![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
id Software's Official DOOM3 Benchmarks
- Thread starter archevilangel
- Start date
as has been pointed out several times in just this thread alone:Tremek said:
http://www.driverheaven.net/reviews/6800andX800XT/test setup.htm
"*NOTE: All optimisations were disabled using the Nvidia drivers; this included tri-linear optimisations. With the ATI drivers it is not possible to disable some optimisations and therefore they are enabled. Whilst this does not provide a true apples to apples comparison between cards it is our opinion that every review product should be tested at the best quality it can provide. It is up to the end user when they have the product to decide on optimisations which suit them. If ATI add an option to their drivers in future revisions we would also choose to disable the optimisations. Therefore it can be assumed that the performance on the 6800 Ultra can be improved over that shown in this review should you require it however this would mean reducing IQ. We suggest you read up on the optimisations which each manufacturer uses before deciding on your final purchase. "
cliffs notes for you:
All optimizations were turned off in the display properties for the 6800, all cheats were left on for the x800.
Hyper_Psycho
2[H]4U
- Joined
- Mar 26, 2002
- Messages
- 2,944
am i the only one that does not care about doom 3 ?
C
CIWS
Guest
Hyper_Psycho said:
Welcome to the discussion. If you had read even a portion of the thread you will have seen a couple of points.
First, if you look even a few posts back you will see the Webmaster is about to have a lot more data to release on how Doom 3 runs on various systems and video cards.
Second, This is a thread specifically about Doom 3 and the benchmarks ID Software did. If you don't give a damn about Doom 3 then why post in a thread about it ?
Finally, Possible driver upgrades are mentioned throughout the thread, but then again ATI has had plenty of time to tweak or develop drivers sets for Doom 3 so one has to wonder what they have been doing all of this time.
Does anyone know exactly what Carmack is talking about in the "Doom 3 Benchmark" article when he states -
On the other hand, the Nvidia drivers have been tuned for Doom's primary light/surface interaction fragment program, and innocuous code changes can "fall off the fast path" and cause significant performance impacts, especially on NV30 class cards.
What is the deal with the NV30 cards? Is this referring to the NV30 cards poor
shader performance?
I am curious as to how the 5700 non-ultra performs against the 9600PRO or 9800SE.
Thanks for any input.
On the other hand, the Nvidia drivers have been tuned for Doom's primary light/surface interaction fragment program, and innocuous code changes can "fall off the fast path" and cause significant performance impacts, especially on NV30 class cards.
What is the deal with the NV30 cards? Is this referring to the NV30 cards poor
shader performance?
I am curious as to how the 5700 non-ultra performs against the 9600PRO or 9800SE.
Thanks for any input.
burningrave101
[H]F Junkie
- Joined
- Sep 9, 2003
- Messages
- 11,825
Tremek said:
Maybe because your comparing to nothing but D3D games lol. There are a ton of OpenGL games and the 6800u is faster in all of them at every setting.
D3D is tied for the performance crown. And everyone stop freaking saying AA AND AF!!! The 6800 has higher performance AA!!! Just look at the benches where only AA is enabled.
The X800's are getting their performance from their Brilinear optimizations. Especially their AF optimizations. Turn them off and the cards will crawl
.They deem it "fair" to compare the 6800 with no optimizations to the X800 which is FULL of optimizations lol.
The 6800 has the same level of IQ as the X800's at just Quality setting. At High Quality setting, which only the 6800's support, they have Full Trilinear which is higher IQ then anything ATI supports.
kcthebrewer
[H]ard|Gawd
- Joined
- Jun 14, 2004
- Messages
- 1,728
nv30 series had a problem running the ARB2 (i believe this is the name) code path. There was a partial precision code path being developed for it, but nvidia's drivers fixed the problem running full precision, so the partial precision code path was dropped.
burningrave101
[H]F Junkie
- Joined
- Sep 9, 2003
- Messages
- 11,825
@trapine said:You are full of FUD burningave
Please show me your proof of The X800 series crawling when optimizations are turned off?As has already been explained to death the X800 optimizations are no where near as agressive as what NV30 optimizations were!
Yea right lol. The R300's were using the same Brilinear filtering tricks that the X800's are using now. ATI just does a better job of hideing them.
Look at ATI's AF performance. They loose almost no performance at all from 2xAF - 16xAF. This is most likely the cause of Brilinear AF optimizations.
People have also made mention of the X800's being benched in Far Cry and their optimizations turned off. They lost more then 30 fps. The person that saw the link is going to shoot it over to me when they find it again.
The 6800's lose almost no performance from their Tri and AF ops off. Their AF ops come OFF by DEFAULT and with both off in Far Cry at 1600x1200 w/ AA + AF, the 6800u loses around 5 fps.
And i've seen other ATI users make comment about ATI's AF optimizations and how they would take a severe hit if they were turned off.
If this wasn't true then why doesn't ATI give the option like nVidia? Huh? Huh?
So until you can provide links showing they lose very little performance with their Optimizations OFF, i dont need any proof that they do. The simple fact that ATI doesn't give the option is proof enough
.@trapine said:You are full of FUD burningave
Please show me your proof of The X800 series crawling when optimizations are turned off?As has already been explained to death the X800 optimizations are no where near as agressive as what NV30 optimizations were!
NV30 sucked, so you're saying that X800 sucks too, just not as much?
Lord of Shadows
2[H]4U
- Joined
- Mar 9, 2004
- Messages
- 3,322
Is it just me or do you think the doomguy should keep his helmet on =)
and it was corny when he pulled a nade on the fatzombie... that could have been a great scene.
(/me if way off topic)
and it was corny when he pulled a nade on the fatzombie... that could have been a great scene.
(/me if way off topic)
- Joined
- May 18, 1997
- Messages
- 58,620
I think if you read this review here, you will see that we are finding the GT is stronger in many games just besides DOOM3. I think we picked the GT over the Pro before all this DOOM3 stuff started publicly and we are sticking to our guns.Tremek said:
"We compare it to a 6800Ultra and Radeon X800Pro playing FarCry, FlightSim 2004, PainKiller, Nascar Thunder 2004, City of Heroes, and Madden 2004. " Looks to be a strong card in a little more than DOOM3....
Sniper_Merc
[H]ard|Gawd
- Joined
- Mar 25, 2004
- Messages
- 1,037
^eMpTy^ said:
They shut off the trilinear and ansi optimizations on the 6800 but couldn't on the x800.
Bustercaps
n00b
- Joined
- Jan 20, 2004
- Messages
- 63
Mike89 said:
I Agree Mike..... One game? So everyone is going to go out and buy a new $400+ Card for a new $50 game... I dont think so.
Lord of Shadows
2[H]4U
- Joined
- Mar 9, 2004
- Messages
- 3,322
Bustercaps said:
I bought a $360 dollar gpu (last season doh!) and am buying a new mb/cpu this season for this so called "one game" But I have the added benefit of everything working for other games. =)
Bustercaps said:
Being that the price has spiked over $60 at Comp-U-Plus since the [H] review of the GT and the Doom 3 benches , I would say that is exactly what is happening.
Sniper_Merc
[H]ard|Gawd
- Joined
- Mar 25, 2004
- Messages
- 1,037
^eMpTy^ said:
Thats what ticked off a lot of people when both the ATI and Nvidia new gen cards were released, ATI insisted on turning off the Nvidia optimizations for benchmarking and then hid the fact that they also had them, it was only revealed that ATI indeed did have optimizations in place after somebody noticed a differance in framerates when colored mipmaps were turned on and off while in game.
The other thing that ticked peeps off is when not only was there a differance in framerates with and without colored mipmaps, is that one of the programs that uses test paterns showed no differance between full trilinear and the ATI optimizations because the optimizations completely shut off when the program was run.
At any rate, thats the short tour of the optimization game. When there turned on in both cards the gap narrows between the two cards in D3D based games, ATI wins some Nvidia wins some, it's when you start seeing Nvidias dominance in OGL that things really start to look like Nvidia has really pulled out the stops this round. (and to keep it on topic, Doom 3 is OGL so there
)I went Nvidia, I've had too many people that I know personally that have had to RMA there ATI cards 6 mo to a year after I suggested that they get them that I will not suggest that they buy ATI again.
Sniper_Merc
[H]ard|Gawd
- Joined
- Mar 25, 2004
- Messages
- 1,037
Grinder66 said:
Doesn't suprise me, hell at this point the GT and the Pro are the only thing's shipping in quanity, with the GT appearing to be more available. Those Doom 3 benches are going to put a strain on that even.
- Joined
- May 18, 1997
- Messages
- 58,620
I just would not know what to tell you about that. I find it interesting that they disable ops on one card and not the other.^eMpTy^ said:
"NOTE: All optimisations were disabled using the Nvidia drivers; this included tri-linear optimisations. With the ATI drivers it is not possible to disable some optimisations and therefore they are enabled. Whilst this does not provide a true apples to apples comparison between cards it is our opinion that every review product should be tested at the best quality it can provide. It is up to the end user when they have the product to decide on optimisations which suit them. If ATI add an option to their drivers in future revisions we would also choose to disable the optimisations. Therefore it can be assumed that the performance on the 6800 Ultra can be improved over that shown in this review should you require it however this would mean reducing IQ. We suggest you read up on the optimisations which each manufacturer uses before deciding on your final purchase. "
The optimizations are there for a reason and unless you can show me how they impact gameplay experience, I don't see turning them off.
I will say this as well, we stopped using standard benchmarks as they were not showing a real world gaming experience, so I find it totally plausible that our data does not show what others show many times. I am sticking with our results though as I feel as though our way of testing truly shows the gaming experience to be expected at home....
el_lance_o
Limp Gawd
- Joined
- Jul 9, 2004
- Messages
- 235
Face it, most people here are gonna get doom 3. I bet almost all of us are, and then we'll see the true performance we get IN GAME than just the new vid cards out, cause most tennage gamers like me don't have an A-bomb to play it on( even though mine's been worked on till it's almost like that, hehe). Oh yeah, we sould make a post for peoples game play smoothness for it once it comes out.
- Joined
- May 18, 1997
- Messages
- 58,620
Also we tested the $299 BFGTech 6800 and it is a beast....
- Joined
- May 18, 1997
- Messages
- 58,620
Rome was not built in a day. From our conclusion page.el_lance_o said:
"Certainly we will have much more information on this in the coming weeks with the DOOM 3 [H]ardware Guide. At that time we will take a look at a much wider array of CPUs, Video Cards, and Motherboards and what overall gaming experience they can deliver."
Also as mentioned earlier in this thread, we will focusing on playability alone in the DOOM 3 [H]ardware Guide.
Also worth noting, this will be done hopefully before the game goes out for sale. So you should have a guide that covers from 8500/GF3 with Athlon 2500/P4 2.4 to currently released hardware. Also we cover the min-spec machine (P41.5/GF4MX440/512MB Ram) very well too. I see where a bunch of folks are freaking out over Ram as the Dell boxes had a LOT in them. 384MB is mine spec and it seems to be realistic, so 512MB in most newer boxes will rock the house.
Jesus you guys need to settle down. This was JUST ONE benchmark with one system setup and from what I understood it was just a flyby demo. If you will remember both the X800 series and the 6800 series cards preform different on different CPU's. The ATI cards seem to preform better on higher end CPU's. While the Nvidia cards are better on lower end CPU's. The benchmark Hardocp used was just on a pentium CPU. The AMD FX-53 CPU is simply the best CPU out right now and I'm going to wait and see how these cards preform on this CPU since that's what I have. That's exactly what you guys need to do also, just wait and see how these cards preform on the same CPU YOU have. All the reviews that I have seen that uses an AMD64 System the ATI cards usually win. When they use Pentium CPU's the Nvidia cards usually win. I think hard ocp actually had a review on this once. If I remember correctly guru3d had a VERY good review on it.
EDIT: Here is the link to the review I was talking about on guru3d. It shows the difference between different CPU's with ATI and Nvidia cards. http://www.guru3d.com/article/article/136/1/
EDIT: Here is the link to the review I was talking about on guru3d. It shows the difference between different CPU's with ATI and Nvidia cards. http://www.guru3d.com/article/article/136/1/
- Joined
- May 18, 1997
- Messages
- 58,620
The DEMO was an ingame recording of actual gameplay. There is no "flyby" mode in DOOM 3 like UT2K4 so they are a bit different. Neither touch the CPU for AI or Physics though, so on that front they are comparable.Vezra said:
While CPU does impact performance, the driving force behind DOOM 3 performance is the video card. DOOM 3 is very forgiving in terms of Ram, CPU, and IO and will play GREAT on a wide array of machines, but for top end high res graphics, you are going to have to have a 6800 series card or a XT-PE.
We did one setup for the OCers out there. Our 754-3000 OCed to 2.45GHz with an OCed 6800. Sweetness.......
- Joined
- May 18, 1997
- Messages
- 58,620
Here is a link to a CPU Scaling article here at HardOCP done with ATI and NVIDIA flagship cards published the day before.Vezra said:
Corrugated
Limp Gawd
- Joined
- Apr 9, 2000
- Messages
- 260
Kyle, Do you know if the Doom3 engine supports SMP?
Thanks
Thanks
- Joined
- May 18, 1997
- Messages
- 58,620
I did not get the chance to ask that question directly to John Carmack, but I did ask and the answer I got from other staffers was "no".Corrugated said:
kcthebrewer
[H]ard|Gawd
- Joined
- Jun 14, 2004
- Messages
- 1,728
Kyle, any word on MP server requirements (recommendations)?
- Joined
- May 18, 1997
- Messages
- 58,620
GF3, TI4600, 5800, 5900, 6800, 680GT, 6800U are all coming this week. As well as 8500, 9600XT, 9800Pro, X800Pro, X800XT. There are probably a couple I missed in there as well._leech_ said:Too bad [H] didn't include any 6800 NU scores
Here are the systems and CPUs we used the cards on. (Not all cards on all systems, but there is some overlap.)
well i wanna see which high end card performs best as well, i kinda hope ATI can keep it close on my system because i like gaiming on my 65inch Widescreen hdtv and id rather pull out the X800Pro and drop in the XT-PE cause im lazy and dont want to set everything backup lol.