NVIDIA Editor's Day Report November 02, 2003 Jakub Wojnarowicz

Overview

Page:: ( 1 / 9 )

There’s a lot you may have heard about NVIDIA Editor’s Day. No doubt there are plenty of reports floating about the conference and what happened. You’ve certainly read reports of game developers showing up to support NVIDIA, and others simply taking the time to market their games. There has no doubt been coverage, honest and disingenuous, favorable and otherwise, of the statements made by NVIDIA’s engineers and reps, as well as their esteemed CEO, Jen-Hsun Huang.

Now, when it comes to finding out what it all means, there has been little discussion. Well, let’s disregard the comments from leading members of the peanut gallery who weren’t even at the event and are simply maintaining their image of a fair and balanced neutrality by kissing the other side’s ass this month. Picking and choosing quotes from a third party and shaping them to suit your own purposes hardly qualifies as professionalism.

[image]

But first, a prelude

Firstly, and this has nothing to do with the conference, it’s time to clear the air of a little popular myth. NVIDIA isn’t 3dfx. NVIDIA’s development as a company doesn’t mirror that of 3dfx. To dispel notions of NVIDIA’s impending doom, the “3dfx curse” and other such nonsense, all one needs to do is look at the financial results for NVIDIA and ATI. Both experienced gains in the last quarter, and while NVIDIA is down relative to its peak, they still maintain a lead that they aren’t keen on relinquishing.

There are some interesting insights to be gleaned from the NVIDIA notables and their attitudes. To further contrast NVIDIA and 3dfx: while 3dfx representatives had an air of hopeful neediness around E3 2000 when they launched the Voodoo5, and were downright desperate a year later when demonstrating the 6000, NVIDIA employees are quite confident and calm.

However, there’s no denying that NVIDIA took the Editor’s Day event very seriously. To suggest that this is a careless response to ATI’s Shader Day and that they disregard their competitors would be a false conclusion. Derek Perez himself expressed surprise that NVIDIA’s CEO, Jen-Hsun Huang, stuck around all day to watch and participate in the proceedings. Mr. Huang, despite his considerable skills in the matter, isn’t a marketing or PR man. As a CEO, he has important duties and his continued presence at the Editor’s Day gives us a hint of the event’s gravity.

The candid nature of the event and the admission that they have been on the defensive for a long time is further proof, and we’ll get to that shortly. Indeed, some may take a more critical view of this and state the opinion that the impressive array of NVIDIA notables was there to because of the seriousness of the matter.



SIDEBAR: I had the immense pleasure of meeting two Japanese writers, who gave me some very interesting perspectives on PC gaming in Japan. My thanks go out to Tsuyoshi Yoshida and Kenji Yoshida (no relation).

Jen-Hsun’s Commentary

Page:: ( 2 / 9 )

A critical look

At lunch, we humble members of the press enjoyed “a fireside chat” with the disarmingly matter-of-fact Jen-Hsun, and Todd Hollenshead, CEO of id. To get a full appreciation of the effect this had, it helps to remember that id has an untarnished reputation in the industry. Nobody is about to accuse John Carmack of fudging numbers or bias when it comes to computer hardware, while id’s financial status and fierce independence makes it rather unlikely they were bought a spot at the conference or compelled through contract to appear. A quick email to Todd Hollenshead resulted in the predictable response:

I went because we like what they're doing from a technical standpoint, they're a good business partner of ours, and they asked if I would come and talk about our relationship and DOOM 3. Frankly, there was little marketing benefit from that sort of event for a game like DOOM 3 and, as I said at the event, we're not under contract with them (apart from NDAs). I was there because I wanted to be there and I like getting an opportunity to interact with groups from the media. I wish I would have had a little more time to spend there.

[image]

Jen-Hsun’s pet Paper Tiger

During the dialogue with Todd, Jen-Hsun made it a point to downplay ATI as a rival despite their strong resurgence which coincided with the release of the R300 and R350. Rather, he was focused on Intel as a competitor, stating that NVIDIA lost its pre-eminent position to Intel because of the embedded graphics hardware market. “Free”, as he stated, “is hard to say no to”. Putting the onus on writers, Mr. Huang suggested including benchmarks of embedded chipsets, in order to demonstrate their inferiority to the public.

That’s fair enough, and in the future, this will be an issue that we, the media, will have to tackle. It’s easy to dismiss embedded graphics now, but if they dominate the market, game developers would be suicidal to not cater to the lowest common denominator. If the sales of add-in cards like Radeons and GeForces drop, there might not be a market for high-end games. It’s an unlikely scenario, but it’s ugly and possible enough to worry about.

Except, that isn’t NVIDIA’s concern right now. NVIDIA has to worry about ATI for now and for the immediate future, no matter how they try to paint matters. And indeed, NVIDIA’s CEO spent considerable time talking about ATI, though not how you might think.



SIDEBAR: Jen-Hsun Huang is NVIDIA’s best PR weapon. He has the amazing talent of being able to focus exclusively on you, speaking in a calm, deliberate fashion (without sounding labored or condescending), answering questions – serious and inane, while there’s a major analyst from a huge Wall Street firm awaiting his attention.

Paper tigers

Page:: ( 3 / 9 )

About-face

To those who haven’t been paying much attention, NVIDIA has been receiving a beating in several media outlets for their “optimizations” to certain driver releases, in order to improve the apparent performance of the NV3x line. There were charges, very fair in many cases, that these optimizations were little more than cheats.

Having had their noses bloodied and learned their lesson, NVIDIA took the opportunity to spell out their new policies at the Editor’s Day and take the high road – relatively speaking. They set up new policies on optimization, a farm of computers to automatically compare image quality between driver revisions and they’re promising more detail in driver release documentation. Here’s the relevant point from the FAQ in the presentation binder:

[image]

Q: How has your driver development changed as a result of the cheating accusations?
A: We will continue to do application specific performance optimizations and compatibility fixes. We met with Futuremark and discussed the optimizations they questioned. Some of the optimizations (such as intelligent color clears and clip planes) they viewed as too aggressive, so we have removed them from our drivers. We have new, stricter guidelines in place to insure that questionable driver optimizations do not get introduced in future driver builds including:

An optimization must produce the correct image
An optimization must accelerate more than just a benchmark
An optimization must not contain pre-computed state

Finger-pointing

One of the more interesting points Jen-Hsun tried to make was that ATI apparently spends far too much of its time worrying about what NVIDIA does. Allegedly, ATI engineers dissect NVIDIA drivers and attempt to spot optimizations that affect image quality. Then, in turn, they passed this information on to journalists at ATI Shader Day.

Not that it matters at all. As far as any gamer should be concerned, both companies keeping tabs on each other on this matter is a Good Thing™. While it may get out of hand as the competition heats up, at the moment it serves consumer interest by keeping both companies honest. In the end, gamers will get improving performance without sacrificing image quality. Engineers on either side are far more likely to spot and prove cheats than simple hardware reviewers. The only downside to this is that the media may become little more than mouthpieces for the companies, and thereby slacken their own independent investigation efforts.

Taking a different perspective on the matter, several developers present had a beef with end-users being able to toggle options that affected the quality of the game. Their point was that they wanted gamers to play the game the way that they envisioned it. That’s a fair point, but as the owner of a game, if I want to play it with all details turned down for maximum framerate, or a fancy driver effect to make it look like a black and white movie, that’s my prerogative.

In the end, the best policy is to make optimizations that affect image quality an option left off by default. It’s up to the gamer to decide whether or not he wants to exchange quality for framerates. The sooner both major chipset makers realize this, the better.



SIDEBAR: Attendees stayed at the W hotel in downtown San Francisco. It’s got nice rooms, but tries to be hip and chic a little too hard. It’s good for a laugh though – the W is home of the $4 snickers bar.

Things we can’t dispute

Page:: ( 4 / 9 )

Batting 0.024

NVIDIA representatives took pains to mention their driver stability and were backed up on this by several developers. They brought up a statistic about tech call support ratios. Despite comprising a large fraction of the chipsets used with the game, only 6 out of 250 calls would be regarding NVIDIA hardware. Unspoken, and likely exaggerated, is the implication that the remaining 244 were about “the red company”.

Having failed to provide the percentage of ATI owners that made the call, NVIDIA would therefore make themselves a seemingly easy target for ridicule on this point. However, although no developer spoke the words, more than one made the implication that ATI users made more than their fair share of calls. Our own experience has shown that the Catalyst drivers are an improvement over ATI’s past efforts, but they’re still somewhat behind the Detonator… err… ForceWare.

[image]

The devs themselves

There are more than enough rabid fans of each company out there that accusations of bias, buy-offs and marketing contracts are sure to fly at the developers and even media who attended the event. Todd Hollenshead already gave his reasons for attending on page two, but two quotes from Randy Pitchford of Gearbox shed more light on the matter.

I came to the conference because I respect nVidia's commitment to software quality (drivers) and I was impressed by their focus on developer support. I want to encourage that behavior amongst the IHV's.

nVidia did not pay me or coach me - I had no idea what their presentation was going to be or what any one else was going to show or talk about. The only thing nVidia requested was that I didn't say anything negative about their competitors. My objective was to talk about how Halo was engineered to use the various hardware on the PC and how nVidia participated through their developer relations program. I hoped that by doing so that the audience would learn that there's infinitely more to making a good graphics card than speed (although speed is important!). I also hoped that the nVidia folks would get the message from me that their focus on driver quality and developer relations is much appreciated from the content provider point of view.

That’s the official line, but in a related email, Randy also stated that “I didn't realize how ugly this is when I agreed to go to the nVidia event.” This is likely the sentiment of most devs involved. They saw a great marketing opportunity and doubtless don’t want to be embroiled in a battle between the top two video chipset manufacturers.

Randy arrived with a very clear and succinct presentation in which he explained how NVIDIA’s developer relations team helped him develop Halo for the PC. Despite the apparent similarities in architecture, a rewrite was necessary to get Halo running at the target 30fps framerate.

Halo’s PC port was the first time that Gearbox ever shared source code with a hardware company. Randy’s personal expectations were very low, though he did have high hopes. It turned out that NVIDIA’s assistance on the shader side of matters resulted in not just, as might be expected, improved performance on NVIDIA cards, but on offerings from both companies. There’s no reason to be cynical about Randy’s statement. Gearbox had even less to gain by going to the conference than id, since their game is already out and selling well.

Harvey Smith from ION Storm had a more casual reply:

the reasons i went:

* good coverage for our game.

* fun with friends in SF. (staying at a sweet hotel...heh.)

* i'm super excited about DX2--it just turned out awesome.

* my schedule is now loose enough where i can travel, relax, promote, speak at conferences, etc.

High Dynamic Range

Page:: ( 5 / 9 )

24 is too much?

Editor's Note: Brandon Bell: Gabe never said that FP16 wasn't enough for HDR though, it's just that their [Valve's] implementation uses floating point render targets which currently aren't supported by NVIDIA drivers. You can see this in the ShaderMark testing with GeForce FX 5700U, where the ATI cards are able to complete HDR tests, but the NVIDIA cards can't run them.

As you can see above, Brandon corrected a misconception of mine. I’ve left the remainder of the article untouched, in order to maintain integrity, but please keep the above in mind when reading the following. We’ll also be looking further into the non-functioning ShaderMark tests. -ed

One of Gabe Newell’s most memorable accusations levied at NVIDIA was that 16-bit is not enough for High Dynamic Range, or HDR. While we were awaiting answers from the developers we polled on this, I decided to see if I can find out for myself just what, exactly, High Dynamic Range is. NVIDIA’s GDC presentation and the SIGGRAPH High Dynamic Range Image Editor pdf can help explain in detail.

First, we have to discuss the dynamic range. A dynamic range is the ratio of the maximum intensity in an image to the minimum detectable intensity, according to the definition used at NVIDIA’s GDC 2002 presentation. What this means for computers, which work solely with numbers, is that in a 3D scene, every pixel is mapped to a brightness on a set scale. Human eyes of course have no such limitation, after a few minutes in the dark we can make out details in a dark tunnel while still seeing light on the other side, and after coming out of the tunnel our eyes can adjust to a sunny day in the park to track and catch a football.

[image]

Is 16-bit enough?

Now that’s the $65,535 question! Is Gabe Newell right and 16-bit HDR is inadequate? Yes and no. Perhaps, for Half-Life 2’s purposes, there will be such drastic changes of light intensity that 16-bit has the same relative weakness that 16-bit rendering does.

Then again, compared to LDR, 16-bit HDR is a vast improvement. An analogy might be to take a look at the earliest 3D games, accelerated with 16-bit Voodoo Graphics cards looked infinitely better than software rendered, 8-bit images. It took until Quake III: Arena and Unreal Tournament before 32-bit rendering became a big deal. In my humble opinion, most people will simply be too struck by the difference between HDR and LDR to notice the limitations of 16-bit. Particularly since you could bet the bank against the fact that most devs aren’t fans of financial suicide and will accommodate both cards.

Taking a look at Doom 3, which has far superior lighting effects when compared to Half-Life 2, it makes me wonder just what exactly Half-Life 2 does that it needs 24-bit HDR. Then again, Half-Life 2 combines outdoor and indoor scenes on the same level, and may require that higher scale.

However, I have little doubt that in a few years, we will see a migration towards 24- or 32-bit HDR. 16-bit may very well be enough now, just as 16-bit 3D acceleration was good enough for Quake II, but progress is a fact of life. This same argument can be used against NVIDIA’s claim that mixed-mode 16/32-bit is better than 24-bit pixel shading. 24-bit is likely good enough for the life expectancy of the current generation of cards, and we might not see a real need for 32-bit until the next generation of games comes around.

Still not sure about the importance of HDR? We got some comments from Tim Sweeney on that and more.



SIDEBAR: I approached many devs, both those that attended the conference and those that did not, for their comments. Few replied, even fewer agreed to be quoted.

Tim Sweeney on the matter

Page:: ( 6 / 9 )

HDR

Tim Sweeney as a more technical explanation of HDR.

Tim Sweeney: The key aspect of HDR is to render scenes with higher color fidelity, generally 16 to 32 bits per color component (instead of 8). This approach enables you to properly represent the full range of lighting and environments you see in a real-world scene.

Once you have HDR, there are lots of effects you can add on top to improve visual fidelity, such as overbrightening (having pixel colors spill into their adjacent neighbors to simulate you're eye's response to bright lights in otherwise dark environments), occular and lens effects (generating growing per-pixel corona patterns), and simulating your eyes' gradual adaptation to bright and dark environments.

You can also hack in overbrightening and lens flares in DirectX7 hardware, or even in a software renderer. Unreal 1 did these sorts of things in various stages of its development. But these are just low-precision postprocessing techniques to add neat effects on the screen; they miss out on the key benefit of HDR, which is the ability to use the higher precision to represent the color range in scenes more accurately. Eye candy: yes; high precision color: no.

[image]

Tim Redux

Page:: ( 7 / 9 )

24-bit is not enough?

Going back for a professional opinion on the “24-bit is not enough” point that NVIDIA had:

FiringSquad: NVIDIA made a big point of saying 24-bit precision is not enough. While this is doubtless true for future apps, is 24-bit precision a limitation to you now, or you expect it to be in the near future?

Tim Sweeney: 24-bit seems fine for UT2004, DOOM 3, and Half-Life 2, so your gaming needs should be satisfied up till mid-2004. After that, who knows? There are lots of CPU applications where 32-bit floating point isn't enough, and you need to use 64-bit floating point for best results.

[image]

ION Storm replies

Page:: ( 8 / 9 )

Questions galore

Like Epic, ION Storm was also generous with their time and they provided answers to some questions as well, this time provided in simply Q&A format. (Note: questions were edited after the fact for brevity and clarity.) Tim Little is the Director of Technology at ION Storm and has been responsible for the development of their own custom rasterizer, Iris, to go along with the rest of the Unreal Engine technology they licensed from Epic.

[image]

Conclusion

Page:: ( 9 / 9 )

The tally

It’s become rather trendy to bash NVIDIA lately. People like winners and people love underdogs. ATI is both right now – they’ve climbed their way out of the abyss and even disregarding the NV30 production delays, their timetable was catching up to NVIDIA’s. NVIDIA, in the meanwhile, is the top dog that stumbled, and it’s always fun taking the king of the hill down. Ignoring all those feelings, putting fanboyism aside, is quite important on getting a clear perspective on matters.

[image]

Final impressions

It’s difficult not to come away with a guardedly optimistic impression of NVIDIA and their prospects, despite my initial cynicism. They had an open forum during the whole conference, answering questions and accusations. They acknowledged past mistakes, and built upon them. Some of NVIDIA’s comments went through the PR processing machine a few too many times, but in general their honesty and forthrightness was admirable.

And none of that would matter at all if the developers hadn’t been there. Yes, most wanted the opportunity to market their upcoming games, but a few didn’t really need that. id Software, not just John Carmack, have a belief in NVIDIA cards. That belief is that the cards look further into the future with their features and design.

It’s tempting to suggest, “who do you think is right – John Carmack or Gabe Newell”? We could investigate the marketing contract between ATI and Valve (pointedly, NVIDIA doesn’t have one with id.) We could make detailed comparisons between Doom 3 and Half-Life 2, to decide which looks better. In reality though, that doesn’t matter. There are certain advantages that each card has.

Once again, it comes down to the numbers that don’t lie (*cough*) – benchmarks. NVIDIA has caught up. It’s taken determined effort by their driver and developer relations teams, but they’ve done it. Are there game-specific optimizations? Certainly. There isn’t a driver out there that doesn’t have them. As Randy Pitchford pointed out, devs will always optimize for the most popular video cards and video card makers will always optimize for the most popular games. Are there still games where NVIDIA lags behind? Absolutely. Quake III (with FSAA and AF) and UT2003 are perfect examples, but on the other hand they have a strong lead in Splinter Cell and NASCAR 2003 now.

What the benchmarks have proven is that NVIDIA’s hardware is as fast as ATI’s, depending on the game. Yes, it does take more work – NVIDIA admitted as much. The NV3X platform isn’t as easy to program fast as R300 and R350 are. On the other hand, people who own GeForce FX cards will be able to take advantage of games that feature longer shader instructions. Of course, whether any cards of this generation will be fast enough to take advantage of the higher precision and longer pixel shader instructions when the games finally arrive, is another matter.

FiringSquad would like to thank Harvey Smith of ION Storm, Todd Hollenshead of id Software, Randy Pitchford of Gearbox Software, and especially the two Tims – Tim Little of ION Storm and Tim Sweeney of Epic.

Got comments? Questions? Opinions? Flames to be thrown? Then don thine asbestos dueling armor, knight, for thou is about to Sound Off! in ye olde news comments.