[pcper] Interview: AMD's Richard Huddy

Page 6 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Erenhardt

Diamond Member
Dec 1, 2012
3,251
105
101
Having SDK and getting a game build are two different things. NV got Tomb Raider build before the game was released. There were last minute changes to the game, which nv couldn't optimize for, while amd worked on those changes in the phase of development. The head start for amd was what made a difference.
Anyway, game was patched later to run better on nv hardware which is quite the opposite to what happens with twimtpb titles - amd proposed changes are rejected due to nv partnering policy.

Shin, repeat Huddy is a spindoctor a couple times more and someone could believe you.
 

ShintaiDK

Lifer
Apr 22, 2012
20,378
145
106
It sounds to me that the AMD guy is making claims that can be verified as true or false.

And it seems he is making false claims.

Why would that be considered a good strategy? I can't figure it out. Seems to me the only benefit would be to whip up controversy and attention, perhaps angering Nvidia fans and pandering to AMD fans.

But if the idea is to sell more video cards, I think it's a lot more helpful to stick to the truth? Apparently this guy is just spouting off about speculations and assumptions, trying to get emotional reactions? Sorry if I'm mischaracterizing, but I don't think this strategy is a good idea or viable long-term.

He did it atleast since 2005 when working for ATI. Back then there was controversy with 3DMark05. Because allegedly Futuremark favoured nVidia by using their DST/PCF extensions. Richard Huddy kept claiming in public that 3DMark05 would benefit from ATIs 3Dc extensions. And thats even after that ATI and Richard already knew that 3Dc couldnt be used.

ATi devrel had actually worked with Futuremark, and after analysis of the algorithm used for bumpmapping, it was mutually decided that 3Dc wouldn't work. The reason being that Futuremark used unnormalized normals, where the amount of 'unnormalization' would act as a scaling factor which would reduce aliasing. 3Dc can only handle normalized normals, so this trick wouldn't work, and image quality would be compromised.
Spindoctors ever think far ahead.

However, I would like to say this applies to all companies (and politicians). They are all full of BS. Its just the amount that may differ a little bit.

But the use of spindoctors obviously work on the majority. Else they wouldnt have armies with these at crazy pay rates. And this forum shows no different. People jump in every time with both legs.
 

caswow

Senior member
Sep 18, 2013
525
136
116
Wow some people are crazily obsessed with this guy and tomb raider . huddy said himself "dig into this stuff dont trust me and dont trust nvidia" and "and if you see us doing something similar go after us!". he is saying himself magazines and online sites have a responsibility to their readers.

you guys are taking this interviews out of context and twist his words around until you like it the way you want to have it.
 

sontin

Diamond Member
Sep 12, 2011
3,273
149
106
Having SDK and getting a game build are two different things.

Nope, there is no difference when you cant change the code.

Wow some people are crazily obsessed with this guy and tomb raider . huddy said himself "dig into this stuff dont trust me and dont trust nvidia" and "and if you see us doing something similar go after us!". he is saying himself magazines and online sites have a responsibility to their readers.

you guys are taking this interviews out of context and twist his words around until you like it the way you want to have it.

Right, they have the responsibility to call AMD out for Tomb Raider. But nobody did it.
And like i said: Huddy said reviewers shouldn't use Gaming Evolved titles in their benchmark parcours because AMD is paying the publisher to implement their features, prevent others from having access to the source or even builds and they prefer games with an inaccurate ingame benchmark.
 
Last edited:

BrightCandle

Diamond Member
Mar 15, 2007
4,762
0
76
We know he lied about TessFX availability to Nvidia.

We also know he is lying about gsync and one extra frame of latency.

We know he isn't telling the whole truth about contracts regards Tomb Raider as Nvidia didn't have access to the source code yet in the MaximumPC he claims they don't put in contract provisions to restrict developers from providing code to Nvidia.

We know that the crysis 2 tessellated water pixture isn't actually true, those texels were culled in the rendered process but that culling is turned off in wireframe - so using that as an example of Nvidia overtesselation is a lie. Batmans cape however is possibly a valid claim.

I would say based on the lies he has been caught in across 2 interviews suggest that the man is inherently dishonest and we should treat any new information he gives us with suspicion and it needs verification from other sources. He is not an honest source of information, that is unfortunately the sad reality that in a couple of days of interview he has tarnished his position and we can't trust what he says, based on things he chose to say that were not true and he knows aren't true.
 

sontin

Diamond Member
Sep 12, 2011
3,273
149
106
Was is the problem with the cape?
Batman:AO runs on a GTX650TI with 60FPS in the ingame benchmark without MSAA.

On the other hand: Dirt:Showdown runs with less than 60fps on a GTX680 with Forward+.
 

BrightCandle

Diamond Member
Mar 15, 2007
4,762
0
76
Was is the problem with the cape?
Batman:AO runs on a GTX650TI with 60FPS in the ingame benchmark without MSAA.

On the other hand: Dirt:Showdown runs with less than 60fps on a GTX680 with Forward+.

It impacts AMD's card because they have lower tessellation performance. The claim is that the subpixel level of tessellation on Batmans cape was not in the developers control, was hidden in gameworks and while it hurts Nvidia cards a little it hurts AMD hardware a lot (40% drop in frame rate). Thus Nvidia is using it to hurt them.

My understanding of how tessellation works is that the model itself provides details of how it is tessellated. The model is definitely within the hands of the developers as they have to choose the tessellation map. But it is a fixed function part where you tell it to what extent to apply the algorithms and quite possibly the quality/extent of tessellation might be gameworks and not in the developers code, I don't know since I haven't seen the gameworks API and since its not open I don't think I can see it. I am not sure that what Huddy has said is actually the case but its plausible, it would require the developers to confirm or not what was and wasn't in their control to be sure about it.
 

sontin

Diamond Member
Sep 12, 2011
3,273
149
106
It impacts AMD's card because they have lower tessellation performance. The claim is that the subpixel level of tessellation on Batmans cape was not in the developers control, was hidden in gameworks and while it hurts Nvidia cards a little it hurts AMD hardware a lot (40% drop in frame rate). Thus Nvidia is using it to hurt them.

He really said that? Wow.
WB Montreal worked with nVidia to improve the cape with Tessellation. There was nothing behind a closed wall. The cape is an asset of WB Montreal.

My understanding of how tessellation works is that the model itself provides details of how it is tessellated. The model is definitely within the hands of the developers as they have to choose the tessellation map. But it is a fixed function part where you tell it to what extent to apply the algorithms and quite possibly the quality/extent of tessellation might be gameworks and not in the developers code, I don't know since I haven't seen the gameworks API and since its not open I don't think I can see it. I am not sure that what Huddy has said is actually the case but its plausible, it would require the developers to confirm or not what was and wasn't in their control to be sure about it.

Nope. The cape is not a "gameworks" feature. It is property of WB Montreal. Everything of it is visible to everyone in the source code of the game.

AMD is complaining that WB Montreal doesn't care about their slow Tessellation performance. However AMD didn't care about Kepler with Dirt:Showdown either. So, what is Huddy's problem again?
 

SPBHM

Diamond Member
Sep 12, 2012
5,059
413
126
I think it's an interesting point, if something like the cape is overall very inefficient, does not add significantly to image quality, but is there because they know it's going to impact more the AMD hardware, it's quite a bad thing to do...

it can get really ridiculous like the water on crysis 2, the concern about what the gameworks dlls are doing seems reasonable.


but from all the things he talked, Mantle on other hardware was the most interesting thing.

as for gsync latency, I remember even nvidia avoiding to claim it had the same latency as no vsync, it's worth investigating more... but freesync is to unknown quantity at this point...
 

BrightCandle

Diamond Member
Mar 15, 2007
4,762
0
76
He really said that? Wow.
WB Montreal worked with nVidia to improve the cape with Tessellation. There was nothing behind a closed wall. The cape is an asset of WB Montreal.



Nope. The cape is not a "gameworks" feature. It is property of WB Montreal. Everything of it is visible to everyone in the source code of the game.

AMD is complaining that WB Montreal doesn't care about their slow Tessellation performance. However AMD didn't care about Kepler with Dirt:Showdown either. So, what is Huddy's problem again?

He is claiming its using Gameworks. It is using PhysX so it must be using gameworks functionality, but I am kind of in agreement with you that my understanding of the tessellation is that its in the developers hands. But I also can't claim to know exactly what the API calls are and whether the tessellation call to DX is originating from the game code or gameworks. It would seem odd that it came from gameworks because PhysX is about movement not increasing detail but its possible that call is happening from PhysX. Not very plausible however, but its another one of his claims.

PhysX is definitely not source code anyone but Nvidia gets to see, so calls made from it to DX would indeed be hidden from all eyes.
 

sontin

Diamond Member
Sep 12, 2011
3,273
149
106
He is claiming its using Gameworks. It is using PhysX so it must be using gameworks functionality, but I am kind of in agreement with you that my understanding of the tessellation is that its in the developers hands. But I also can't claim to know exactly what the API calls are and whether the tessellation call to DX is originating from the game code or gameworks. It would seem odd that it came from gameworks because PhysX is about movement not increasing detail but its possible that call is happening from PhysX. Not very plausible however, but its another one of his claims.

The cape is a normal game asset. WB Montreal is using standard DX11 Tessellation to make it look better.
API calls are coming straight from the game code.

He is just complaining about it because AMD's hardware sucks with geometry performance.

And with Hairworks it is the same: nVidia's technique is using Tessellation to produce around 500k hairs in CoD:Ghost. And they can do something like this with Hairworks: https://www.youtube.com/watch?v=X7RQ4SKBGKk
 
Last edited:

3DVagabond

Lifer
Aug 10, 2009
11,951
204
106
Exactly. Spindoctors are per definition untrustworthy.

But they hold the power of rhetorics to bend the weak minded into their will.

Its quite simple. Always look with a critical view on all information. Just like any historican would with their sources.

Read Blackened's post. He at least spent some of his post disputing a statement made, "TressFX". He then quoted a source that at least has dealt with Huddy. You, once again, are merely name calling with zero to back up what you say. All we can do is examine your credibility in regards to AMD and you are definitely nVidia biased, so where does that leave anyone to decide on your statements.

@Blackened, Off to work, I will respond tp you, but NZ is about as far off timezone wise as is possible with most of the Western world.
 

Mand

Senior member
Jan 13, 2014
664
0
0
All we can do is examine your credibility in regards to AMD and you are definitely nVidia biased, so where does that leave anyone to decide on your statements.

Actually, no, you do not get to attack other members in order to attempt to discredit their statements.
 

blackened23

Diamond Member
Jul 26, 2011
8,548
2
0
He is claiming its using Gameworks. It is using PhysX so it must be using gameworks functionality, but I am kind of in agreement with you that my understanding of the tessellation is that its in the developers hands. But I also can't claim to know exactly what the API calls are and whether the tessellation call to DX is originating from the game code or gameworks. It would seem odd that it came from gameworks because PhysX is about movement not increasing detail but its possible that call is happening from PhysX. Not very plausible however, but its another one of his claims.

This entire tessellation thing is ridiculous. Nvidia and AMD create features, they do not create games. They are not responsible nor have any control of the amount of tessellation used. The unfortunate truth is that NV has had better tessellation performance than AMD for some time now. This was VERY pronounced during the Cayman era, because frankly, the Cayman absolutely sucked at tessellation. Tahiti improved it markedly as did Hawaii, but it is still not as good as the hardware tess used by NV. Aside from this: tessellation is one of the best DX11 features because it greatly increases the realism and detail of any given polygon.

So this Huddy clown is saying that, essentially, that developers should take pity on them and use less tessellation to appease their hardware. Really? REALLY? Is this guy joking or what? And then he somehow tries to make it sounds like this is Nvidia's fault. Like nvidia programs games. Yeah, let's just pretend that Nvidia programmed Watch Dogs. I guess NV programmed Watch Dogs from the ground up if we all are to believe AMD's marketing.

AMD also already addressed this issue by having per game tessellation control, with user control in their CCC application. AMD's tessellation may be slower by default, but with performance drivers they can raise it or the user can use AMD's optimized tessellation which lowers quality but improves performance. This was used very often with Cayman, although Tahiti and Hawaii are not nearly as bad as Cayman was at tessellation. Cayman was a dog at tessellation, period. And someone remind me what happened when Crysis 2 DX11 was released? Did AMD's marketing whine about Crysis 2 using too much tess? Sound familiar?

As far as Batman: AO goes, to my knowledge, AMD already released performance drivers which brings Hawaii to near parity with NV or reasonably close. But here we are again at the heart of AMD's problem. Too much money wasted on marketing and the payroll of guys like Hallock and Huddy, and not enough money where it matters: software engineering. So they take forever to release a performance driver for Batman: AO which does bring AMD hardware to near parity with NV. So while they were taking forever to release that performance driver, they play the blame game instead of doing what they SHOULD be doing: getting stuff done on the software side of things. When is AMD going to get a clue and put the money in the proper areas? Not marketing, not the blame game, but software development. That's AMD's problem. Their hardware is fantastic. Software needs more funding to make that hardware realize its potential. Marketing? Their marketing is detestable.

Any and all tessellation, and the amount used, is entirely at the discretion of the developer. Nvidia does not go in and program games, they did not program the amount of tessellation used, that is entirely up to the developer. So again this Huddy (and his sidekick Hallock) are trying to create some sort of smoke screen where they want people to think that NV are actually programming the games themselves. That is completely eyeroll worthy. The fact that Huddy even brought this up exposes him for what a schmuck he really is.
 
Last edited:

monstercameron

Diamond Member
Feb 12, 2013
3,818
1
0
It is interesting that this huddy guy doesnt have any credibility the moment he rejoins AMD, or so some posters would have us believe.
well another poster added to the ol' ignore list.
 

AnandThenMan

Diamond Member
Nov 11, 2004
3,980
595
126
This entire tessellation thing is ridiculous. Nvidia and AMD create features, they do not create games. They are not responsible nor have any control of the amount of tessellation used.
It is VERY naive to think that Nvidia doesn't have any political influence on what their game dev partners do.
So this Huddy clown...
Stopped reading your post there.
 

f1sherman

Platinum Member
Apr 5, 2011
2,243
1
0
We all know who's who on this forum.
There is no need to call bias; 50% of this subforum is about bias and IHV preference.

The fact is AMD is in yet another PR offensive. I guess that's not a bad move because it brings them in spotlight.
But I'd rather see them in tech offensive.


$@*& can't believe they are still on "over tessellation".

HOW ABOUT MAKING YOUR GPUS DO TESSELLATION BETTER , instead of crying omg too many details, not optimal visual gains vs performance hit
 
Last edited:

blackened23

Diamond Member
Jul 26, 2011
8,548
2
0
The proof is in the AMD performance drivers for Batman: AO. This is a recurring theme. AMD has released a performance driver that brings AMD to near parity with NV in terms of Batman: AO performance. But the issue is that AMD spends millions on marketing and not nearly enough on what matters: software development.

So this creates a situation to where AMD's software guys don't' have the resources to do their job, the software guys can't release a performance driver for new games in a reasonable amount of time, and that compels their marketing guys to go to the media and play the blame game.

AMD's performance in Batman: AO is now fine after their driver addressed it. The problem is it took AMD forever and a day to release a performance driver for it. Too much money on marketing, too much money on social media, and not enough money on software development.

Here's hoping AMD gets rid of their marketing staff, drops their social media marketing nonsense, and gets some more software development staff. Like I said, their software guys are the REAL jobs at AMD that are important jobs. Do you seem them whining? No, they just do their thing but they apparently don't have the manpower and money to do it properly. That money is apparently going to marketing. If AMD fixes that, they won't have to whine/lie to the media as they do. And this tessellation thing is truly hilarious. The fact that Huddy suggested that developers go easy on them because THEY CANT CREATE PERFORMANCE DRIVERS in a reasonable amount of time and aren't (apparently) willing to dedicate more die size to hardware tessellation, is truly hilarious. What a class act this Huddy guy is. Sarcasm intended.

What's stupid about this is that AMD's graphics hardware is great, always has been. Not always without fault (R9-290 reference cooling as an example) but performance wise, has been great and competitive. AMD has always been competitive with NV even without not always having the performance crown. But I find the PR spin as of late to be fairly ridiculous. It's about time they dedicate more resources to the software, the real issue, instead of the PR spin game.
 
Last edited:

f1sherman

Platinum Member
Apr 5, 2011
2,243
1
0
Did Nvidia cry omg too many compute, too much SSAA, too much Global Lightning for no visual gains at all?
 

PPB

Golden Member
Jul 5, 2013
1,118
168
106
Guess that "clown" is the reason you are able to play games in yer almighty DX. Some people are just being plain disrespectful, just because they went to the company that isnt paying them to post in this forum, lol.

Also, you have to be totally ignorant of 3d modeling to think that straight concrete wall-rails need that much tesselation. Tip: they really dont.
 

blackened23

Diamond Member
Jul 26, 2011
8,548
2
0
Did Nvidia cry omg too many compute, too much SSAA, too much Global Lightning for no visual gains at all?

The forward+ lighting used in Codemasters games, brought up earlier, is a good point. Did NV whine about that? No, no I don't think so. Even though that feature is completely broken and cripples performance on NV hardware. It took time but later performance drivers brought forward+ performance up to a reasonable level on NV hardware. Without the whining.

I just respect companies that fix their crap without whining, it's more respectable than a company that relies on the PR spin game.

This strategy also apparently isn't working for AMD. You'd think for all this PR nonsense as of late, that they would have results to show for it. But they're still stuck at 35% discrete GPU market share.
 
Last edited:

Grooveriding

Diamond Member
Dec 25, 2008
9,110
1,260
126
Given how upset Mr. Huddy has made particular individuals, I think he's on the right track as far as his role with AMD.
 

f1sherman

Platinum Member
Apr 5, 2011
2,243
1
0
The forward+ lighting used in Codemasters games, brought up earlier, is a good point. Did NV whine about that? No, no I don't think so. Even though that feature is completely broken and cripples performance on NV hardware. It took time but later performance drivers brought forward+ performance up to a reasonable level on NV hardware. Without the whining.

I just respect companies that fix their crap without whining, it's more respectable than a company that relies on the PR spin game.

This strategy also apparently isn't working for AMD. You'd think for all this PR nonsense as of late, that they would have results to show for it. But they're still stuck at 35% discrete GPU market share.


Even 50% would not change much on $$ front. Not when they are going for throat with NV for existing customers (which barely keeps the two of them alive), instead of widening total customer base.
Or while they're getting pummeled in mobile and pro which are almost an order of magnitude more profitable. But hey that's OK - look over here Desktop, we'll go full force there.

And they do this all the time, the lesser margin market there is, the more aggressive and willing to fight they are.
I'm surprised they're not in sand digging by now.
 

SirPauly

Diamond Member
Apr 28, 2009
5,187
1
0
Personally allow developers to decide what is best for their applications and for their customers!
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |