Crysis: DX9 vs. DX10 @ [H]

I've run Crysis (High/maxAA/1680x1050) on my machine under Windows XP x86 and x64... i did notice a jump up when I booted into my x64 partition over the x86 partition (but this was only on the flight deck of the aircraft carrier when the boss comes out)..

I did NOT however notice any drop-off when going from x64 Windows XP (DX9) to Windows Vista x64 SP1 (DX10). it still performs rock-solid.

So by that rationale DX9-XP-32bit runs SLOWER than DX10-Vista-64bit... are you saying they should test in XP-32bit-DX9 (common whore) or XP 64bit DX9 (rare)???

my guess is that overall XP64 will give the best results... but not a noticible difference from Vista64..

and i know i excluded Vista x86... but who cares it was a moronic release for morons.

BTW.. people.. must we use IQ for Image Quality.. it's rather confusing dontcha think??

Thanks for this, I was wondering if there was any noticeable difference between XP & Vista 64bit (although most games arent very optimised for 64bit). I have XP Pro 32/64bit and Vista Business 32/64bit. I am currently dual booting XP32/64 with 64bit being a permanent clean os for gaming (except up-to-date and with an AV), and 32bit being my sloppy os with programs installed that have been used only once for one purpose at the time. Now that I've recently upgraded my ram to 4GB and have two 8800GTS's on the way, I'm planning on dual booting XP64 and Vista64, playing dx9 games on XP and dx10 on Vista.

I've played around with a XP64 and Vista Ultimate 32bit install on a core2 E8400 system and it performed very poor and sluggish compared to XP. I was hoping that it had something to do with the Ultimate bloatware. Your post gives me hope that I may actually enjoy dual booting with Vista 64 :p
 
Thanks for this, I was wondering if there was any noticeable difference between XP & Vista 64bit (although most games arent very optimised for 64bit). I have XP Pro 32/64bit and Vista Business 32/64bit. I am currently dual booting XP32/64 with 64bit being a permanent clean os for gaming (except up-to-date and with an AV), and 32bit being my sloppy os with programs installed that have been used only once for one purpose at the time. Now that I've recently upgraded my ram to 4GB and have two 8800GTS's on the way, I'm planning on dual booting XP64 and Vista64, playing dx9 games on XP and dx10 on Vista.

I've played around with a XP64 and Vista Ultimate 32bit install on a core2 E8400 system and it performed very poor and sluggish compared to XP. I was hoping that it had something to do with the Ultimate bloatware. Your post gives me hope that I may actually enjoy dual booting with Vista 64 :p

The thing with Vista is that you really REALLY need to give it some time to learn your computing behaviour. At first it'll feel somewhat slow and sluggish (and is really dependant on RAM, so the more the merrier). Once it's done indexing, and tuning it's prefetch settings, etc it's WAY faster than XP from a desktop usage perspective.

Having used Vista x64 since RTM I would never go back to XP. It feels like too much of a downgrade. My wife has also used Vista since about that time. I asked her if she'd like to go back to XP. Her response was basically HELL NO.
 
Ever tried Call of Duty?
There low 20 FPS dosn't matter like in other games...
+60 FPS does NOT hold true to all games, do some research.

It was just a general rule of thumb that was used back in the quake2/3 days that 60 was the goal and 30 was the minimum playable number. It does hold true for most games that is for sure....
 
I've played around with a XP64 and Vista Ultimate 32bit install on a core2 E8400 system and it performed very poor and sluggish compared to XP. I was hoping that it had something to do with the Ultimate bloatware. Your post gives me hope that I may actually enjoy dual booting with Vista 64 :p
Ugh, again with the 'bloatware' fud.
Vista takes time to 'break in'. My laptop (see sig) took about a month or so (I don't use it much) before it didn't spend the first20 minutes thrashing my HD and readyboost drive. Now, it's crazy fast and boots in ~40 seconds, which is very similar to some of the SSD based systems out there.
 
Awesome, that is good to know. I knew the prefetching and indexing would take a little time to ease in, but this was on a brand new hard drive. I've also cleared my XP prefetch many times with little slowdowns so I didn't expect that to factor in, but I guess Vista just takes longer than XP. And even when applying all of the updates (right before SP1 was released) it took 1.5 to 2 hours to update with a campus connection that gets something over 1MB/s down and 900KB/s up. It was sitting at the applying update screen for a very long time, so long that I checked the task manager twice to make sure it didn't lock up. Yet I hear people saying all of the updates take little over 30 mins.

The system was a E8400, nforce 780i, 4GB RAM, two 250GB seagate 7200.10's. On the other hand I have seen a core2 dell running Vista Home 32bit with very few slowdowns, and that's why I assumed it was Ultimate edition because it comes bundled with all the extras. I guess I should get Vista installed soon to get more day to day experience with it.
 
So it's clear... as long as I play Crysis on Windows XP
ATI 3870 is the winner by far.
Thanks fot the review!
 
Yea with older drivers and before the patch came out. Just like with Crysis, you did this article to see what improved drivers and game patches would net us. Now, I am not expecting you to include Bioshock, I'm just pointing out how your response could be confusing.

I don't think it's very confusing, you look at the date of the article....

Besides, it's stated in the conclusion that Nvidia cards don't take a hit going from DX9 to DX10, and they did a good job of showing the differences between the two in image quality and graphics.
 
I don't think it's very confusing, you look at the date of the article....

Besides, it's stated in the conclusion that Nvidia cards don't take a hit going from DX9 to DX10, and they did a good job of showing the differences between the two in image quality and graphics.

Yes I did. September of 2007 which was before a bunch of driver versions and the bioshock patch- the same mindset the article stated for Crysis. Maybe you didn't understand the sentiments of more than half the posters in this thread, the hit from DX9 to DX10 is better represented if you use Windows XP as the DX9 variable. It is better represented in the sense of the particular direction and questions that this particular article is trying to find an answer to.
 
Kyle, could you add Dx9 text on windows Xp to that?
It's probably a troublemaker but.. i'm curious about it, and from what i have read, several here think the same.

Thanks :)
 
I didn't read the 7 pages of posts, but I did read the original article --- somebody may have already commented on this...

Do I understand correctly that the DX9 vs. DX10 graphics differences were compared without all settings being on the highest possible settings? (Very High) I understand that "playable settings" was the intent here ---- BUT to say that DX10 isn't that much different than DX9 when all settings are pretty much medium doesn't reveall the full story of the visual differences between DX10 and DX9. Crysis programmers have stated that DX10 settings don't even turn on unless shaders and shadows are set to the highest setting in DX10 - if I remember correctly. So there really shouldn't even be much difference between DX10 and DX9 with medium or even "high" settings, but if you put DX10 card on Very high and suffer the poor framerates to get some screen captures, and compare those to DX9 on high (because very high isn't even available with DX9) Then you'll REALLY notice some differences - I know this is the case - because I've done these tests on my own machine. DX10 looks clearly better - even if unplayable at those very high settings.

The article has merit to tell Nvidia users to run DX10 because there isn't enough of a hit to worry about and to tell ATI users to run DX9 because at playable settings they won't be able to tell a difference between DX9 and DX10 but they will be able to see the performance hit.

But an aside note might be made: that DX10 looks better than DX9 at full very high settings ---- if/when the hardware is capable. In my opinion the graphical variance in DX10 and DX9 is not disappointing - thus DX10 isn't a disappointment. The hardware needs to catch up with the software again.
 
Modred, that paragraph does not say whether those users are on XP or Vista. Logic dictates that a Vista user doesn't need those benchmarks, because they can easily run them themselves. An XP users, OTOH, doesn't have Vista and thus doesn't know if DX10 has caught up to what their boxes can do or not.

It matters, and it should have been the point of the article.
Not at all. The point is, if your XP DX9 performance is close to the DX9 performance in the article, then you can expect it to stay the same (assuming you have similar hardware). Consequently, since the DX10 performance is pretty much = to the DX9 performance (on Nvidia cards) you can expect similar results.

If, however you have better performance than the DX9 in the article you can expect DX9 hits, and since the DX10 performance is pretty much the same, worse performance in DX 10.

And the point of the article is NOT XP versus Vista. It's not addressed because it doesn't matter. In XP you don't have the option of DX10. The point of the article was to address whether or not [H] should use the DX10 or DX9 pathways while evaluating cards. If there had been a difference between teh two there owuld have been an argument for using one over the other. i.e. if DX9 was better, you could say that it would be a better option for gamers.
However, since using DX10 provides some (if barely any) IQ benefit and little or no penalty as far as fps, it remains the logical choice between the two.

This was not an evaluation of DX9 vs DX10 as a gaming platform, rather their appropriateness as testing platforms.

Now, If you want [H] to do a DX10 vs DX9 (and vista vs xp at teh same time) evaluation, ok, but that is a different argument. That is a different article and should not have been included here.
 
And the point of the article is NOT XP versus Vista. It's not addressed because it doesn't matter. In XP you don't have the option of DX10. The point of the article was to address whether or not [H] should use the DX10 or DX9 pathways while evaluating cards. If there had been a difference between teh two there owuld have been an argument for using one over the other. i.e. if DX9 was better, you could say that it would be a better option for gamers.
However, since using DX10 provides some (if barely any) IQ benefit and little or no penalty as far as fps, it remains the logical choice between the two.

This was not an evaluation of DX9 vs DX10 as a gaming platform, rather their appropriateness as testing platforms.

Now, If you want [H] to do a DX10 vs DX9 (and vista vs xp at teh same time) evaluation, ok, but that is a different argument. That is a different article and should not have been included here.

which is all fine and dandy, i can see your point...but whats the first thing you think of when you hear DX 10 vs DX9? i know what i think.....Vista vs. XP....because even though DX 10 isn't an option in XP, it doesn't mean that DX 9 shouldn't be represented by XP....

i whole-heartedly agree with the statement...if if own XP - i'm running DX9....if i own Vista - i'm running DX10....basically, without using XP, we're seeing how Vista can handle DX9....

IMHO i think (even though it's NOT an article of XP vs. Vista) XP should have represented DX9 and Vista as DX10 because that is how the AVERAGE gamer would play their games.

that said...was it a helpful article? yes, would it have been better if they included XP for DX9? yes....but at the end of the day, it's not destructive criticism.
 
@Modred189

I think the facts are that Crysis and most other games out now still run better under winXP.

And what people are trying to drive home here in this thread is that if [H] had done a dx9 compare using winXP the performance would have been 10-15% better for dx9. ;)

And anyone who thinks dx9 cant look as good as dx10 in Crysis is fooling themself... just because certain settings are locked out for dx9 by default does not mean you cant turn them on yourself.
And get the same IQ but with alot better performance.

Just my .02
 
I think you can turn almost all of the so called DX10 features on in DX9 with some config edits, but there are still certain things that are truly DX10 specific and give it that edge
 
The only conceivable issue I think anyone can have with the article is that only one title was tested out. Perhaps a couple of others could be tested for a "part 2" of the article?

(maybe Bioshock?)
They already did that.Dont you click on the links?
 
Very Cool Brent, I forgot all about that review! Thanks!

Any Way you could Link the Bioshock, Lost Planet and Call of Juarez Reviews at the bottom of the Crysis Review Page?

Uhh again, they did. Despite that the links should be named anything bit quote: "here, here and here"
They should had placed the actual names of the games.
 
You answered quite a few Questions for me over DX9 Vs DX10 gameplay. Great Review as always, and it just confirms my belief that I saw no appreciable difference either.
 
@Modred189

I think the facts are that Crysis and most other games out now still run better under winXP.

And what people are trying to drive home here in this thread is that if [H] had done a dx9 compare using winXP the performance would have been 10-15% better for dx9. ;)
WHO CARES.
The point of this article was whether [H] should use DX9 or DX10 Crysis in their evaluations. XP has absolutely NOTHING to do with that question.
Period.
And anyone who thinks dx9 cant look as good as dx10 in Crysis is fooling themself... just because certain settings are locked out for dx9 by default does not mean you cant turn them on yourself.
And get the same IQ but with alot better performance.

Just my .02
A: You can enable some of the eye candy, but DX9 cannot support the contrast capabilities, nor can you enable the particle-to-surface interaction capabilities. (The smoke example)

B: Did you read the article where there was a 2-3 fps difference? That's not "alot better performance"

PS: It's -- a lot--, not --alot--
 
WHO CARES.
The point of this article was whether [H] should use DX9 or DX10 Crysis in their evaluations. XP has absolutely NOTHING to do with that question.
Period.

but it does matter because when i think of DX9 gaming, its associated with XP...really it's not trying to find out which OS is better...i'm not running Vista to run DX9, i'm running XP (the "I" in this is your average gamer) so wouldn't it make sense to do it that way?
 
but it does matter because when i think of DX9 gaming, its associated with XP...really it's not trying to find out which OS is better...i'm not running Vista to run DX9, i'm running XP (the "I" in this is your average gamer) so wouldn't it make sense to do it that way?

Maybe some of the [H] readers can post some stuff up showing it.
I saw a few posts back that someone mentioned doing this.
 
but it does matter because when i think of DX9 gaming, its associated with XP...really it's not trying to find out which OS is better...i'm not running Vista to run DX9, i'm running XP (the "I" in this is your average gamer) so wouldn't it make sense to do it that way?
No, not in this article! This article is about [H]'s testing procedures, not Vista vs XP gaming.
And what about all te hdx9 games released that I play in Vista?

Trust me and all the others posting here dude.
The performance increase in winXP is alot more than 2-3fps :p
It could be a 25% difference and I could care less. That's not whatthe article is about.

Again, if you guys want a DX9XP vs DX10 Vista article, ok, ask for that, but it fdoesn't belong in this article.

Again, i'll quote from the article:
"Lately, we have been receiving some reader feedback asking why we are running Crysis in DX10 when the DX9 path produces better performance, with the same image quality."
 
So basically this entire article sums up what we already know, don't waste your time playing DX9 in Vista. What a revelation. I'm sure the 2 people on the planet using DX9 in Vista appreciate it.

As for the rest of us, can we get a useful DX9 vs. DX10 article that uses XP vs. Vista? That'd be better.
 
Ok, Im at work right now and wont be off till midnight EST. When I get home, Im going to bench Crysis on XP running the custom config that enables most of the DX 10 settings to be enabled. Unless anyone has any objections, Ill take screenies of the PC specs in N-tune and then screeny Crysis with Fraps or something running. If anyone has any other suggestions as to how to better show fps while doing this, please let me know. I want this to be strictly honest so no one can cry foul when I post the results.
 
I'd quite like to see XP DX9 vs Vista DX9 and Vista DX10 article.
I did find this article to be useful though and well done to [H] for making the effort to do it.

Its not just an afterthought to make an article like that, lots of preparation and very hard work involved and then making sure that its clean and clear enough so your readership doesnt rape you in the forums.
Despite all that, many have still called the article crap and that has to be disheartening as it isnt crap, Its just not for those wise asses that cant ask for more to be included without putting [H] down.
I'd like to see how much it costs and how long it takes them to do the article themselves.
 
So basically this entire article sums up what we already know, don't waste your time playing DX9 in Vista. What a revelation. I'm sure the 2 people on the planet using DX9 in Vista appreciate it.

As for the rest of us, can we get a useful DX9 vs. DX10 article that uses XP vs. Vista? That'd be better.

QFT

and modred...HOW MANY TIMES DO I HAVE TO SAY TO YOU IT'S NOT A BATTLE OF THE OS's....XP = DX9 VISTA = DX10....so YES this article....at least TACK IT ON

how can you prove your testing methods in using either DX9 or 10 from this one game anyways? at least put dx9 with xp in as well....can you not at least agree with that?
 
Decent read, I guess, but it really didn't tell me much. I agree with others, it wasn't an extensive enough test to make any real conclusions. Sure it may have fulfilled the premise of the article itself, but clearly it wasn't what the vast majority of users were looking for. If we all had Vista and compatible hardware we could do the benchmarks ourselves and it would be a lot more fun. Personally, I usually read benchmarks of hardware/software which I do not own, as to make an informed purchase. If I had owned the OS and the game, I could just play it! Seriously, think about it for a second.

But the real issue here is that Microsoft thought they could sucker everyone into Vista by making dx10 not run on XP, as well as devs like Crytek supporting this artificial limitation with locked game settings. That and the fact that the industry is going backwards. New OSes which are slower than the old ones; New video cards slower than the old ones; Hell, even LCD refresh rates can't match CRT quality. Something is just wrong here. Thats the problem, not the article itself. In a perfect world we wouldn't need to benchmark this stuff. The next-generation should *by default* by faster and better then the old generations. Thats the whole point.
 
Ok, as promised, here are my results.

Here you see what my system specs are in the n-view utility. Notice the only OC I have is a slight one on the CPU. The GPU's and vid mem are both running at stock speeds as are the streaming processors.



CrysisPCspecsforH.jpg




Here we have the two in game screens that pertain to the graphic settings. Notice the AA in the first pic. In the second, all the settings say 'Custom'. For anyone who is unfamiliar with this, these are the config hacks that force all the DX-10 bells and whistles on DX-9. A premade config file is available here:
http://www.moddb.com/games/183/crysis/addons/9312/crysis-dx10-graphic-features-under-dx9-minimod
and ore info here:
http://ve3d.ign.com/articles/news/35301/Crysis-Very-High-Settings-in-DX9-Improved


First settings screen

Crysisspecs2forH.jpg



Last settings screen

CrysisgamespecsforH.jpg






So that is the rig Im running and the specs the game is running at.

For the first shot I wanted to get a steady base reading so I found a good spot on the side of a mountain in some trees that overlooked the bay.

CrysisFPS1forH.jpg


FRAPS is in the bottom right hand corner....

Next I wanted to start to add more geometry to the mix.


CrysisFPS2forH.jpg



Again, FRAPS is in the bottom right hand corner.



After this I decided to load up one of the first ice levels where you have to constantly find a heat source. I thought this would make for a good tax on the cards due to the distortion of the containment (?) field around the perimeter.


CrysisFPS3forH.jpg



The next two pics are to show exactly just how much of the DX10 lighting and shaders DX9 can really do.....

CrysisFPS4forH.jpg


I love this one....

CrysisFPS5forH.jpg



These last one is to show more of how the game performs with more structural detail.

CrysisFPS6forH.jpg





Can we go back and look at those FPS again?

pic 1: 51

Pic 2: 47

Pic 3: 44

Pic 4: 37

Pic 5: 49

Pic 6: 41



You would have to have IBM's newest version oc Deep Blue along with an 8 way SLI rig set up to run these settings at these FPS on Vista....Ok, maybe not quite that bad, but you get the idea.
 
Not really surprising to be perfectly blunt. Most people who know anything about Crysis know that most of the benefits from the DX10 path is to enable the very high settings in the menu, since you did not compare the very high settings to anything below that there can be no real genuine comparison of the detail present in DX10.

If any point should be made here it should be that we need better hardware to push these DX10 effects, Nvidia need to get off their ass and push out the door a card thats twice as fast as the 8800GTX like when the 8800GTX beat the 7xxx range by 2x the speed. Quite frankly the 9800GTX is a joke being no faster than the 8800's

I can't help but feel that the article trys to make the point that "in real world gameplay testing we see no benefits of DX10, therefore theres no benefits to DX10" that if we can't use it in "real world" test, it doesn't exist at all.

I know a few people who have done what I did when playing Crysis, and thats to dump the resolution down a few notches, and then sacrafice shader and shadow details, or other settings, to enable them the excess power to turn up some of the other settings to very high (where DX10 effects maybe seen)

I think this kinda comes back to bite you guys and your testing methadology again, where you have this strict "these settings are best" opinion of the game when in fact lots of users are going to use completely different configurations to you, some will no doubt extend some of the graphical quality settings to "very high".

In fact if you'd spent any real time reading your games section in your forum you'd see many different crysis configs posted which are hand tweaked to enable the higher end effects by disabling a lot of the other less noticeable effects.
 
and modred...HOW MANY TIMES DO I HAVE TO SAY TO YOU IT'S NOT A BATTLE OF THE OS's....XP = DX9 VISTA = DX10....so YES this article....at least TACK IT ON
Dude, lol, you just say it's not about the os, but you bring the Os up in your sentence, above. Make up your mind.
how can you prove your testing methods in using either DX9 or 10 from this one game anyways?
because their testing bed is a Vista system, and since this article addressed questions about their testing methods, why would you talk about XP? You might as well include Linux DX9 in Wine.


Don't get me wrong, I would like to see a DX9 XP, DX9 Vista and DX10 vista comparison. It would be very enlightening for many XP users.
But again, that has NOTHING to do with this article, and is an article all of its own.
 
I love how all Vista users do/say anything to avoid comparing XP DX9 preformance vs Vista DX9/D10 preformance...but perhaps the numbers would be to painfull?

If we look at Steam, this paints this picture:
http://www.steampowered.com/status/survey.html


Windows XP 81.39 %
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png

Windows Vista 14.57 %
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png

Windows Vista 64 bit 2.55 %
bit.png
bit.png
bit.png

Windows 2003 64 bit 0.72 %
bit.png

Windows 2000 0.62 %
bit.png

Other 0.15 %

No matter how you try and hide it, XP is STILL king...by a long shot...actually this shows that Vista 64 is overhyped...gamers use XP, not Vista..and NOT Vista 64 bit!
 
I love how all Vista users do/say anything to avoid comparing XP DX9 preformance vs Vista DX9/D10 preformance...but perhaps the numbers would be to painfull?

If we look at Steam, this paints this picture:
http://www.steampowered.com/status/survey.html


Windows XP 81.39 %
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png

Windows Vista 14.57 %
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png
bit.png

Windows Vista 64 bit 2.55 %
bit.png
bit.png
bit.png

Windows 2003 64 bit 0.72 %
bit.png

Windows 2000 0.62 %
bit.png

Other 0.15 %

No matter how you try and hide it, XP is STILL king...by a long shot...actually this shows that Vista 64 is overhyped...gamers use XP, not Vista..and NOT Vista 64 bit!
I would say it is under-adopted, not overhyped.
But good point. And it validates it's own article to compare them.
 
Windows Xp 64 bit too many people on the ass and most enthousiests assume Vista 64 will have the same compatability problems. There is a VERY low level of consumer education about Vista 64. The average consumer doesnt know or care about it either.

Vista should have just been 64 from the start in all flavors.
 
Windows Xp 64 bit too many people on the ass and most enthousiests assume Vista 64 will have the same compatability problems. There is a VERY low level of consumer education about Vista 64. The average consumer doesnt know or care about it either.

Vista should have just been 64 from the start in all flavors.
Do they even make 32bit cpus any more?
 
Back
Top