graphics card you are using.

In any case, drawing things to the screen doesnt provide a very comprehensive test of all the functions of the processor. You're only testing a few specific features of the processor. A real game would use many more.

I agree..

The next CSC (2) shouldn't hace this problem anymore, because I no longer show that 2D graphics.. I just let the program to see how many times you machine can do 1+1 :) That should fix that "problem", right?