backtop


Print

Since I recently had the chance to test two "Conroe" chips, I'd like to share my experiences with those exciting new CPUs

Like many other sites out there, German hardware site K-Hardware (which I'm writing for) was also provided with samples of Core 2 Duo / Extreme chips. Of course Intel sent me a Core 2 Extreme X6800 and a Core 2 Duo E6700 along with a D975XBX Revision 304. In Germany the CPU manufacturers tend to give out the faster models only, since they'd like to sell more of those with higher margins on them.

Performance-wise Conroe leaves no question. I already expected it to win just about every benchmark and it came true during the actual testing. AMD only managed to win in memory bandwidth and latency tests as well as in ScienceMark 2 - Molecular Dynamics. Other than that, Intel won everything and most of the time leaving a big gap between the new chips and AMDs FX-62. Even a E6600 outperforms AMDs enthusiast chip by a good margin. On top of that, Conroe overclocks great, which will make it the #1 choice for enthusiasts.

Now enter the bad things... Yes, there are a few bad things about Conroe. First off is temperature. While Intel managed to significantly cut power consumtion, the core temperature during load is almost as high as with Netburst-based Pentium XE 955 and about 10°C higher than compareable AMD chips, which boast the same power consumption figures when measured at the wall outlet. Just for reference, my X6800 went up to 66°C, while the E6700 remained at 60° thanks to lower core voltage. A possible explanation is the higher thermal density with 65nm, but those Netburst-chips were 65nm as well, so I'm not satisfied with that. I would have expected lower thermals.

Another thing I didn't like was the way how they introduced some new instructions unofficially dubbed SSE4. Upon request Intel told me that those instructions were internally named MNI (Merom New Instructions), but will not receive some recognizeable identifier. It simply doesn't exist in Intel's marketing. SSE4 doesn't exist, albeit those instructions are there. I wonder how people will specify that you need those instructions, when there is no official way to call them. Maybe people will simply stick with SSE4, which is already used in a range of CPU information tools. The CPUID flag is there, it just lacks an official name.

Another not so nice thing is motherboard compatibility, something Intel will never get right. You always need a new chipset and even if you have that, some nasty VRM could be in the way. But this problem is already well-known and people have apparently accepted to get a new mainboard everytime they buy a new Intel CPU. Of course I'm also concerned about availability and I'm not alone with that either.

That being said, I still regard the Core 2 Duo processors as a great choice and I can only hope that those people, who want to have one of those, will get hold of one rather sooner than later. I also believe that such strong competition will be good for AMD, as they are now forced to be more competitive.

Those interested could check out my review here.




"I'd be pissed too, but you didn't have to go all Minority Report on his ass!" -- Jon Stewart on police raiding Gizmodo editor Jason Chen's home
Latest By Marcus Pollice
AMD Talks Details on K10
April 14, 2007, 11:46 AM
Intel Readies Merom-L Core
April 7, 2007, 9:16 PM
NVIDIA Holds Back "G84, G86" Cards
March 23, 2007, 7:08 PM
CeBIT 2007: GeForce 8600 Spotted
March 14, 2007, 1:19 PM
NVIDIA Not Present CeBIT
February 25, 2007, 7:36 PM
NVIDIA's Current Driver Mess
February 12, 2007, 7:32 AM







Latest Blog Posts
Amazon Fire HD 8
Nenfort Golit - Jun 19, 2017, 6:00 AM






botimage
Copyright 2017 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki