Go Back   nV News Forums > Hardware Forums > Benchmarking And Overclocking

Newegg Daily Deals

Reply
 
Thread Tools
Old 09-16-03, 08:45 AM   #169
digitalwanderer
 
digitalwanderer's Avatar
 
Join Date: Jul 2002
Location: Highland, IN USA
Posts: 4,944
Default Re: Re: You're right, I was misunderstanding.

Quote:
Originally posted by Evan Lieb
My personal credibility or AT in general? Either way, you are in the minority and are being overly critical and sensational.
My apologies, I meant AnandTech's credibility...I really don't know you well enough to comment on yours but by your posts here I must say I'm getting a negative vibe already.

I'll agree that I might be in the minority on that currently, but just because I'm in the minority it don't mean the minority ain't right....as nVidia was kind enough to prove recently.
__________________
[SIZE=1][I]"It was very important to us that NVIDIA did not know exactly where to aim. As a result they seem to have over-engineered in some aspects creating a power-hungry monster which is going to be very expensive for them to manufacture. We have a beautifully balanced piece of hardware that beats them on pure performance, cost, scalability, future mobile relevance, etc. That's all because they didn't know what to aim at."
-R.Huddy[/I] [/SIZE]
digitalwanderer is offline   Reply With Quote
Old 09-16-03, 08:51 AM   #170
Clay
Registered User
 
Clay's Avatar
 
Join Date: Oct 2002
Posts: 4,993
Default Posting Elsewhere?

Evan,

I'm curious (and too busy to check around myself at the moment)...are you making similar posts at other sites as you have here? If not, why not?
Clay is offline   Reply With Quote
Old 09-16-03, 08:57 AM   #171
Solomon
 
Join Date: Sep 2002
Location: In a House
Posts: 502
Default

Quote:
Originally posted by Evan Lieb
Of course, but that doesnít mean theyíre right. Iím not here to argue the legitimacy or technical accuracy of any reviews. Quite frankly, if some other web site or person wants to constructively criticize an AT review, Iíll be all ears, but that doesnít mean Iíll necessarily agree with their assessment.
Since you brought up this, What about your explainations and reasons behind the Anandtech GeForce FX 5900 review? Ever since that review, Anandtech has slid down hill IMO. The Quake 3 numbers were so far fetched and no reason to why the charts were like that.

From I recall Anand never did anything to rectify it and now it seems the Q3A part of the review was just ripped out. So besides the smoke and mirrors. What was the reasoning behind the numbers on the GeForce FX 5900 review?

Regards,
D. Solomon Jr.
*********.com
Solomon is offline   Reply With Quote
Old 09-16-03, 08:59 AM   #172
The Baron
Guest
 
Posts: n/a
Default

Evan, the simple fact is that benchmarks have become the Only Real Test of a video card's quality anymore, and it is this attitude that has caused this atmosphere where companies believe that they can cheat in benchmarks and get away with it.

To be blunt--it's the fault of the sites where one guy reviews two cards a week. It's one thing to review two cards in a week that are based off the same chipset, use the same drivers, and have the same clockspeed, but reviewing a 5600 Ultra, then a 5900 Ultra, then a 9800 non-Pro, then a 9600 Pro, then a 5900 non-Ultra... the reviewer does not gain a level of familiarity with the cards that would ultimately benefit the consumer more than six hundred benchmarks.

Quote:
I am working on a solution, as I have alluded to in a few posts over the last couple of months. I am not ready to announce what I am doing, but the above is what I am attacking. The problems can be resolved.
Email me when it's ready, Skuzzy.
  Reply With Quote
Old 09-16-03, 09:00 AM   #173
digitalwanderer
 
digitalwanderer's Avatar
 
Join Date: Jul 2002
Location: Highland, IN USA
Posts: 4,944
Default

Quote:
Originally posted by Solomon
Since you brought up this, What about your explainations and reasons behind the Anandtech GeForce FX 5900 review? Ever since that review, Anandtech has slid down hill IMO. The Quake 3 numbers were so far fetched and no reason to why the charts were like that.

From I recall Anand never did anything to rectify it and now it seems the Q3A part of the review was just ripped out. So besides the smoke and mirrors. What was the reasoning behind the numbers on the GeForce FX 5900 review?

Regards,
D. Solomon Jr.
*********.com
Yeah, I remember bumping heads with some mods over at Anand about that one on their forums too....and the numbers stood without any explanation of their invalidity too.
__________________
[SIZE=1][I]"It was very important to us that NVIDIA did not know exactly where to aim. As a result they seem to have over-engineered in some aspects creating a power-hungry monster which is going to be very expensive for them to manufacture. We have a beautifully balanced piece of hardware that beats them on pure performance, cost, scalability, future mobile relevance, etc. That's all because they didn't know what to aim at."
-R.Huddy[/I] [/SIZE]
digitalwanderer is offline   Reply With Quote
Old 09-16-03, 09:11 AM   #174
Bert
a+d(L.N)+s(N.H)^n
 
Bert's Avatar
 
Join Date: Apr 2003
Location: Blue Circle
Posts: 69
Default

Quote:
Originally posted by Skuzzy
Doing tri-linear on the first texture stage only is not a bad methodology.
It is bad when I specifically enable it for another texture stage. I don't particularily care about the defaults, but the driver should not override settings the application does explicitly request.

I have an application that needs to control texture filtering exactly, and I was surely surprised when the Radeon (old one, it's been a while ago) did only bi-linear when I specifically requested tri-linear. Turned out the driver disabled it because of some other setting (aniso, IIRC).
Bert is offline   Reply With Quote
Old 09-16-03, 09:27 AM   #175
Skuzzy
Bit Bumper
 
Join Date: Aug 2002
Location: Here
Posts: 782
Default

I was playing a bit of devil's advocate Bert.

I whole heartedly agree and believe there should be a switch in the video card settings to enable/disable exactly what filtering methods should be used and the driver must not deviate from those settings.
Tri-linear, as DX defines its usage, is supposed to be always used in mip-map transitions, but MS added a slight caveat to this in another section of the specs. Almost looks like an afterthought, but it does state that if the hardware is not able to run tri-linear, then it is ok if the hardware/driver do not honor it, and I quote (from the DX9 docs):

"Because of limitations in much of today's hardware, few display adapters can perform trilinear mipmap interpolation through the multiple texture blending interface offered by IDirect3DDevice9. Your application can use multipass texture blending to achieve the same effects, or degrade to the D3DTEXF_POINT mipmap filter mode, which is widely supported."

ATI and NVidia are within thier rights to disable tri-linear in a multi-texture operation, according to the API. Enabling AF in the app, according to DX specs again, does not insure tri-linear will be used either.
__________________
Stuff Happenz!

Last edited by Skuzzy; 09-16-03 at 09:30 AM.
Skuzzy is offline   Reply With Quote
Old 09-16-03, 09:47 AM   #176
muzz
 
muzz's Avatar
 
Join Date: Feb 2003
Posts: 816
Default

Quote:
Originally posted by digitalwanderer
Yeah, I remember bumping heads with some mods over at Anand about that one on their forums too....and the numbers stood without any explanation of their invalidity too.
Yep there were issues with a review there a little while ago, jjbirney,myself and may others brought it to their attention, Evan said they were looking into it, and then NOTHING ( for like 2 weeks), next time I went back to the thread mysteriously disappeared... DELETED.
I started another thread about it, and of course that was ignored as well.
Don't question, just listen and believe what we say......

Ya right.
__________________
muzz
muzz is offline   Reply With Quote

Old 09-16-03, 10:02 AM   #177
Bert
a+d(L.N)+s(N.H)^n
 
Bert's Avatar
 
Join Date: Apr 2003
Location: Blue Circle
Posts: 69
Default

Quote:
Originally posted by Skuzzy

Tri-linear, as DX defines its usage, is supposed to be always used in mip-map transitions, but MS added a slight caveat to this in another section of the specs. Almost looks like an afterthought, but it does state that if the hardware is not able to run tri-linear, then it is ok if the hardware/driver do not honor it, and I quote (from the DX9 docs):

"Because of limitations in much of today's hardware, few display adapters can perform trilinear mipmap interpolation through the multiple texture blending interface offered by IDirect3DDevice9. Your application can use multipass texture blending to achieve the same effects, or degrade to the D3DTEXF_POINT mipmap filter mode, which is widely supported."

ATI and NVidia are within thier rights to disable tri-linear in a multi-texture operation, according to the API. Enabling AF in the app, according to DX specs again, does not insure tri-linear will be used either.
Wow, so this is stating correct filtering does only need to be applied to the first tex unit. If you want to have it on all units you have to multipass.

There is no such loophole in the OpenGL spec, though. The texture color is unmistakably specified as a lerp between mip levels adjacent to the LOD factor lambda. There is some allowance for how to determine lambda, though, to allow using euclidic or linear distance calculation. With some stretch this could be interpreted as allowing the pseudo-trilinear mode with only a small blending range, but falling back to point-sampling is definetly not.
Bert is offline   Reply With Quote
Old 09-16-03, 10:07 AM   #178
digitalwanderer
 
digitalwanderer's Avatar
 
Join Date: Jul 2002
Location: Highland, IN USA
Posts: 4,944
Default

Quote:
Originally posted by muzz
Don't question, just listen and believe what we say......

Ya right.
Yup. Where was their big concern about fair reviews/coverage back then? What changed between then and now to make them more interested in it?

Who will watch the watchmen?
__________________
[SIZE=1][I]"It was very important to us that NVIDIA did not know exactly where to aim. As a result they seem to have over-engineered in some aspects creating a power-hungry monster which is going to be very expensive for them to manufacture. We have a beautifully balanced piece of hardware that beats them on pure performance, cost, scalability, future mobile relevance, etc. That's all because they didn't know what to aim at."
-R.Huddy[/I] [/SIZE]
digitalwanderer is offline   Reply With Quote
Old 09-16-03, 10:13 AM   #179
The Baron
Guest
 
Posts: n/a
Default

Quote:
Originally posted by digitalwanderer
Who will watch the watchmen?
Enthusiasts. As soon as someone receives any piece of hardware for free solely for the purpose of review, they cease to be an enthusiast and become a journalist.

Why do you think I asked for some R3xx results with AF on my DM-Icetomb benchmark? I saw some numbers that didn't make sense (at least according to other sites), so I asked the community to do a sanity check. And I was right. That's why there is a community.
  Reply With Quote
Old 09-16-03, 10:27 AM   #180
digitalwanderer
 
digitalwanderer's Avatar
 
Join Date: Jul 2002
Location: Highland, IN USA
Posts: 4,944
Default

Quote:
Originally posted by The Baron
Enthusiasts. As soon as someone receives any piece of hardware for free solely for the purpose of review, they cease to be an enthusiast and become a journalist.
Sanity check: So I'm still safely an enthusiast until I publish a review?
__________________
[SIZE=1][I]"It was very important to us that NVIDIA did not know exactly where to aim. As a result they seem to have over-engineered in some aspects creating a power-hungry monster which is going to be very expensive for them to manufacture. We have a beautifully balanced piece of hardware that beats them on pure performance, cost, scalability, future mobile relevance, etc. That's all because they didn't know what to aim at."
-R.Huddy[/I] [/SIZE]
digitalwanderer is offline   Reply With Quote
Reply


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 01:28 AM.


Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright ©1998 - 2014, nV News.