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

Newegg Daily Deals

Reply
 
Thread Tools
Old 05-25-03, 09:18 PM   #73
Sazar
Sayonara !!!
 
Sazar's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 9,297
Default

Quote:
Originally posted by Behemoth
this is one way of shuffling instructions that manually reduces workload:

original program:

do a;
do b;
if (c == true) {do d;}
else { do e;}
do f;
do g;
do h;
do i;
do j;


by studying the scene, c is always true, instructions other than a,d,h have no visual effect in this scene, let me make a faster version of a mathematically functionally visually identical program of above by "shuffle instructions":

do a;
if (c == true) {do d;}
else {do e;
do b;
do f;
do g;
do i;
do j;
}
do h;



original program does 8 instructions and 1 comparison per pass.
my mathematically functionally visually identical program does only 3 instructions and 1 comparison per pass. what an optimization!

shuffle instructions can be very cheaty imho

p.s. if my post offend you in any way, i am sorry, i am just here to share my point of view
how come you never came up with such detailed work dealing with nvidia's 'optimizations' as you call it ?
Sazar is offline   Reply With Quote
Old 05-25-03, 09:26 PM   #74
Spotch
 
Join Date: Dec 2002
Location: Dallas,TX
Posts: 352
Default

Quote:
Originally posted by Sazar
nice response sean...

here's my thing with kyle though...

fair enough if he is not aware of the situation.. he can reserve judgement on nvidia's side of matters...

but... the speed with which he attacked (and I would consider it an attack purely based on the terminology used) ET and the fact that he practically dismissed the document put forth by futuremark is not really a sign of an un-biased 'journalist'

seeing the way he reported the quake issue even with the evidence @ hand and the handling of this matter... it is very apparent there appear to be different standards... whether it is because of bias to a particular IHV or whether it has anything to do with his rather apparent dislike for 3dmarko03 (and therefore perhaps futuremark ? ) I can't say...

but the fact of the matter is we have evidence of something going on... which when disabled renders a product in a not so attractive light... and instead of making ANY comments on this... kyle is making rather thinly veiled accusations at the people reporting this matter rather than discussing anything related to nvidia's side of things... ie how these optimizations affect us...

whether kyle likes 3dmark03 or not is NOT in question... nvidia's use of 'optimizations' is... and yes... while kyle does have a distinct style of reporting... it is disturbing to see the way he has handled this matter...

fair enough... if more info is required... so be it... kyle could have taken the high road and done like anandtech and toms and NOT implicated himself in this matter in the manner he has...

comments made about kyle are based on HIS actions thus far... and are not based on rumours or innuendo... they are based on cross-referencing what we.. the consumers... are seeing on various sites/the futuremark .pdf/independent testers

I am willing to wait to see what happens... I have no problems with that... I just wish kyle had decided to do the same...

ergo... I will 'trust' kyle perhaps when I see what he has to say on the full matter and to see if he manages to do this without throwing out petty and utterly useless accusations concerning other sites and benchmark devs...

-edit- btw nvidia has taken the same route as kyle in attacking futuremark and implying ati's connection rather than taking anytime whatsoever to address the fact that their drivers have been shown to be acting improperly in a benchmark...
NVIDIA has taken the same route as Kyle? Your joking right? Kyle has NV's cork up his sellout arse and thats the fact jack. I wouldn't be surprised if they put him up to his original Futuremark attack with promises of gold and jewels. (And more special Doom 3 like benchmarks) He is in this for the money... he is no longer a journalist, he is a billboard with NV Biach stamped on his forehead! But thats just my opinion, yours may differ.

Spotch is offline   Reply With Quote
Old 05-25-03, 10:19 PM   #75
Behemoth
radeon 9800 pro
 
Behemoth's Avatar
 
Join Date: Apr 2003
Location: Darkness Falls
Posts: 841
Default

Quote:
Originally posted by StealthHawk
No, you said we should wait for official word from nvidia before condemning them. We have official word from one company on what they did, and it's not nvidia. And ATI never said they were cheating. CatalystMaker said it was an OPTIMIZATION, and he was against removing it because he felt it was valid.
ok you didnt even believe what i said, here is the quote:
Quote:
possibly bug.
possibly nvidia cheat.
possibly futuremark cheat.
its on page 8 near the bottom in a thread called my thoughts on "optimization" in this forum. i would wait to see it proved, i didnt rule out the nvidia cheating possibllity as welll.
Ati never said they were cheating, but nvidia was even better, nvidia said nothing!!! but ati admitted they "shuffle instructions" which alone looks cheaty enough to me already. at the time i said i would wait to see more nvidia evidences, i didnt know there was official word from any company on what nvidia did.

Quote:

What guess? Did you even read the .pdf audit from Futuremark? They said that they DETECTED nvidia's shader programs, and they tell you where each shader program was found.
the guess refered to ET guessed clip planes were manually inserted.


Quote:

What Futuremark evidence against ATI? There is NONE. No claims, no screenshots, NOTHING. They said ATI lost 8% in GT4 and they would investigate further. They did not accuse ATI of doing anything. ATI are the ones who said they changed a shader.
ok if you dont like them to be called evidences, lets call them whatever you like, however i believe what futuremark said about ati on 2 things.
1. "the test was also detected"
2. "and somehow altered by the drivers"
both on page 4 of futuremark audit report, actually they are one sentence but they tell me 2 things.
these are what i would consider as much better evidences than a few screenshots plus a guess, whether they should be called evidences or not that is not important to me, i only know they are much better than a few screenshots plus a guess.
and again, ati even admitted they "shuffle instructions", this plus futuremark's 2 points, maybe its just me, are much more different than a few screenshots and a guess.


Quote:

How would they prove it's a valid optimization? By showing you their code compared to the original code? Is that even legal? Can they show you Futuremark's original code legally?
i dont know how would they prove it, but if they incorporate the so called questionable but valid optimization properly in their next driver, and the performance wont drop 8% without doing driver detection and codes alteration, its already enough a proof to me.

Quote:

So you don't want to believe the guess of ET, B3D, or Futuremark, but we're supposed to believe your guess....show me the logic. Your guess is based on what observations or evidence?
i believe ET,B3D,Futuremark to some extent.
i have already showed you my logic, observations and evidences, believe what you will, i am just sharing my point of view.

Last edited by Behemoth; 05-25-03 at 10:37 PM.
Behemoth is offline   Reply With Quote
Old 05-25-03, 10:34 PM   #76
Behemoth
radeon 9800 pro
 
Behemoth's Avatar
 
Join Date: Apr 2003
Location: Darkness Falls
Posts: 841
Default

Quote:
Originally posted by Sazar
how come you never came up with such detailed work dealing with nvidia's 'optimizations' as you call it ?
because i have realised my understanding of "shuffle instruction" is so different to many others in here, its so interesting i would like to share with you all guys
if nvidia brought up such terms, i would probably do the same depending on how people understand it, and of coz people are not gonna believe nvidia if they say "shuffle instructions"
Behemoth is offline   Reply With Quote
Old 05-25-03, 11:11 PM   #77
Sazar
Sayonara !!!
 
Sazar's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 9,297
Default

Quote:
Originally posted by Behemoth
because i have realised my understanding of "shuffle instruction" is so different to many others in here, its so interesting i would like to share with you all guys
if nvidia brought up such terms, i would probably do the same depending on how people understand it, and of coz people are not gonna believe nvidia if they say "shuffle instructions"
nvidia did not shuffle instructions though

btw... how is it possible for futuremark to cheat ?

it is THEIR benchmark... THEY are not an IHV that is running hardware using the software... I have no idea what you mean by futuremark cheat...

further.. ET did not GUESS... they had enough evidence to suggest it was put in place on purpose... ie... code chcked applciation being run and ran optimizations based on that info...

couple of other things..

you read through the futuremark report... they have provided the FILES that nvidia's drivers changed and replaced in order to run the benchmark...

there have been pictures posted on various websites incl this one showing the difference in precision as well in gt4 in the water.. and we have already seen teh sky

concerning the optimizations... read up in these forums about what uttar has to say as well as checking the frontpage of www.rage3d.com for sweeny's explanation of shaders and optimizations...

perhaps that will shed some light

enjoy..

Last edited by Sazar; 05-26-03 at 12:04 AM.
Sazar is offline   Reply With Quote
Old 05-25-03, 11:14 PM   #78
Sazar
Sayonara !!!
 
Sazar's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 9,297
Default

Quote:
Originally posted by Spotch
NVIDIA has taken the same route as Kyle? Your joking right? Kyle has NV's cork up his sellout arse and thats the fact jack. I wouldn't be surprised if they put him up to his original Futuremark attack with promises of gold and jewels. (And more special Doom 3 like benchmarks) He is in this for the money... he is no longer a journalist, he is a billboard with NV Biach stamped on his forehead! But thats just my opinion, yours may differ.

since kyle came out with the first jabs @ ET and futuremark... it is only fair to say nvidia is taking the same route as kyle don't you think ?

after all nvidia's first word on this has been this week... a few days ago... kyle has been discussing ET's ethics and futuremark's test and its validity far before nvidia decided to release a statement...

ergo.. I am not joking... I am posting with consideration to the precedence established by kyle in this matter considerng 'optimizations and cheats'
Sazar is offline   Reply With Quote
Old 05-25-03, 11:34 PM   #79
Spotch
 
Join Date: Dec 2002
Location: Dallas,TX
Posts: 352
Default

Quote:
Originally posted by Sazar
since kyle came out with the first jabs @ ET and futuremark... it is only fair to say nvidia is taking the same route as kyle don't you think ?

after all nvidia's first word on this has been this week... a few days ago... kyle has been discussing ET's ethics and futuremark's test and its validity far before nvidia decided to release a statement...

ergo.. I am not joking... I am posting with consideration to the precedence established by kyle in this matter considerng 'optimizations and cheats'
There is a difference between precedence and coincidence. Even assuming Kyle's bed buddy at NVIDIA didn't talk him into posting the reaction on [H], it is still incorrect to say that a multi-billion dollar corporation took the same road as some hack on a website. Besides, I am certain Kyle's mind was made up for him by NVIDIA and he, being an uncouth young lad, recklessly blurted it out like a kid who couldn't wait to tell the whole school about his first lay.
Spotch is offline   Reply With Quote
Old 05-25-03, 11:48 PM   #80
Behemoth
radeon 9800 pro
 
Behemoth's Avatar
 
Join Date: Apr 2003
Location: Darkness Falls
Posts: 841
Default

Quote:
Originally posted by Sazar
nvidia did not shuddle instructions though

btw... how is it possible for futuremark to cheat ?

it is THEIR benchmark... THEY are not an IHV that is running hardware using the software... I have no idea what you mean by futuremark cheat...

further.. ET did not GUESS... they had enough evidence to suggest it was put in place on purpose... ie... code chcked applciation being run and ran optimizations based on that info...

couple of other things..

you read through the futuremark report... they have provided the FILES that nvidia's drivers changed and replaced in order to run the benchmark...

there have been pictures posted on various websites incl this one showing the difference in precision as well in gt4 in the water.. and we have already seen teh sky

concerning the optimizations... read up in these forums about what uttar has to say as well as checking the frontpage of www.rage3d.com for sweeny's explanation of shaders and optimizations...

perhaps that will shed some light

enjoy..
i believe futuremark is possible to cheat. but why would futuremark want to cheat? please forgive me i dont want to go over it once again.
thanks for your kind info on the couple of other things. i am sure they may change my view on nvidia.

Last edited by Behemoth; 05-25-03 at 11:57 PM.
Behemoth is offline   Reply With Quote

Old 05-26-03, 12:04 AM   #81
Sazar
Sayonara !!!
 
Sazar's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 9,297
Default

Quote:
Originally posted by Spotch
There is a difference between precedence and coincidence. Even assuming Kyle's bed buddy at NVIDIA didn't talk him into posting the reaction on [H], it is still incorrect to say that a multi-billion dollar corporation took the same road as some hack on a website. Besides, I am certain Kyle's mind was made up for him by NVIDIA and he, being an uncouth young lad, recklessly blurted it out like a kid who couldn't wait to tell the whole school about his first lay.
it is possible.. no doubt...

my point was not to make out that nvidia and hardocp are bed buddies...

that is your assumption... and I will leave it at that...

my point in the initial post followed from the rest of what I posted...

implication was that nvidia has done like kyle in not addressing the issue @ hand but rather pointing out other things that should have no bearing on the matter @ hand...
Sazar is offline   Reply With Quote
Old 05-26-03, 12:05 AM   #82
Sazar
Sayonara !!!
 
Sazar's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 9,297
Default

Quote:
Originally posted by Behemoth
i believe futuremark is possible to cheat. but why would futuremark want to cheat? please forgive me i dont want to go over it once again.
thanks for your kind info on the couple of other things. i am sure they may change my view on nvidia.
yes quite... I still don't have a clue what you mean by futuremark cheating...

they are not testing themselves so what is the cheating you are talking about ?
Sazar is offline   Reply With Quote
Old 05-26-03, 12:22 AM   #83
Behemoth
radeon 9800 pro
 
Behemoth's Avatar
 
Join Date: Apr 2003
Location: Darkness Falls
Posts: 841
Default

Quote:
Originally posted by Sazar
yes quite... I still don't have a clue what you mean by futuremark cheating...

they are not testing themselves so what is the cheating you are talking about ?
when things are not proved, i usually dont rule out any possibilty, if i think futuremark is capable of cheating, why not put it on suspect list as well
Behemoth is offline   Reply With Quote
Old 05-26-03, 12:25 AM   #84
Sazar
Sayonara !!!
 
Sazar's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 9,297
Default

Quote:
Originally posted by Behemoth
when things are not proved, i usually dont rule out any possibilty, if i think futuremark is capable of cheating, why not put it on suspect list as well
yes.. but in order to put in on the suspect list you need to have some evidence...

I STILL can't understand the possible logic you are using to suggest that futuremark is cheating... what are they cheating in and how can they possibly achieve this ?

Sazar 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


Similar Threads
Thread Thread Starter Forum Replies Last Post
"GPU has fallen off the bus" error on 650M unless a CUDA program run first amonakov NVIDIA Linux 0 06-18-12 06:34 PM
Getting the proprietary nvidia driver to run with Debian 3.0 r0 (woody) Katchina404 NVIDIA Linux 9 01-12-03 08:49 AM
On The 333 Boards Does Anyone Run At 166fsb john19055 CPUs, Motherboards And Memory 22 07-30-02 07:39 PM

All times are GMT -5. The time now is 12:12 AM.


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