What do we want to see in a review?

[Ali]

What does everybody thing needs to be looked at when the Parhelia is finally reviewed?

Im thinking along the lines of:
Quake3 at low res/detail
Quake3 at highest res/detail possible

That will show us what type of performance drop we get. And show a little bit about driver maturity.

Then I would like to see that UT2003 demo thing that Anandtech has, showing what it can do for modern games.

I suppose 3dMark, just so we can tell how powerful the PS/VS are.

There is a AA comparison done here:

http://www.beyond3d.com/articles/wildcatiii/index.php?page=page4.inc

And I would like to see the Parhelia put through these tests with both FAA and 4XFSAA

The hardware AA fonts should be looked at in comparison to the Microsoft software ones, and see if the gamma correction makes much difference.

I would also like to see some benchmarks that have low/average/high fps readings done, at different resolutions and details levels. This would be very interesting to see in relation to the competition.

Surround gaming I think would be VERY hard to review. What else could you say except it looks cool. Without the reader having 3 monitors to try for themselves, its a little pointless. (my opinion anyway).

2D quality will need to be looked at. Im assuming its going to be leaps and bounds ahead of everything else though.

Im also curious as to how noisy the fan is on the card. A lot of people now days are going to the trouble of making their computers as quiet as possible, I know I am (water cooling without OCing!), and I would find a one paragraph impression interesting.

What else should be looked at?

Ali

[DukeP]

How easy it is to pry the standard cooler off?


What about high resolution output? 1600x1200x32bit 2d speed and clarity...

Video playback.
Output to TV/Video.

Oh. And all of the above

~~DukeP~~

[mute]

NO QUAKE 3...

Use jedi knight 2 , or RTCW if you want to bench the quake 3 engine

no test below 1024x768
high detail no aniso , no FAA , no fsaa
and then a test with high detail , highest aniso , and FAA and/or fsaa to see the perfomance drop

here's a list i think you should use

ut2003
3dmarks2001
codecreatures
morrowind

[Marshmallowman]

year I agree JK2 and RTCW, no quake 3

ut2003,
and anything that uses new hardware
aquanox?
xisle tech demo(or is that nvidia only)
carmageddon TDR

[Dr Mordrid]

I'd love to try codecreatures out, but their site links to download.com and download.com says it's no longer available. Got another link?

Dr. Mordrid

[Dr Mordrid]

Nevermind. Got it at MajorGeeks;

http://www.majorgeeks.com/article.php?sid=1256

Dr. Mordrid

[Drizzt]

I'm with BuddMan. Use Morrowind for the test, too.

[Ali]

Well, I still think there is a place for Quake 3.

It doesnt do a good job on stressing video cards to the max anymore, but for me it will show the difference between my G400 and the card Im almost certainly going to upgrade to.

If someone with Windows98 installed on their computer, who also has a highend CPU tested Quake3 with the G400 release drivers, and again with the newest drivers and worked out a $age increase over many resolutions, this would make a good comparison to the %age increase we could expect from Parhelia when the drivers mature.

For me personally, I dont realy care about how fast it goes in a benchmark like Codecreatures, because I dont have a reference for it (from G400). I agree we need benches with modern games/apps, but the older ones are still needed to relate it back to older hardware.

I would also be interested in how it performs on different platorms. eg AMD lowend (Duron 900 or something), AMD highend (XP2100+), Intel lowend (Celeron 900ish) and the fastest P4 available.

I know the lowend CPUs arnt really that lowend in the grand scheme of things, but if you can afford a Parhelia, then you probably have at least a 900Mhz PC.

The %age drop when different settings are enables (anso, AA etc) mapped against the same data for other cards (8500, Geforce3&4Ti, G550) would be great to see. you would also need quality comparrisons here too though, due to different implementations of the features.

Ali

PS, Im not going to do a review, as I dont have a website, the hardware, the time, or the inclination. Im just interested to see if the people who do do reviews will give us whet we want to see.

[GT98]

I personally like MOH:AA better then RTCW...is there that big of a difference between the two?

[oblisk]

For games:

Pick 2 or 3 modern game engines and run the cards through there paces. That is really all that needs to be done for games.

For 3d Apps:

ProE
AutoCad
3DSMax
Maya

There is another app i know which should be benchmarked/infomation given on limitations etc. alas i cannot remeber it for the life of me.

Video Editing/Capturing:

Premeir?
Im not sure of any good benchmarks in this area but there should be some

2d

Again i have no idea what benchmarks are availible

Benchmarks should also be done on low, medium, high and
workstation machines (and with both amd and Intel parts)

I.e.

low end:
P3-600/K7-600

Medium:
Athlon(duron) 1ghz/ P4(P3/cel)-1.2~

High:
Athlon XP 2100+/P4-2.4

Workstations:
Dual-MP / Dual P4 Xeon

Though this will not give a Comprehensive benchmark, it should be quite thourough (sp?) and more rounded than the majority of benchmarks out today.

[BuddMan]

That's odd. I even posted in this thread and the post was deleted. Even Drizzt responded to it. I wonder what's going on????

Anyway, back to topic, I think they should have Morrowind tested. I mean, the game's a beast and can bring any system down to it's knees. I'd like them to do comparison's with other cards on how it runs the game with everything at max including the use of pixel shaders.