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
MrKaz - Thursday, March 9, 2006 - link
Why Ati drivers where modified to recognize the Conroe processor?What was the problem if the processor didn’t get recognized?
I mean if the FX60 could say Unknown Processor and have no problems, (my mobile 2600+ says the same and runs OK).
What problem would have come from the drivers by not recognizing the processor?
Some drivers optimizations there? Like NVIDIA/ATI?
rodrigorras - Wednesday, March 8, 2006 - link
The real big question here is why Intel is showing your own benchmark...In My opinion if conroe is a mega boost from P4 and AMD64, why intel did not gave it to anand and let anand configure it and benchmarked on your own way.
(optimized drivers for Conroe...it's just not fair comparison, next time try with AMD64 optimized drivers)
Think about...Theses numbers are just fake...if they were true imagine how many sites would been receving engeneering samples...!!!
Always remenber that Intel is a marketing Fud Master...
Bye!
Rodrigo Rios
rodrigorras - Wednesday, March 8, 2006 - link
Of course remenber that new Games even with low resolution the bottleneck is videocard and not CPU...So do not expect 40% boost of frames only with a cpu upgrade...
let's be logical
ProperMethodz - Thursday, March 9, 2006 - link
http://www.7hmgaming.com/forum/showthread.php?t=37...">http://www.7hmgaming.com/forum/showthread.php?t=37...JarredWalton - Thursday, March 9, 2006 - link
"For instance, Tomshardware ran a benchmark on the FX-57 where it was pulling 183.4 FPS in UT2k4 on an NVIDIA Geforce 6800 GT and the AnandTech review the FX-60 was only pulling 160 FPS. These numbers by Tom are almost a year old. This means some serious fishy shit is going on."Different cards, different systems, different maps. Some levels in HL2 score 80-100 FPS on high-end machines, and other levels score 160+ FPS. 160 FPS on one level means nothing other than that level isn't particularly complex.
vdhd - Wednesday, March 8, 2006 - link
Hi guys, I just read through all the comments and I would like to make a few of my own.1) I have no doubt Conroe will be a good processor when it eventually comes out.
2) This isn't about Anand vs Rahul, both of us will tell you that we need to see the platforms in house to give the real benchmarks. Perhaps passing the crown over to Intel is a bit premature at the moment (6 months premature), but we'll see how it plays out down the road.
3) Anand is one of the most respected journalists in our industry, and I hold him in the highest regard. Anyone suggesting that I am calling him out needs to read my blog again.
4) To the guy who is suggesting to sue Anandtech based on his writing - get a grip, please. I suspect AMD will post another mind blowing quarter, short term thinking gets you nowhere in this business.
5) Lastly, I made it clear on Dailytech as well - Voodoo sells both AMD and Intel. Around 80% of our notebooks go out with Intel processors. As far as workstations and desktops go, we'd be doing our customers a dis-service by making poor recommendations. In fact if we had to choose Intel for our desktops currently we prefer the Pentium M for over the Pentium 4.
6) I have also been accused of being biased towards Nvidia (many many times). If you've read my battles with ATi over the years you'd know that I am only telling it from the inside. When you are dealing with these companies it's completely different from what you all see on the outside.
I am biased towards great technology no doubt, that's why we were first in line to launch Crossfire on RD580 - I love it! I am also looking forward to Nvidia's upcoming launch, but that doesn't take away any positive feelings towards ATi at the moment.
Thanks, I hope that clears some things up.
Rahul Sood
http://www.rahulsood.com">www.rahulsood.comwww.rahulsood.com
MiLLeRBoY - Friday, March 10, 2006 - link
Rahul, you fool. I found your blog post from Tom's Hardware's forum. You posted detailed analysis of the BIOS issue, but then Anandtech posted updated results for the new BIOS and the results didn't change much. lol.I found your blog from here: http://forumz.tomshardware.com/hardware/funeral-AM...">http://forumz.tomshardware.com/hardware...MDftopic...
Here is a direct link to your post: http://voodoopc.blogspot.com/2006/03/if-only-they-...">http://voodoopc.blogspot.com/2006/03/if-only-they-...
Delete that blog post, lol. It makes you look ridiculous.
JarredWalton - Thursday, March 9, 2006 - link
Welcome, Rahul!Well, this is certainly a big can of worms that has been opened, but looking at it objectively, I would have to say that "faking" anything more than a 5-10% performance loss (on the AMD system) is going to be very difficult. While I would say there's potential for Intel to also fake a 5-10% performance boost for Conroe, that would be very unlikely at this point. If the RD480 chipset isn't the best chipset in the world, what about 975X running Conroe six months before launch.
I would say the biggest issue is with the early declaration of victory, but really we're talking about minor semantics. Intel will almost certainly regain the performance crown with Conroe; I'd be absolutely shocked if they didn't. Really, I'm a bit surprised it even took this long. Until Conroe launches, of course, they don't actually have the lead. :)
IDF is always about what Intel has planned for the future, and right now their future is looking good. What have they shown in the recent past? Super cooled CPUs running at 5 GHz that they didn't really benchmark much. For them to step out and demonstrate a Conroe 2.66 GHz chip against an *overclocked* FX-60 just OOZES confidence. Letting anyone (let alone Anand) play with the system and publish results? That's not typically what you do if you're trying to hide something. If they wanted to hide stuff, they'd be showing slides of performance numbers, and they'd have super-cooled Conroe chips running as fast as possible to come out on top. Pitting an air-cooled CPU against an FX-60 just doesn't smack of difficulties at all.
What's going to be far more difficult is selling us all on the need for quad-core processors come 2007. Dual-core? Sure, that's reasonable, since people will often do two tasks at the same time. Beyond two cores it gets difficult to see home-desktop uses for more processing power. Oh, sure, I'll be glad to see WMV9, DivX, etc. encoding speeds increase substantially, but I really don't do any of those tasks all *that* often. (And GPU acceleration can probably do it a lot faster anywat.)
Oh, one last question: six months? Where did that come from? That puts Conroe launch in September; I thought Conroe was coming in July, which would make it more like 4 months. Did Intel push Conroe back and I missed it? :(
Frumious1 - Wednesday, March 8, 2006 - link
This is in response to http://voodoopc.blogspot.com/2006/03/if-only-they-...">Rahul Sood's rantThe BIOS chip used is programmable, so 6.00PG is simply the revision of the chip being used. The question is *how* the chip was programmed, and the screen shots don't show. I'm not sure what Rahul and others are going on about, but I've got a DFI RDX200. It runs fine, even with the old BIOS. Where did he find out that "the DFI bios version 'D49C-32' they are running is from 10/11/05" - I don't see that posted anywhere, but maybe I missed it? Anyway, I'll be happy to try and run some benchmarks with the old and new DFI BIOS if I can manage to get $1000 worth of GPUs. (Damn, I knew there was some reason I couldn't generate similar results!)
There are only http://labs.anandtech.com/products.php?sfilter=400">five other CrossFire boards currently available, and arguably the DFI RDX200 is the most enthusiast friendly. http://labs.anandtech.com/products.php?sfilter=460">RD580 boards are starting to show up, but you can't reasonably blame Intel for not having an ASUS A8R32-MVP in the system, since it just showed up within the past three days.
So, the motherboard choice wasn't an intended slight. How about the BIOS version? I've got an Opteron 146 in my DFI board, overclocked to 2.8 GHz. I don't have X1900 cards to benchmark in it, but the differences between the 10/11 and 12/15 BIOS revisions was incredibly small. Hey - the CPU was at 2-2-2-5-1T timings instead of 2-2-2-8-1T, and perhaps 8 tRAS is better as well? The difference between minor memory tweaks is never more than a couple percent for the same timings.
An "unknown" CPU is also rarely a problem; all that means is the BIOS doesn't specifically know the model name. Hmmm - what's the model name for a 2.8 GHz FS dual core chip again? I don't know it either, though I suspect it will be FX-62. Cool 'n Quiet set to disabled by default? Well, unless CnQ is malfunctioning, the CPU will run at full speed in stress situations. Besides, Intel may have set it to disabled manually - after all, they almost certainly set the memory timings manually!
All told, I've seen nothing in any of the critiques that amount to a serious indictment of the results. Best case, the AMD FX-62 system might have been 5% faster with an appropriate BIOS and a few other tweaks. 5% is NOTHING! I repeat: NOTHING! Start a stopwatch and then count to 60 in your head without looking at any clocks. See if you can come within 3 seconds of the real time. In contrast, see if you can come within 15 seconds of the real time, because that's about where Conroe will be in all likelihood.
Let's also not forget that many of the AMD supporters (VoodooPC among them) have a lot to lose if Intel suddenly takes the lead again. They have to worry about sales between now and Conroe launch, they have to worry about new products, they have to start working with someone they snubbed for a couple years (Intel), etc.
Whom should I trust more: Anand that runs a website that caters to the enthusiast - a site that has been continually recommending AMD processors over Intel for the past two years! - when he says that Conroe appears poised to trounce anything AMD will have for the next nine months at least... or the word of a man that *wasn't* there to benchmark the system, sells 4 Intel desktop setups (vs. 16 AMD configurations - and 1 Intel laptop vs. 3 AMD laptops), and he stands to profit or be hurt by what enthusiast sites might say about an upcoming CPU transition?
Personally, I'll just wait and see what happens when everything finally ships. My bet is that Anand gets proven right (again), and Intel comes out on top (after their 3 year fall from grace). And if Intel manages to launch that Extreme Edition at anything near 3.0 GHz, AMD is pretty much screwed in terms of being "performance champion" until they get 65 nm chips shipping en masse. Even then, I'm not sure they can close the gap without a new architecture. After all, K8 is just K7 with 64-bits (who cares?), SSE1/2, and an integrated memory controller; great performance but it's certainly not a revolutionary design.
DSaum - Wednesday, March 8, 2006 - link
If Rahul Sood's http://voodoopc.blogspot.com/">http://voodoopc.blogspot.com/ is a true account of the errors in the Anandtech Conroe/AMD review then Anadtech should be prepared for a lawsuit by AMD shareholders like me who lost value as a result. It appears that Anadtech is either incompetent or malicious.