DirectX 9.0 Vs 10.0

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Good god, some people have no idea about computer architecture:

While the DX11 binaries will have some "impovements" (probably updated DX10 .dll's) to are said to speed up DX10 execution, DX10/10.1 cards WILL NOT BE ABLE TO RUN DX11! This is due to incomplete hardware support, and Microsofts own Display Driver Model mandating that to be a DX compliant card, you must support EVERY feature for that DX version. Even though ATI cards have a tesselation unit, due to lacking support for a handful of other DX11 features, the advantages of DX11 will not be taken advantage of, even though the card itself if perfectly capable of tesselation.

You want DX11? Spend $500 on a new GPU. As I've held for a while though, until XP dies, don't expect to see much.
 
In the name of the greatest OS that has ever trod this earth, I draw the line in the dust and toss the gauntlet before the feet of tyranny, and I say XP now, XP tomorrow, XP forever"

Jammin' with yer inner Wallace...
 
I expect to see Direct 9.0c and Windows XP gone soon (I know that this might not happen soon) so that the PC games can fully be optimize for DX 10 and DX 11 but today, most games aren't fully DX 10 games because DX 10 is not fully use/optimize and it is still partially DX 9 in order to be compatible with DX 9 and allowing many users who are still stuck with Windows XP to buy these new games. The PC game market is still depended on DX 9.0c and XP users since many are still stuck with Windows XP + DX 9.0c so most games will not show much improvement from DX 9 to DX 10. That is why we don't notice a lot of difference of graphic improvement with the current DX 9 + DX 10-compatible games. 😉
 
DX10 wasn't that big a technology anyway, more of an under the hood update. Tesselation for DX11 is the only major feature I see for DX11, and I personally have questions about what type of performance hit using Tesselation will cause.
 
Funny so many speculations but nothing substantial.
Use Google and search DX11 new API.
DX11 is more about the implementation of the new API.

It's funny thought to see how people pull facts out of their arses and argue their believe and opinion as they are nothing but the absolute truth. 😀
 
Sorry for the late reply to the thread but I was in Deadmonton and then making the most of the last few days of snow. :sol:



I understand what you're saying, and I SEE your point, I just don't aggree with it anymore than I did with with it last time, nor those who said it about Win98SE when XP came out and made the Win2K/XP path king 2 years later. The difference being, if you're limited in time and space, then it's far FAR easier to drop the XP path than to drop the DX11 Vista tack on. XP+DX9 is far FAR more of a resource hog to developers for your scenario than than just the move to add DX 11. But you don't see that it's not an either/or option, it's something where DX11 most of the resources are suken costs to DX10, so the effort and ROI is easily justifiable, XP is title and market dependant. You make it sound as if it's an imposibility and as if it's anywhere close to the situation we had at the Vista / DX10 launch.

[/quote]Until XP drops below 20% market share, it makes no sense to sell a product with a handful of extra features that could potentally lose 20% in sales as a result of coding to a DX10+ standard.[/quote]

Unless that tiny market costs you a significant amount of dev time and delays yours product to market. If Crytek, Dice, Epic or Valve were launching a new title the effort required to code for XP's single DX9 path would be a huge investment of time and resources, while the DX11 tack on would be part of the standard development and give people a way forward for many (like unlockable content) whereas the XP fallback is an unused option for 80+% of the market for your scenario.

Throw in the lag in the general public aquiring DX11 hardware, and the difficulty of having to ensure three seperate graphic API's work (Plus three OS codepaths, XP running DX9 and Vista/7 running either DX9E, DX10, DX10.1, or DX11) and you see why its so easy to simply stick with DX9.

No, it's not, you make it sound like adding DX11 and Vista are optional, and they're not Vista is required for any new game period, thus a sunken cost, so the difference at worst is XP+DX9 vs VISTA's added DX10.1/DX11 work. And the two don't come close to comparing for resources. BTW, what's this third OS codepath? Win7 will be like the XP/Win2K path, interchangeable.

Using your logic, we'd be seeing a heck of a lot more DX10.1 games, considiering 10.1 is such a minor addition to DX10.

No we wouldn't, because there was a disconnect on the adoption of DX10.1, nV didn't want to play with ATi & S3 so that's created a much cleaner break than DX11 while where both ATi & nV's next hardware will be equal in that respect. Based on your logic there should be NO DX10.1 features in any titles, because there's more than 20% of users on XP and much more than 20% on DX9 hardware let alone DX10.0.

In short, developers for games could care less about "the enthusiast crowd". They want to make money, and coding to DX9, and spending only minimal time on advanced DX features, is the best way for developers to do that.

In short, you're dead wrong!
Many big title developers DO care about the Enthusiast crowd, and the Performance crowd, both of which BUY games in large numbers and are the CORE of PC gaming. Those who would reject a game because it requires Vista aren't those who consistently buy new titles and pre-pay to make sure they have it. Those are the people the developers want, not those who stick to XP and DX9 hardware forever and buy their titles when they go down to $9.99 in the discount bin. And even for those gamers, by the time and DX11 title reaches the discount bin that gamer would likely be moving to Win7 and a cheap DX10 card anyways. This is the same split as Win98SE or AGP vs PCIe, eventually they simply become irrelevant to new income.
Those developers that would focus their efforts more on the XP crowd are the traditionally low requirement strata like SIMs, WOW, and EA Sports franchises, that rely on the sales of a low-end install base who wouldn't be classified as PC 'gamers' so much as people who play a game or two on PCs, not the big title launches of new games / franchises which have always pushed the envelope. And most of those low-end people not in the WOW market are moving to consoles anyways.
I'm not saying there won't be XP titles and those titles that support XP and DX8 & 9 vs anything DX10 based, but those are the same ones that basically just dropped Win98 support, like NHL08 had last year and NHL09 just switched to XP/Vista-only.
What you're saying is not that balance of understanding, you say that no-one will put that minimal effort into DX11, going so far as saying there's no point to DX10, which is ridiculous! Just as ridiculous as hoping for a DX10 port for XP from a random 'group' as if they have more time and motivation for free than the developers who actually see ROI. [:thegreatgrapeape:5]

You talk about DX11 as if it's this huge undertaking, and about Vista as if it were still the pariah it was at launch, when in reality the bigger barrier for developers is not taking on DX11 to their already MUST have Vista workload, so much as the far bigger effort of adding XP support to the those dwindling ranks who are not expanding but gettting smaller all the time. Not every game, nor even most of the games will be DX10.1 or DX11, however, many of them will be, and unlike what you're saying, it's not going to be as hard a step into DX11 as it was into DX10+Vista.

Anyone who would consider themselves a PC GAMER who buys new games when they're new (not playing just CS for the past decade) now has at least DX10 capable hardware, and likely either has Vista or has been waiting for a reason to switch once it improved enough. That time has come, and it will be only more ripe months from now when DX11 hardware shows up.
 


Yes, directX10 cards are capable of running directX9 applications.
 
If the majority of the gaming market was *enthusiests* as you claim, I doubt the source engine would be the most popular game engine yet created, as the games on it would hardly interest that particular crowd. Heck, the majority of the PC gaming crowd still runs CS:S, HL2, or WoW, hardly taxing games...

And again, DX11 faces the same issue DX10 did. Same overall architecture, but with the split API intact. While not anymore difficult to implement, programmers hate having to maintain code. You are more or less required to have DX9 as a result of XP, and if you want a higher DX, you need to use DX10 as the result of all the hardware support (compared to DX11+, which will have the standard 2 year hardware upgrade cycle for the majority of users).

DX11 is an issue of the split API, code maintainability, and even program size (I've noticed some DVD's getting close to the 9GB barrier...). While XP mode will get some users to switch to 7 (until they find out they also need to upgrade to 2GB of RAM, which kills the idea as far as I'm concerned), XP will be around for some time, because M$ still hasn't given users a definative reason to upgrade from XP; no extra stability, no major features, no significant speed increase, etc.
 


is it? Unreal3 is the most popular game engine across multiple platforms. most games that use source are the ones from valve themselves.
 



...and of these titles, exceedingly few *new* sales. Rather it's the same people who keep playing the same games they're been playing for a while. But it's new sales that drive development. As pointed out above - the $9.99 bargain bin is not where PC game developers are aiming to be. Not to mention the cracked/pirated titles, which generate no money for anyone.

Were is the profit?






Gee.... I would have sworn game size was more due to the ever more detailed graphics, effects, and AI, rather than something as low level as the API. But what do we know... :sarcastic:

And using a RAM argument?? Dude, you're really reaching now...
 
You misunderstand. How many games did Crysis sell? Or even Unreal 3? (Remember, we're talking PC here). There are very, very few games designed exclusivly for enthusiests, and those games almost never sell well. And for the record, the most used engine is probably Source, the engine for most new sales would be Havok (as much as I want it to die, but thats another arugment). Neither of these engines can really be described as "cutting edge" by any stretch of the imagination, and last I checked, L4D is quite popular...

Remember, its total sales that matter most. Value for PC games drops quickly, so its how many you can sell the first few weeks, and then how long you can continue to sell the game. 5 copies of CS:S at $9.99 is worth just as much as one sale of Crysis at $59.99.

And my point with the 9GB was in regards to the physical limitations of DVD disks, and has nothing to do with RAM. More disks cut into profit margins, so if its a close shave, its those minor improvements that get cut. And the API has nothing to do with that size, rather, the IMPLEMENTATION of the API. Its one thing to create a .DLL for DX11 that allows for tesselation; its another to actually implement it in the games .exe. A minor factor, sure, but a factor non the less.
 





Actually sir - I feel it's you who is misinterpreting things: Firstly, I responded to the examples you provided, no more, no less. Please don't change the tune to 'Crysis' in mid-argument. Besides, Crysis sold a million copies before January: http://www.tomshardware.com/forum/99063-13-crysis-sales-exceed-million So I would hardly call that "almost never sell".



Most importantly, though, it's not about games which have already been written. It's about what's going to be written next.


Regarding the RAM comment:

While XP mode will get some users to switch to 7 (until they find out they also need to upgrade to 2GB of RAM, which kills the idea as far as I'm concerned),


I was responding to your complaint about a computer needing 2GB of RAM. Never mind that XP games run on Vista and 7, and that you can't have hardware acceleration in a virtualized environment anyhow. The XP virtualization provided for free as a download for Win 7 exists simply to provide backwards compatibility to small businesses. No gamer in his/her right mind would consider using it because lf the lack of hardware acceleration, and companies with the larger support packages have access to a much heavier duty virtualization engine anyhow.


And thank you for admitting that the API takes up verly little space.
 
Seriously, how much would it cost for dics, a pittance i woudl imagine considering the quantities involved.

How many pc gamers would complain about inserting a second disc?

None, if they are an actual pc gamer, games used to come on multiple floppies, then multiple cd's so who cares about dvd's.

Only it costs more to burn in two DVD's instead of one. Multiply that by a couple million, and you see an effect two disks has on the bottom line.
 
If they charged 50 cents more for the games, theyd be making profits on having 2 discs, so to me, itd make more sense to charge that 50 cents, than to cut into the games abilities. Packaging shouldnt have as large an impact on a products viability, period.
 
In the context of a Win 7 article: http://www.anandtech.com/systems/showdoc.aspx?i=3557&p=5


Although it is being released for Vista too, Windows 7 marks the official introduction of Direct3D 11. As a superset of Direct3D 10.1, Direct3D 11 adds support for tessellation, multi-threaded rendering, and GPGPU abilities via the Compute Shader. We’ve already covered a great deal on Direct3D 11, so please see our introduction article on it for more details.

Next up we have the Windows Display Driver Model 1.1, which in turn is the lynchpin for several other graphics related features. WDDM 1.1 itself is not particularly impressive, but it’s what it allows that is. For all practical purposes, all we need to know about WDDM 1.1 is that it’s a minor revision of WDDM that requires DX10-class hardware (rather than DX9-class on WDDM 1.0) and as such allows the operating system additional features.

So what can you do with WDDM 1.1? For starters, you can significantly curtail memory usage for the Desktop Window Manager when it’s enabled for Aero. With the DWM enabled, every window is an uncompressed texture in order for it to be processed by the video card. The problem with this is that when it comes to windows drawn with Microsoft’s older GDI/GDI+ technology, the DWM needs two copies of the data – one on the video card for rendering purposes, and another copy in main memory for the DWM to work on. Because these textures are uncompressed, the amount of memory a single window takes is the product of its size, specifically: Width X Height x 4 bytes of color information.


Furthermore while a single window may not be too bad, additional windows compound this problem. In this case Microsoft lists the memory consumption of 15 1600x1200 windows at 109MB. This isn’t a problem for the video card, which has plenty of memory dedicated for the task, but for system memory it’s another issue since it’s eating into memory that could be used for something else. With WDDM 1.1, Microsoft has been able remove the copy of the texture from system memory and operate solely on the contents in video memory. As a result the memory consumption of Windows is immediately reduced, potentially by hundreds of megabytes.

What makes this even more interesting is how this was accomplished. With WinXP, GDI+ was partially accelerated, but this ability was lost to little fanfare when the DWM was introduced for Vista and thereby made GDI+ acceleration impossible, pushing all GDI work back to the CPU. This in turn is responsible for the need for a local copy of GDI and GDI+ windows and the increased memory usage of Vista, as reading data back from the video memory for the CPU to work on is too slow to be practical. The solution as it turns out is that by reintroducing GDI acceleration, the amount of read-backs can be minimized to the point that a local copy of the texture is no longer necessary. Now GDI isn’t completely accelerated, only the most common calls are, but this is enough that when the least common calls trigger a read-back, the performance hit is negligible.
 


That was kind of my impression. I thought DX10 was more or less a stepping stone to something greater/advanced. With everything, time will tell.
 


Well with Crysis you'd have to talk PC for launch sales because there was nothing else and it did damn fine sales for PC outselling most other major titles, including the much anticipated Spore. Unreal3 was just a dissapointing game, but games built off the UE3 engine are quite popular and wide ranging.

There are very, very few games designed exclusivly for enthusiests, and those games almost never sell well.

I didn't say exclusively for enthusiasts, thats' why they HAD DX10 + DX9 SM2/3, but not DX8 support, and why the new split would be DX11, DX10+, DX9 SM3 because you catter to the Enthusiast AND PErformance crowd that do spend $100-200 on a Graphics card every couple of years. And they do sell well, just like COD4, Bioshock, Crysis, Gears of War all sold very well on PC. They don't sell as well as on console, but like I already mentioned those are not the PC gaming crowd, and making a game crap to run on a GMA950 is not going to attract more people to PC gaming because you increased the 'possible' install base 100 fold.

And for the record, the most used engine is probably Source, the engine for most new sales would be Havok (as much as I want it to die, but thats another arugment).

Now you prove you don't know WTF you're talking about, I suspected it in the PhysX discussions, but that just proves it. Havok is not a GAME engine it's a physics engine that tacks on to others like Source, like Gambryo, etc.
And considering your limited understanding of the source engine and it's split upgrades, you wouldn't appreciate the difference between the various Unreal Engines which likely have more titles than any 2 other engines combined, but being specific to UE3 it might not have as many sales as the overall source catalogue, but for a split around the same time of developement. Source needs an update, just like UE3 did, and with UE4 coming the Source toolset is beginning to look a little old and limiting due to it's reliance on the old.

Remember, its total sales that matter most. Value for PC games drops quickly, so its how many you can sell the first few weeks, and then how long you can continue to sell the game. 5 copies of CS:S at $9.99 is worth just as much as one sale of Crysis at $59.99.

Once again, proving you know nothing of which you speak. Other than just the basic math of 5x9.99 means you lose 4 cents upfront.
Anywhooo, 5 copies of a title for $10/piece that shipped to stores @ $59.99 do not provide the same margin or profit as 1 copy sold at launch. The cost of making the game and shipping it, and putting it on the shelf, and all of tha mean that the profit margin is reduced to near nil on those 5 that have also not gained interest for the devloper, whereas the $60 launch title, paid for itself, it's shipping, the marketing, and everything else that becomes a sunken cost, and then the developer has all of the interest they earn on that money versus some future sale. They are nowehere near the same. Only on a no additional cost distribution system like Steam would there be the same $ figure returned, but once again with the $60 upfront earning interest or paying down accrued debt.
I doubt Epic would care to collect $2 from someone still wanting to buy UT2K3, it wouldn't be worth the effort.

And my point with the 9GB was in regards to the physical limitations of DVD disks, and has nothing to do with RAM. More disks cut into profit margins, so if its a close shave, its those minor improvements that get cut.

9GB of compressed data, which often expand to about 20GB on computers, and 2 DVD costs pennies compared to everything else. Heck the packaging likely cost more than 2-3 pressed DVD inside.

Seriously, for all the reasons you come up with, none of them make sense for a relatively minor add to add DX11 support compared to these strawmen you create as if developers would never make titles for enthusiasts as if have a lagrer market versus a more profitable market mattered most.

Selling 100 million copies of a game at $5 a pop sounds awesome and might make it the most $ figure sales ever, but if it costs you $4.50 to get those titles to the public, it's nowhere near as impressive as selling 10 million copies at $30 a pop with a $5 a disk cost figure due to the added DX11 cost.

Once again, you try to maximize profits, but you're making it sound like it's a huge effort to add DX11, and as DICE is saying, it's really alot easier than they thought.

Figure out how you explain that statement than there is no point to DX10, because it's ridiculous. And saying DX11 split and implementation are the same as the one facing DX10, ignores what a pain it was for DX10 & Vista.

Anywhooo, I'm done for now; I'm off to watch the hockey games, then off to Sunshine to go SKIING !!! :sol: