Rage3D take a look at Assassin's Creed D3D10.1 support

Thursday 17th April 2008, 11:54:00 PM, written by Rys

Rage3D have taken a look at the PC version of Assassin's Creed, a game I've been into for a long time on 360, to check out what's new when you run the game under Vista SP1.  You'll remember that SP1 brings D3D10.1 support.  Alex checks to see if the two are related somehow.

The answer is a resounding yes, with both performance and image quality improvements to be had if you have a D3D10.1 GPU (R6-class ATI only) and the Microsoft operating system that everyone loves to hate, along with its latest collection of fixes, improvements and additional features.

You can find the full article over at Rage3D, where they take a look at HD 3870 X2 performance, Quad Crossfire and the good old HD 2900 XT (a card I have a serious soft spot for, even if nobody else ever did).

http://www.rage3d.com/articles/assassinscreed/


Discuss on the forums

Tagging

amd ± radeon, d3d10.1, rage3d, morgoth, rocks


Latest Thread Comments (116 total)
Posted by Jawed on Sunday, 11-May-08 21:55:44 UTC
Tech Report has updated the article: http://www.techreport.com/discussions.x/14707 Don't know why they bothered, but there it is. Jawed

Posted by Silent_Buddha on Sunday, 11-May-08 23:09:28 UTC
It's hard not to buy into the conspiracy theory that UBIsoft will hold back a DX10.1 patch until Nvidia can get a DX10.1 class product on the market (if they ever do).

I'm sure ATI is more than willing to help them with any efforts with regards to "fixing" the DX10.1 path, especially considering it makes them much more competitive with the Nvidia offerings.

An interesting developement of this whole broohaha. That article at TGDaily (I think that's the one) mentions several developers praising DX10.1 codepaths in their upcoming games and commitment to shipping with DX10.1.

EDIT - had to go find it. http://www.tgdaily.com/content/view/37326/98/

Quote
“Of course it removes the render pass! That's what 10.1 does! Why is no one pointing this out, that's the correct way to implement it and is why we will implement 10.1. The same effects in 10.1 take 1 pass whereas in 10 it takes 2 passes.”
Now if only we knew who that developer was. :)

Regards,
SB

Posted by Demirug on Monday, 12-May-08 00:15:06 UTC
Quoting Silent_Buddha
Now if only we knew who that developer was. :)
It wasn’t me.

The statement isn’t the full truth either. You can eliminate this second pass with D3D10 (even with D3D9), too. But as you need to use an additional render target for this job this would require additional work on your shader code. The Direct3D 10.1 solution for this problem is much easier to implement as you just make a copy of the depth buffer at the right point.

Depending on your overdraw factor it could be a good Idea to keep this additional pass (but with MSAA active) to fill the depth buffer for Early-Z when it comes to the expensive shader rendering.

Posted by silent_guy on Monday, 12-May-08 01:28:33 UTC
Quoting Silent_Buddha
It's hard not to buy into the conspiracy theory that UBIsoft will hold back a DX10.1 patch until Nvidia can get a DX10.1 class product on the market (if they ever do).
Conspiracy theories are definitely more fun, but why is it so hard to believe that both companies are telling the truth?

Here's a different theory: AC has been completed. The majority of the players who will buy the game have already done so. The A-team at Ubisoft has moved on to newer projects. The B-team stays behind to fix up urgent bugs and keeps the game in maintenance mode. (This is a common mode of operation in a lot of companies, BTW. It's a great way to let junior programmers/engineers gain experience before graduating to the A-team.) All additional funds spent won't result in meaningful additional sales and are basically a net loss, so effort is being kept to the bare minimum and rather than spending time fixing bugs in two render paths, they just limit themselves to the one that works for both brands.

Definitely more boring and straightforward than the conspiracy angle, which is exactly why it's way more likely...

It's one thing to risk your reputation for the gold standard of GPU benchmarks, it's a whole other thing to do so for AC (solid, but not exactly a blockbuster) after it has already been released, especially if the competition is still not blowing away your equivalent cards despite the 20% performance increase.

Posted by cho on Monday, 12-May-08 02:00:03 UTC
http://www.pcinlife.com/article_photo/9800gx2/screenshot/3870x2/_no_rt_assassinscreed_dx10_1680x1050_4aa_by_game_16af_by_driver_27fps.png maybe this is a example that "a problem that caused the post-effects to fail to render properly" ? (btw, i had found out why the fps is so slow on Radeon HD 3870X2, that was caused by the Rivatuner'S OSD)

Posted by Arnold Beckenbauer on Thursday, 15-May-08 11:06:46 UTC
http://www.fudzilla.com/index.php?option=com_content&task=view&id=7355&Itemid=1
Quote
Today, Ubisoft launched a new patch for the PC version of Assasin’s Creed. The patch will fix a couple of bugs in the game, among which the bug where Alt-Tab key combination would crash the game. The patch will also introduce post-effects improvements for cards with DX10.1 support and add support for x64 versions of Windows.

You can download the patch here and the complete list of fixes is as follows:

...

- Fixed broken post-effects on DirectX 10.1 enabled cards

Posted by Morgoth the Dark Enemy on Thursday, 15-May-08 12:05:25 UTC
Quoting Arnold Beckenbauer
http://www.fudzilla.com/index.php?option=com_content&task=view&id=7355&Itemid=1
It doesn't fix anything with the 10.1 path, it simply removes it.

Posted by Arnold Beckenbauer on Thursday, 15-May-08 12:18:39 UTC
Quoting Morgoth the Dark Enemy
It doesn't fix anything with the 10.1 path, it simply removes it.
Unfortunately you are right.
http://www.pcgameshardware.de/aid,643633/Test/Benchmark/Assassins_Creed_Patch_102_im_PCGH-Benchmark-Test_-_Stellungnahme_von_Nvidia/
http://www.pcinlife.com/article/graphics/2008-04-02/1207107328d527_4.html

Great work...
But the broken PP effects are fixed.

Posted by Morgoth the Dark Enemy on Thursday, 15-May-08 12:58:52 UTC
Quoting Arnold Beckenbauer
Unfortunally you are right.
http://www.pcgameshardware.de/aid,643633/Test/Benchmark/Assassins_Creed_Patch_102_im_PCGH-Benchmark-Test_-_Stellungnahme_von_Nvidia/
http://www.pcinlife.com/article/graphics/2008-04-02/1207107328d527_4.html

Great work...
But the broken PP effects are fixed.
They're not fixed as they were never broken through the DX10 path to begin with. It's like saying:dude, I just chopped off your arm because you had a nasty tatoo. But the tatoo is gone.

They also didn't fix the AA quality in DX10, which is inferior to both 10.1 and DX9. And there are reports on UBis boards that the same bugs from before the patch are there, so they didn't seem to fix/improve much if anything with this particular update, they just removed 10.1 for whatever reason they chose to.

Posted by Wesker on Thursday, 15-May-08 15:49:02 UTC
Quoting Arnold Beckenbauer

http://www.fudzilla.com/index.php?option=com_content&task=view&id=7355&Itemid=1...

You can download the patch here and the complete list of fixes is as follows:

...

*- Fixed the performance improvements on ATI DX10.1 capable cards.*
Nice! That performance boost was really bugging me. :smile:


Add your comment in the forums

Related amd News

AMD Phenom X3 released; reviewed
Stanford University release Folding@Home client for R6-family ATI GPUs
Official: AMD layoffs 10%; misses Q1 guidance.
AMD release FireGL V7700 with DisplayPort support
AMD release new Phenom X4 processors with B3 silicon
AMD RV670 price cuts & 128-bit Radeon HD3830?
Breaking: AMD to layoff 5% of its workforce, says The Inq
AMD Shangai die size disclosed; Fudzilla hints at 0MiB L3 chip
AMD unveils embedded products based on RV610/690G/Sparta
AMD beats estimates and nears break-even