PDA

View Full Version : Parhelia got owned by Xabre 600!



mikeul
2nd March 2003, 10:39
http://www.hardwaretidende.dk/hard/artikelimages/18022003-7.png

Notice how the Xabre 600 64 Mo eats the Radeon 9500 64 Mo, Radeon 8500 128 Mo, Parhelia 128 Mo, etc. for breakfast
And notice the Parhelia is LAST, even BEHIND THE GEFORCE4 MX!

3dfx
2nd March 2003, 11:08
Parhelia was never designed for gaming at that rez tbh, that's what FAA 16x is for ;)

Tomasz
2nd March 2003, 11:49
Parhelia was never designed for gaming at that rez tbh, that's what FAA 16x is for
Still, it is pretty sad that the Parhelia can't even muster half the fps of cards that cost 1/4 of its price. Hell, even the Xabre 400 owns it. I sure hope this is due to some driver issues. Otherwise, this means that the Parhelia's design is very inefficient.

Kruzin
2nd March 2003, 13:01
How did any of those cards score at 3072x768?
Oh yea...

KeiFront
2nd March 2003, 13:02
Probably driver issues: http://forums.murc.ws/showthread.php?s=&threadid=35525&highlight=rtcw

mikeul
2nd March 2003, 13:27
"Hell, even the Xabre 400 owns it. I sure hope this is due to some driver issues."

Maybe, but Matrox is more well known for good drivers than SIS...

Tomasz
2nd March 2003, 13:33
Maybe, but Matrox is more well known for good drivers than SIS...
True, but Matrox has always had problems in OpenGL.

K6-III
2nd March 2003, 13:40
Rather, it is issues that the Parhelia has with the Q3 and Q2 engines...

borat
2nd March 2003, 15:23
Originally posted by 3dfx
Parhelia was never designed for gaming at that rez tbh, that's what FAA 16x is for ;)

i got one and i thought the whole idea with it was that it performed rubbish at 800x600 but you could crank up the res and eye candy without taking a hit, 16xfaa is a kop out if it cant perform decent frame rates at resolutions without and aa.9i love the card as it plays the games i use fine but i am not going to defend it on this one)

borat
2nd March 2003, 15:24
Originally posted by Kruzin
How did any of those cards score at 3072x768?
Oh yea...

right on!!!:):):):):):):):):):)

DGhost
2nd March 2003, 17:23
Originally posted by Kruzin
How did any of those cards score at 3072x768?
Oh yea...

you know... according to that chart the Parhelia should run ~20FPS (less probably) at 3072x768... while all the other cards would remain playable if they could run at that resolution...

K6-III: what Thomaz said is more accurate in this case... its not just the Q2/3 engines that are having issues, there are problems in OpenGL that are affecting way more than just those...

3dfx: i wish the Parhelia wasn't designed to run at those resolutions. unfortunately, that is not the case. especially considering that 1600x1200 is less demanding than the 3072x768 surround gaming mode that Kruzin felt the need to bring into this....

Chrono_Wanderer
2nd March 2003, 19:36
I notice my following example has many flaws in them ;) but o well it's an idea...

1600x1200
= 1920000 pixels

3072x768
= 2359296 pixels

So... a 1600x1200 scene is approx. 81.4% of a 3072x768 scene...

now let's take the second worse card on the chart, the Readon 8500LE 64MB for comparison...

42.5fps * 81.4%
= 34.6fps...

34.6fps vs. 22.9 fps. for the R8500LE to render 2359296 pixels...

**sniff sniff** :(

Kruzin
2nd March 2003, 21:19
What good does it do to calculate "theoretical" possibilities, when the card can't do it?
Even if P got 5 FPS, it would be 5 more than the others can dream of doing.

Kruzin
2nd March 2003, 21:29
You know what...I don't why this thread is even here.
It's clearly a benchmark thread, so it belongs in the benchmark forum.
Off ya go...

DGhost
2nd March 2003, 22:14
Kruzin - who gives a shit if you can play the game on 3 monitors if it is unplayable at that resolution?

Chrono_Wanderer
3rd March 2003, 02:19
Originally posted by Chrono_Wanderer
I notice my following example has many flaws in them ;) but o well it's an idea...

heh. i noticed the problem with the theoretical estimation. ;)

edit: oh yea. i guess the est. sort of reflects on how well the core was built.

P5ycho
3rd March 2003, 04:57
Originally posted by Kruzin
What good does it do to calculate "theoretical" possibilities, when the card can't do it?
Even if P got 5 FPS, it would be 5 more than the others can dream of doing.

Just look at what you've just posted, you'd rather play a game at an obviously unplayable framerate just because your can do it and no other card can... well i'd switch to a healthy 60+ fps on a singlehead card anyday if that were my case!

WyWyWyWy
3rd March 2003, 07:43
SOmething is wrong there.
Even GF4mx is faster than 9500?!

Pace
3rd March 2003, 08:52
Originally posted by Kruzin
What good does it do to calculate "theoretical" possibilities, when the card can't do it?
Even if P got 5 FPS, it would be 5 more than the others can dream of doing.5fps would be completely unplayable though. So, none of those cards would be any good for SurroundGaming.

An average of 23fps is pretty much unplayable as well - does anyone play RTCW at playable frame rates and, if so, what settings?

Bit by bit the Radeons get more and more tempting...

GNEP
3rd March 2003, 08:55
Actually, anyone fancy double checking that score with their P and RTCW? I would, but don't have RTCW...

KeiFront
3rd March 2003, 10:52
Here we go, thread has been moved to the benchmark section so lets benchmark ;):

AMD XP 1800@2000, SDRAM, Parhelia A128R (not oc)
Latest Parhelia driver: 1.3.0.43.
All settings at highest.

Checkpoint
1600x1200 32Bit without FAA
average of 3 benchmark runs: 52.1 FPS

1600x1200 32Bit with FAA
average of 3 benchmark runs: 34.5 FPS

Surround gaming
3072x768 32bit without FAA
average of 3 benchmark runs: 49.1 FPS

3072x768 32bit with FAA
average of 3 benchmark runs: 34.4 FPS

K6-III
3rd March 2003, 11:52
Based on your results, looks like the review was very flawed...

KeiFront
3rd March 2003, 12:43
Probably, depends on which drivers version they used. Or maybe game settings.

mikeul, what's the source of your picture. So we can see which driver revision they used.

Novdid
3rd March 2003, 13:23
I have a R9700Pro and play the game at 1600x1200x32, some form of FSAA and Aniso. Never bothered to bench it, if it just hovers around 50 I'm happy (but I get more than that;)). That's in single player. In multi player I'm more into speed and back down to 1024x768 4xAA and 16xAniso. Some times I'm tempted to run the game (or any other game for that matter) with 6xAA just because it just looks so gorgeous.:)

KeiFront
3rd March 2003, 13:32
I found the source of the picture and I'm currently rebenching.

borat
3rd March 2003, 13:33
i think you will find that surround gaming does not actually hinder the card as much as you think as when running the 2 side screens thaey are not usually subject to intense action but merely displayin surroundings so the graphics card is not so highly stressed, also i dont know how much they put on the card but things like the physics of the game also remain constant, i certainly can play smoother in ut2003 at 3072x768 than at 1600x1200, aniso and faa of course:)

KeiFront
3rd March 2003, 13:37
http://www.digit-life.com/articles2/digest3d/index.html You can download the config they used.


Return to Castle Wolfenstein (MultiPlayer) (id Software/Activision) - OpenGL, multitexturing, Checkpoint-demo, best test settings, S3TC OFF, configs are here

Still getting a lot more than their score though:
1600x1200 32 bit
2945 frames, 80.2 seconds: 36.7 fps with FAA
2945 frames, 52.8 seconds: 55.8 fps without FAA

http://www.digit-life.com/articles2/digest3d/0103/itogi-video-q3std-w2k-ath1600.html

K6-III
3rd March 2003, 14:42
Figured as much. Digitlife always buchers the Parhelia's performance. Wonder why....

borat
3rd March 2003, 15:20
hmm and crank up the anti aliasing on the xabre and radeons and what do we get?:)parhelia bashing is not really deserved, i mean its perfectly playable with faa at high resolutions even in games which the drivers suck with.

borat
3rd March 2003, 15:24
how about this graph with shite drivers too:
http://www.digit-life.com/articles2/digest3d/0103/itogi-video-ut2003-w2k-aaa-ath1024.html

Novdid
3rd March 2003, 15:38
Originally posted by borat
hmm and crank up the anti aliasing on the xabre and radeons and what do we get?:)parhelia bashing is not really deserved, i mean its perfectly playable with faa at high resolutions even in games which the drivers suck with.

I agree that the digit-life guys paint a bad picture of the Parhelia, but you have to take into consideration that those boards with lesser AA modes costs 1/4 of the P...

UberLord
3rd March 2003, 17:12
Originally posted by borat
hmm and crank up the anti aliasing on the xabre and radeons and what do we get?:)parhelia bashing is not really deserved, i mean its perfectly playable with faa at high resolutions even in games which the drivers suck with.

I dunno about the Xabre's, but the 9700 pro's AA isn't to be sniffed at. Admitedly the quality of the P.'s AA is superior to any other card for the home user, but the 9700 Pro's x4 AA isn't that far off IQ-wise AND it can get much higher FPS using AF+AA than the P. can.

So yeah, against the ATI 9700 Pro, the P. bashing is deserved :rolleyes:

GNEP
4th March 2003, 01:34
Originally posted by KeiFront
Still getting a lot more than their score though:
1600x1200 32 bit
2945 frames, 80.2 seconds: 36.7 fps with FAA
2945 frames, 52.8 seconds: 55.8 fps without FAA

http://www.digit-life.com/articles2/digest3d/0103/itogi-video-q3std-w2k-ath1600.html

Hmm... thought as much :rolleyes:

Thanks KeiFront. Yours is a similar machine to their test bed too by the looks of things (isn't 2GHz about what the 2600+ runs at?)

KeiFront
4th March 2003, 09:22
Originally posted by GNEP
Yours is a similar machine to their test bed too by the looks of things (isn't 2GHz about what the 2600+ runs at?)
Sorry for the confusing :speechl:, my machine is running at 1.67Ghz (AMD XP1800+@2000+) with SDR ram, their processor is running at 2.13GHz (AMD XP 2600+) with DDR ram.

Stupid ratings ;).

GNEP
4th March 2003, 09:30
Ahh! And you've changed your sig to clarify now I see... :) So basically they should be getting a LOT more than they did in their "test"?

KeiFront
4th March 2003, 09:46
Originally posted by GNEP
Ahh! And you've changed your sig to clarify now I see... :)

Yep, I changed my sig from XP 1800@2000 to XP 1800+@2000+ :bandit: (stupid ratings ;))


Originally posted by GNEP
So basically they should be getting a LOT more than they did in their "test"?

They certainly should got better scores than me.

I also tested Ut2003, 1024x768, AntiAliasing (FAA) + Anisotropy (2X)
Flyby: 50.043545 FPS
Botmatch: 28.260815 FPS

borat
5th March 2003, 02:00
Originally posted by UberLord
I dunno about the Xabre's, but the 9700 pro's AA isn't to be sniffed at. Admitedly the quality of the P.'s AA is superior to any other card for the home user, but the 9700 Pro's x4 AA isn't that far off IQ-wise AND it can get much higher FPS using AF+AA than the P. can.

So yeah, against the ATI 9700 Pro, the P. bashing is deserved :rolleyes:

yes the 9700 pro is faster and its quality is there with parhelia, but parhelia is perfectly playable at these settings, i never said that 9700 pro was not fasster than it, when you buy parhelia you are buying into triple head, image quality and decent gaming performance, couple this with the fact that it was realeased months before the 9700 and it isnt that bad, the reason it gets a bashing is because everyone tests it without faa on and say "ooohh the geforce/radeon can produce 30000 fps at this setting but the parhelia can only do 100" turn everything on and it runs silky smooth.if i was buying now i think id go radeon but im happy with parhelia so ill wait till its due its upgrade.

SteveC
5th March 2003, 06:12
It's not just the results that are a bit low for the Parhelia, but the core speed! Only the GF3 has a lower core speed. How I'd love to see a Parhelia running at 350Mhz Core speed....

borat
5th March 2003, 06:16
Originally posted by SteveC
It's not just the results that are a bit low for the Parhelia, but the core speed! Only the GF3 has a lower core speed. How I'd love to see a Parhelia running at 350Mhz Core speed....

very true, although if it came top of all the charts it could tun at 1mhz and no one would care.

Kastuvas
6th March 2003, 05:24
How do you run the benchmark on RTCW?

KeiFront
6th March 2003, 09:48
First download the checkpoint demo (rtcw-checkpoint.zip) you can find it at: http://www.3dcenter.de/downloads/rtcw-checkpoint.php

Create a new directory in your RTCW "main" folder and call it "demos" e.g. "X:\Program Files\Return to Castle Wolfenstein\Main\demos"

Extract the checkpoint.dm_60 file into the "demos" folder

Now start RTCW multiplayer and bring up the console by hitting tilde key (`).

type /timedemo 1 (enter)
type /demo checkpoint (enter)

Let the demo run (it will take about a minute) and when it returns to the main menu hit tilde (`) again to see the framerate.

Kastuvas
6th March 2003, 17:17
Thank you KeiFront,
I got 29.1 Fps @1280x1024x32 (no FSAA, no anisotrpic)

KeiFront
7th March 2003, 11:15
Your're welcome. Pretty good score with that processor and mobo.