Spring IDF 2006 Conroe Preview: Intel Regains the Performance Crown
by Anand Lal Shimpi on March 7, 2006 3:58 PM EST- Posted in
- Trade Shows
Gaming Performance
The one area where AMD has been the clear leader for years has been in gaming performance - Conroe changes everything.
Updated: In Don MacDonald's keynote he also provided us with another reference point for Conroe's performance, this time under Call of Duty 2. We have no idea what settings they ran at but the results we saw were Conroe at 111 fps and a Pentium Extreme Edition 3.73GHz scoring 90 fps. But the most interesting gaming tests are below:
First off we've got Quake 4 running the 1.0.5.0 patch at 1280 x 1024 with High Quality settings. The only demo available was Intel's own demo but nothing looked out of the ordinary with the recording. We tested with both r_useSMP enabled and disabled, first the SMP disabled numbers. Updated: The Quake 4 scores have been updated as mentioned in our follow-up article.
With SMP disabled, Conroe holds a 25% performance advantage over the 2.8GHz Athlon 64 X2. Enabling SMP provides a similar 24% performance advantage.
Next up is a Half Life 2 Lost Coast demo, once more an Intel supplied demo but there's only so much you can do to a demo recording to make it favor one CPU maker over another:
Conroe's performance advantage extends to 31% under Half Life 2, talk about a complete role reversal here.
We finish off this page with Unreal Tournament 2004 and a 20% performance advantage for Conroe.
Facebook
Twitter
RSS
220 Comments
View All Comments
JumpingJack - Tuesday, March 7, 2006 - link
Conroe will be out in the first week in July. Merom will be out end of Aug/start of Sept. with Woodcrest to follow shortly after.kalaap - Tuesday, March 7, 2006 - link
Well with Intel picking up Apple as another customer, they probably need a little more time to ramp up volume shipments.JackPack - Tuesday, March 7, 2006 - link
Apple is small potatoes. They sell around 1.25 million Macs per quarter.Intel manufactures around 40 million processors in the same period.
Questar - Tuesday, March 7, 2006 - link
They may, remember Yonah shipped six months early.Intel also likes to have lots of product in the pipeline when something is launched. All the board people and system makers have to be ready. That's why you can buy a Dell ot HP system on the day Intel launches a chip.
I'll bet you that right now Conroe in being manufactuered on production lines.
Doormat - Tuesday, March 7, 2006 - link
Well thats what I'm thinking - I kinda expect Merom and Conroe in time for WWDC in early August (pushed back from its usual July timeframe). The fact that it was pushed back a month makes me a little suspicious on that Apple is waiting for Intel to annouce and ship the parts.Justin Case - Tuesday, March 7, 2006 - link
At the resolution and settings level they tested most games (ex., FEAR), the game is GPU-limited, not CPU-limited. In other words, if the graphics card, chipset and driver really were matched, the game would perform at exactly the same speed, and the difference would be in the CPU load (a faster CPU would show a lower load, because it would spend more time waiting for the GPU to finish rendering each frame). I have a feeling the drivers had some strategic "tweaks"...Also, bear in mind that a 20% performance increase is to be _expected_ from a product to be released four months from now (in other words, it's simply following Moore's law).
But anyway, if you trust benchmark results coming from Intel (or AMD), you might as well believe Apple's marketing. I'll wait for the real product, and independent testing, thank you very much. I still remember Intel's claims about the original Pentium 4 (3 times faster than the Pentium III, they said) and Prescott (a.k.a. Pres-hot!), not to mention the Itanium, the Paxville Xeons, etc., etc., etc....
DarthPierce - Tuesday, March 7, 2006 - link
Your inability to grasp what being GPU limited means or when it occurs astounds me. an example of GPU limited would be you're running FEAR at 1600x1200 on an ati 9600. In this example, any cpu (that doesn't suck horribly) would give roughly equally bad scores.If the CPU is changed and the score changes, that itself shows that a task is not GPU limited.
Think!
Anand Lal Shimpi - Tuesday, March 7, 2006 - link
We were limited to 1280 x 1024 because of the LCDs we were testing on, but a dual X1900 XT setup isn't going to be GPU limited at 1280 x 1024.As I posted above:
"As far as I could tell, there was nothing fishy going on with the benchmarks or the install. Both systems were clean and used the latest versions of all of the drivers (the ATI graphics driver was modified to recognize the Conroe CPU but that driver was loaded on both AMD and Intel systems).
Intel told us to expect an average performance advantage of around 20% across all benchmarks, some will obviously be higher and some will be lower. Honestly it doesn't make sense for Intel to rig anything here since we'll be able to test it ourselves in a handful of months. I won't say it's impossible as anything can happen, but I couldn't find anything suspicious about the setups. "
Take care,
Anand
Questar - Tuesday, March 7, 2006 - link
Use your head, if it was GPU limited, there would not be a performance increase between the systems with a different cpu. The vid card would cap the frame rates.BTW, do you really think a x1900 crossfire system is GPU limited at 1280x1024?
AndreasM - Wednesday, March 8, 2006 - link
You should read his post more carefully. What he is saying is, that because F.E.A.R. is GPU-limited these results cannot be correct, because there should be no performance increase. I do however agree that it probably isn't GPU-limited at that resolution.