WTF?? 2.4P4 slower than Athlon XP 2000!

feylya · 17-09-2003 7:49pm #1

I have a 2.4b P4 on a MSI 865PE Neo2 mobo, 512MB RAM and a GeForce FX 5600 Ultra. I ran AquaMark3 and I got 15,300. My brother ran it and got 18,300 on a Athlon XP 2000 and a GeForce 4 Ti. Can any body sort me out?

weemcd · 17-09-2003 8:34pm

athlons are better, imo

feylya · 17-09-2003 8:36pm

1) I didn't ask for your opinion
2) Look up some benchmarks.

SouperComputer · 17-09-2003 9:37pm

swap the GFX on the machines & use the same version of dx and see what happens.

Jam · 17-09-2003 9:49pm

Isn't Aquamark 3 a pure dx9 benchmark? if you've been keeping up with the HL2 news, FXs do pants when it comes to dx9. So I think you're comparing apples to oranges. The GF4 is running in dx8.1 while the FX is computing all the addional pretties dx9 offers so it gets less FPS.
Maybe?

RicardoSmith · 17-09-2003 9:53pm

Benchmarks are almost never telling the full story. Swap the Vid Cards and see whats the difference. be interested in hearing the result.

feylya · 18-09-2003 7:12am

I might try swapping the graphics cards. However, in AquaMark, there's a cpu test. He beat be by about double! Something is a bit dodgy and I don't know wtf it is! I'm praying Detonator 50 will be as good as it should be.

SouperComputer · 18-09-2003 9:47am

i dont know anything about aquamark, but insofar as the CPU benchmark, could it be a case of commercial influence on AMD's part. Double is a lot though, AMD\P4 debate aside.

Personally, based on subjective experience so far I havent seen any benefit of either chip.

Perhaps a bottleneck on your end with RAM or you need more aggressive BIOS settings?

feylya · 18-09-2003 10:14am

Any chance you could post some ram timings? I have ****e ram, just the stuff I got with the Komplett upgrade pack thing.

Sir Random · 18-09-2003 12:20pm

Aquamark is primarily GPU based and the ti4600 is a better card than the FX5600.
I suppose you're aware of the FX problems with DX9 so maybe that's slowing up the Aquamark or did you use the leaked 51.75 drivers?

EDIT: Just noticed the cpu marks, what cpu score do you get? how does it compare to all the other P4 2.4s?
My P42.4b scores 8239 cpu marks

jesus_thats_gre · 18-09-2003 12:27pm

Originally posted by weemcd
athlons are better, imo

Jaysus so bloody helpful.

Even a 1.7 P4 with a GF 4200i is beating him at thit stage.

feylya · 18-09-2003 1:44pm

Just for an experiment, I clocked it back to 2.4 (was at 2.6) and it got 19000+. My house mate who has a 1.7ghz P4 and a Geforce 4 ti got 20,000.

CivilServant · 18-09-2003 3:14pm

What drivers are you using?

Definitely check what your cpu is running at using wcpuid or cpu-z. These proggies will tell you exactly what your cpu is doing.

To get best performance people recommend clean install of os.

feylya · 18-09-2003 4:17pm

Detonator 45.23 The cpu is running at 2.4 or 2.6 depending on what I have it clocked at. It may be the ram but I don't know.

Surely my machine should beat the 1.7 p4, even with the less than great graphics card.

feylya · 18-09-2003 4:19pm

Originally posted by Sir Random
Aquamark is primarily GPU based and the ti4600 is a better card than the FX5600.
I suppose you're aware of the FX problems with DX9 so maybe that's slowing up the Aquamark or did you use the leaked 51.75 drivers?

EDIT: Just noticed the cpu marks, what cpu score do you get? how does it compare to all the other P4 2.4s?
My P42.4b scores 8239 cpu marks

Eh I think I'm getting about 3000. I'll say it again - WTF?

CivilServant · 18-09-2003 4:54pm

My athlon at 2.1Ghz only scores around 6500 cpu marks and my graphics ti4400 oc to 4600 scores ~8500. Total 23,000. Maybe mine should be higher also. Not too worried though cos it's a synthetic benchmark. Plays ut2003 fine, with everything on max.

On a side note what 3dmark2001 do you score with that setup? If you download the 51 detonators will improve your score cos nvidia (the cheatin bolloxes) have switched from 32bit fp to 16bit fp. With a slight loss in imagine quality, but it will improve your score no doubt.

feylya · 18-09-2003 4:58pm

I can't remember what I got with 2001 (I'll run it later). With 3d Mark 2003 (yes, I know it's bollox), I got 2701

Not too shabby.

CivilServant · 18-09-2003 5:12pm

Whoops ... I got the gfx and cpu scores mixed up. ~8500 for cpu and ~6500 for gfx. That sounds right acutally. No more tweaking needed thank ****.

I almost get 14k rock stable with 2001, over 14k won't play ut2003 without crashing. 1800 for 3dmark2003.

I got 2701 Not too shabby.

Not bad for that card. Why didn't u get a radeon instead?

feylya · 18-09-2003 5:14pm

Could only afford the GeForce. It was about €150 (and Komplett sent me the Ultra by accident) and the Radeon was more expensive. Didn't have much money at the time so I just went for it. I could sell it for €200 I suppose

CivilServant · 18-09-2003 5:25pm

I recommend buy&sell

feylya · 18-09-2003 5:27pm

Ah, I'm sure there a 1st monkey in the college who wants to impress his mates by having a ****e GFX card. I'll think about it anyway.

feylya · 18-09-2003 6:34pm

I just ran 3d Mark 2001 SE - 8500. I hang my head in shame. Running @ 2.6GHZ with GFX card @ 350 mhz chip with 750mhz memory, running Detonator 45.23 on Windows XP Pro sp1 with DirectX 9.0

CivilServant · 18-09-2003 7:56pm

Hmmm, u should be gettin 10k for new €150 card. I started at 12k before some ocing and tweaking the cpu and gpu. 2.4c should be able to clock up to 3Ghz no prob. Just take it easy and watch those temps, u should get there.

feylya · 18-09-2003 8:22pm

it's only a 2.4b. I don't know what's the deal. Could you post pictures of your aquamark scores please? Just interested.

CivilServant · 18-09-2003 9:12pm

Memory's gone to shreds in me old age, here's what I really got.

2.1Ghz Athlon
Geforce 4 Ti4800SE 320/660

feylya · 18-09-2003 9:15pm

There's something really not right with my computer. I'm being beaten by Athlon's and GeForce 4's.

/me goes off and shoots himself

RicardoSmith · 18-09-2003 9:24pm

Heres some benchies I did abouyt a year ago. I didn't faff about with the memory settings timings or anyhting. The scores weren't consistant toward the end for some reason.

3D Mark 2001(SE I think) (1024x768 32bit colour)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Axia 1ghz at 1ghz (133/100) + GF2Ti450 (451/250) = 3768
Axia 1ghz at 1.4ghz (140x10) + GF2Ti450 (451/250) = 4432
Axia 1ghz at 1.4ghz (140x10) + GF2Ti450 (481/275) = 4577
Axia 1ghz at 1ghz (133/100) + GF4Ti4200 (default) = 6578
Axia 1ghz at 1.3ghz (133x10) + GF4Ti4200 (default) = 7452
Axia 1ghz at 1.4ghz (140x10) + GF4Ti4200 (default) = 7696

P4 1.6a at 1.6ghz (100fsb) + GF4Ti4200 (default) = 8241
P4 1.6a at 2.0ghz (125fsb) + GF4Ti4200 (default) = 9312
P4 1.6a at 2.1ghz (??fsb) + GF4Ti4200 (300/550) =10297

P4 1.8a at 1.8ghz (100fsb) + R9000 Pro (default) = 6901
P4 1.8a at 2.4ghz (133fsb) + R9000 Pro (default) = 7440
P4 1.8a at 1.8ghz (100fsb) + GF4Ti4600 (default) = 9378
P4 1.8a at 2.4ghz (133fsb) + GF4Ti4200 (default) = 9699
P4 1.8a at 2.4ghz (133fsb) + GF4Ti4600 (default) =10370
P4 1.8a at 2.4ghz (133fsb) + GF4Ti4600 (default) =11428
P4 1.8a at 2.4ghz (133fsb) + GF4Ti4600 (default) =10422

P4 1.8a at 2.4ghz (133fsb) + GF4Ti4200 (300/550) = 10120

CivilServant · 18-09-2003 9:24pm

Geforce 4 is an excellent performer when it comes to dx 8 graphics. Aquamark isn't really (in my opinion and others) a fully fledged dx 9 benchmark anyways. It only uses 7 PS2 shaders, which if memory serves was only about 10% of the total shaders used.

I'm being beaten by Athlon's and GeForce 4's.

Athlons rock!:D

feylya · 18-09-2003 9:25pm

AquaMark isn't a DirectX 9 test! And athlon's aren't better. ARGH!

CivilServant · 18-09-2003 9:36pm

Athlons aren't better but they DO rock!

Aquamark 3 is being labelled as a dx9 test when it's really more of a medley of dx 7,8 and 9 tests. If you want a good dx9 benchmark run nature on 3dmark2003, X2 demo or Gun metal demo.

feylya · 18-09-2003 9:39pm

Nature has been run. It looks nice. The Pixel Shaders 2.0 test looks even nice

CivilServant · 18-09-2003 9:49pm

Hoho! I can't even watch that eye candy! :rolleyes:

/me waits for HL2 n 9800

feylya · 18-09-2003 9:58pm

Mmmm, hl2. Hopefully the new Rel 50 drivers might do something for me.

RicardoSmith · 18-09-2003 10:07pm

Benchies are one thing. In game FPS are another. Often they give very different scores. So don't go by benchies alone.

feylya · 18-09-2003 10:12pm

But benchies are cool for pushing your machine to the max and for gloating rights. Unlike someone I know who doesn't like benchmarks cos "they're not real" or graphical demos because "they're not the game".

RicardoSmith · 18-09-2003 11:08pm

I suppose if you get your kicks from doing benchies thats fair enough....

feylya · 18-09-2003 11:12pm

Originally posted by RicardoSmith
I suppose if you get your kicks from doing benchies thats fair enough....

Yes, yes I do.

feylya · 19-09-2003 12:56am

WTF?!?!?!?!

netman · 22-09-2003 10:38am

dude, that score is pretty low for your setup.

i'm not sure if it's the 5600 ultra, or there's other issues on your pc, but i'm getting 44086 in aquamark 3, and it's not top-of-the-range pc, just a decent setup: 2.6c with radeon 9700pro.

try checking the CPU clock with something like wcpuid or cpu-z.
check the memory timings (cpu-z will give you those).
try the latest drivers from nvidia.

if nothing else, dump your card and get a 9500 non-pro modded to 9800 for roughly the same price. if i knew of the mod at the time when i was buying mine i would have went for that.

feylya · 22-09-2003 12:07pm

Do you have hyper threading enabled on that chip? I only have a 2.4B. Although, my cpu score is being beaten by a P4 1.7Ghz. And I had to overclock to 2.7Ghz just to get that 21000!! Somethings not right...

CivilServant · 23-09-2003 6:03pm

Yeah p4 'c' chips have hyperthreading enabled. I've heard this benchmark likes plenty of bandwidth. And since netman has a 'c' chip defaults are 200Mhz fsb and he's probably pushed it a little further. Mine was at 185 fsb so that helped a little too.

netman · 24-09-2003 10:08am

it's hyperthreading enabled, and i have dual channel memory.

haven't really noticed any difference in games with HT enabled or disabled, but maybe aquamark gives you brownie points for turning it on

feylya · 24-09-2003 10:10am

I think I've narrowed it down to the ram bottle-necking everything. Everything else is good components. I shall now officially start the "Buy Feylya 1 GIG of Decent RAM" charity

feylya · 08-10-2003 5:13pm

Just thought I'd let you know that the leaked Det 52.13 drivers give a decent increase. Just got 22,000 in Aquamark3 running @ 2.7 ghz.

WTF?? 2.4P4 slower than Athlon XP 2000!

Comments