Anand's Benchmark Statements in his 9800 XT Review [Archive]

View Full Version : Anand's Benchmark Statements in his 9800 XT Review

Deathlike2

10-01-03, 03:25 PM

While looking at this article... I've seen so much BS.. it's not remotely funny..

I didn't exactly like the fact he used non-official Dets... but it would be nice if we had access to these dets (any Det after the 51.75) to see for ourselves the differences (even the creator of Rivatuner has had access to the 52.10s, Anand had the 52.14s)...

Too much supposed optimizations that somehow seem "illegitimate to both ATI and NVidia" with some emphasis on ATI... but I'm not going to complain about that too much here...

Note the use of Neverwinter Nights:

http://www.anandtech.com/video/showdoc.html?i=1890&p=13

Neverwinter Nights was designed with NVidia extensions in mind... it's a rather unbalanced comparison to say the least (although people could make arguments with the Glide/Direct3D/OpenGL benchmark days, that's not the scope of my discussion)

Obviously NVidia would win this...

Note the use of Splinter Cell:

http://www.anandtech.com/video/showdoc.html?i=1890&p=15

He fails to mention that AA is not useable in this benchmark.. yet he mentions it in his Halo benchmark...

http://www.anandtech.com/video/showdoc.html?i=1890&p=9

Maybe I should stop reading Anandtech... this is the most flawed review ever.. honestly if Anandtech was to get better at all is to consult Beyond3D for bits and pieces of info that they are missing (or need looking into)

In any case... I'd like to see what you guys think...

ReDeeMeR

10-01-03, 04:42 PM

I dont know why people were praysing anandh, they always seemed biased to th big boys to me. And usualy skip the important bits showing the bad parts from either nvda or intel.

StealthHawk

10-01-03, 05:26 PM

Anand's reviews always have weird quirks. See the NV35 review with incorrect Quake3 FSAA numbers. See the Splinter Cell numbers where Anand accuses ATI's FSAA of not working properly(he doesn't realize that FSAA cannot be enabled without errors and NVIDIA's driver disabled FSAA from being turned on).

ChrisRay

10-01-03, 05:28 PM

Note the use of Neverwinter Nights:

http://www.anandtech.com/video/showdoc.html?i=1890&p=13

Neverwinter Nights was designed with NVidia extensions in mind... it's a rather unbalanced comparison to say the least (although people could make arguments with the Glide/Direct3D/OpenGL benchmark days, that's not the scope of my discussion)

I disagree here. I think whether the game is biased towards the hardware or not. I think its worth noting that hardware runs paticular games better than others.

extensions or not. As long as this is not the only game used. I see this as a good thing.

I would have "Loved" a NWN benchmark when I replacing my Ti 4200 with my 9500 Pro. it woulda been a wakeup call at the time.

StealthHawk

10-01-03, 05:47 PM

I agree, a game is a game. Whether the game itself is biased towards one IHV is irrelevant. If it's a popular game people will play it. Reviews should be geared towards showing potential buyers how a card will perform in the games they play. If no games are tested that people actually play the review is worthless. Just because NWN performs better on NVIDIA cards doesn't mean people with ATI cards don't play it. NWN is popular, and I think we need more depth in reviews besides the use of just first person shooters. Not everyone plays only first person shooters. It may have been true that in the past only FPS games could make a graphics card flex its muscles, and that for other games high end cards were useless. But with the advent of AF and FSAA high quality video cards are useful for all genre of games.

Deathlike2

10-01-03, 05:48 PM

Well.. the only reason why Neverwinter Nights runs better on NVidia cards than ATI cards.. is the extensive use of NVidia extensions...

Using NVidia extensions defeats the purpose of trying to make the game work for everyone... as when ATI are attempting to run these extensions... it will fail miserably and perform poorly (same goes for ATI extensions on NVidia hardware)

Anand SHOULD have mentioned that this game was optimized towards NVidia..

You (ChrisRay) need to read up on WHY Neverwinter Nights runs poorly on ATI hardware.. and it's not because ATI's hardware is inferior.

StealthHawk

10-01-03, 05:51 PM

ChrisRay

10-01-03, 08:28 PM

Originally posted by Deathlike2
Well.. the only reason why Neverwinter Nights runs better on NVidia cards than ATI cards.. is the extensive use of NVidia extensions...

Using NVidia extensions defeats the purpose of trying to make the game work for everyone... as when ATI are attempting to run these extensions... it will fail miserably and perform poorly (same goes for ATI extensions on NVidia hardware)

Anand SHOULD have mentioned that this game was optimized towards NVidia..

You (ChrisRay) need to read up on WHY Neverwinter Nights runs poorly on ATI hardware.. and it's not because ATI's hardware is inferior.

Uhh. Are you implying I dont know why? I find that attitude from you obnoxious. First we need to know that DOES before we can even get into why.

Knowing WHY is irrelevent if you dont know that it DOES. And this is a very important piece of information I had been unaware at the time. So Please dont patronize me that I dont research products for I buy them. There was little to no information about NWN and the r300 architecture at the time.

Secondly. It's not completely about the extensions. The game wasnt even making full use of the r300 architecture at the time. The 8500 Ran this game about 2x faster than a 9500 Pro w/o AA/AF enabled. That is Disturbing. And that has absolutely nothing to do with Nvidia. Because my 8500 was supposed to be an indication of how I thought the game would perform on the 9500 pro I was buying. Guess What. it wasnt It was in fact half the speed of my 8500 AIWDV at the standard settings. Do you not find that paticularly disturbing?

Knowing this would have averted a decision I made at that time Because I was heavily into the game. After I got the 9500 Pro. I literally had to stop Playing that game for almost 6 months. Or put my 8500 or Geforce 4 TI bacl in.

Why is the question you ask after the fact. Not before the fact. Why becomes irrelevent if you play said game anyway. If it's not going to be improved upon. And does it really matter? Who cares if the game is optimised for Nvidia. Because frankly. Its Irrelevent. Because the game is producing the exact same quality on my Geforce 4 TI as it does my Radeon 9500 Pro. So what difference does it make? In the end all I care about is how the game looks and play. And the only Image Quality differences I saw was actually poorer on the 9500 Pro than my Geforce 4 Ti (this was in regards to the reflective pixel Shaded water, This could have been fixed, But we're talking about then and not now right?) Where the Geforce 4 Actually produced better image. And Ran the game faster.

Isnt that what this is all about? Faster frame rates. Higher Quality? Your complaints about a game being optimised for are pretty ridiculous. Since Quality is the same and if not better.

P.S I am referring my 8500 All in Wonder DV which is clocked at 230/390. Which is significantly slower than a standard 8500.

Now I may have seemed a bit annoyed by your reply./ But I have educated myself. But Nothing at the time could have prepared me for what I saw when I first loaded that game. Completely unplayable frame rates.

ragejg

10-01-03, 10:10 PM

///OFF TOPIC///

Originally posted by ChrisRay
P.S I am referring my 8500 All in Wonder DV which is clocked at 230/390. Which is significantly slower than a standard 8500.

ChrisRay: remove the stock hsf (screwdriver wrapped in 2 layers of paper towel levered on all 4 corners), remove the blue crap with a razor blade, gently lap the gpu, sand teh hsf smooth, and do a 50/50 mix of AS3 and epoxy... It'll get you 270 core on that mutha... It'll definitely help... I have an AIW8500DV as well... nice card...

///BACK ON TOPIC///

Rogozhin

10-01-03, 11:59 PM

I came to the conclusion that my oc'd msi g4ti4400 wasn't any better than my oc'd radeon 8500LE.

just a difference of opinion.

rogo

ChrisRay

10-02-03, 12:15 AM

Originally posted by Rogozhin
I came to the conclusion that my oc'd msi g4ti4400 wasn't any better than my oc'd radeon 8500LE.

just a difference of opinion.

rogo

What? In Neverwinter Nights? Wouldnt surprise me as the two cards have always performed very similarly in this game. it was the r300 derivitive lines that had a buttload of trouble with it.

StealthHawk

10-02-03, 04:14 AM

For more Anand review "weirdness" see my post here (http://www.nvnews.net/vbulletin/showthread.php?s=&postid=205813#post205813).

Specifically, the Homeworld2 benchmark. Look at this stuff.

Benchmark1
no FSAA/AF
gfFX5600Ultra: 25.1
4x FSAA/8x AF
gfFX5600Ultra: 38.4 (+53%)

Benchmark2
no FSAA/AF
NV38: 43.8
gfFX5600Ultra: 15.5
4x FSAA/8x AF
NV38: 44.3 (+1%)
gfFX5600Ultra: 25 (+61%)

You'd think either the author or the editor would look at the graphs for errata like this and catch it.

ChrisW

10-02-03, 04:25 AM

Originally posted by StealthHawk
For more Anand review "weirdness" see my post here (http://www.nvnews.net/vbulletin/showthread.php?s=&postid=205813#post205813).

Specifically, the Homeworld2 benchmark. Look at this stuff.

Benchmark1
no FSAA/AF
gfFX5600Ultra: 25.1
4x FSAA/8x AF
gfFX5600Ultra: 38.4 (+53%)

Benchmark2
no FSAA/AF
NV38: 43.8
gfFX5600Ultra: 15.5
4x FSAA/8x AF
NV38: 44.3 (+1%)
gfFX5600Ultra: 25 (+61%)

You'd think either the author or the editor would look at the graphs for errata like this and catch it.
Good one! The new drivers are so good that the 5600U is actually a lot faster with fsaa and aniso enabled then when disabled!

druga runda

10-02-03, 06:29 AM

Originally posted by ChrisW
Good one! The new drivers are so good that the 5600U is actually a lot faster with fsaa and aniso enabled then when disabled!

:rofl

miracle drivers indeed...

that image quality is overplayed give me FPS :nana: ... faster with AA and AF on than off ... :afro:

theultimo

10-02-03, 07:04 AM

Huh :confused:

Doesnt make ANY sense....

Deathlike2

10-02-03, 12:35 PM

There's obviously something wrong.. I'd suspect the numbers are switched.. but I'm not sure myself..

In any case.. Anand really should be consulting more reliable sources (eg. Dave Baumann from Beyond3D) to look over some potential probs...

ChrisRay.. I'm sorry for coming on like that...

As far as I know, Bioware is currently working on resolving performance issues on ATI cards.. and a beta patch (to my knowledge) is available to improve performance...

The thing that the average person is NOT aware of hurts them... as normally we'd think that it's the vender's fault (in this case, ATI) for having poor performance "naturally" when it isn't the cause.

ATI_Dude

10-04-03, 04:10 PM

Originally posted by Deathlike2
While looking at this article... I've seen so much BS.. it's not remotely funny..

Note the use of Neverwinter Nights:

http://www.anandtech.com/video/showdoc.html?i=1890&p=13

Neverwinter Nights was designed with NVidia extensions in mind... it's a rather unbalanced comparison to say the least (although people could make arguments with the Glide/Direct3D/OpenGL benchmark days, that's not the scope of my discussion)

Obviously NVidia would win this...

According to Bioware NWN isn't particularly optimized for nVidia chipsets. The engine is dated and just happens to run better on nVidia hardware.

Originally posted by Deathlike2
Note the use of Splinter Cell:

http://www.anandtech.com/video/showdoc.html?i=1890&p=15

He fails to mention that AA is not useable in this benchmark.. yet he mentions it in his Halo benchmark...

He also fails to mention whether the nVidia cards use shadow buffer mode or projector shadow mode. Only nVidia chipsets (GeForce 3/4/FX) support shadow buffer mode which is much more demanding than projector shadow mode since it calculates dynamic shadows for multiple light sources. ATI cards don't support shadow buffers and run in projector mode. Unless projector mode was forced on nVidia cards in the benachmark the results are invalid since it gives ATI an unfair advantage.