Dual core processors work best when software can run in parallel on them. So-called 'multithreaded applications' benefit from an additional CPU core because subroutines can be allocated to different arithmetic and logic units. Administering the threads carries an overhead, though, which means that dual core processors are never exactly twice as fast as their single core counterparts.
Chip-makers AMD and Intel have released dual core processors aimed at users who need high arithmetic performance and use mainly multithreaded applications. Programs such as CAD/CAM and audio or video processing benefit particularly from a second processor core. However, AMD and Intel's dual core chips for this market cost between $500 and $1,000, and are therefore much too expensive for the mass market. .
Dual core and the office
Relatively little multithreaded software is used on standard office and home computers, so the purchase of a high-end dual core processor is rarely justified. Having said that, mainstream users can benefit from dual core technology. If several applications are active at the same time and certain tasks are stalled, then a dual core chip is worth having. For example, a hard disk defragmenter may be running in the background, leaving insufficient resources for a foreground application like a presentation. Similar effects can occur when antivirus or anti-spyware scans are active in the background. In these circumstances, a dual core chip can be very helpful even on a standard office PC.
When you consider what's going on beneath the surface of a typical office PC in a larger enterprise, it's arguable that a dual core processor can be justified here, too. Applications installed by the IT department can create a multitude of processes -- these may not always be active, but it's almost inevitable that at some point a crucial productivity task (finishing a presentation, for example) will be held up by a lack of computing resources.
Now that both Intel and AMD have affordable dual core processors available (the $241 Pentium D 820 and the $328 Athlon 64 X2 3800+ respectively), there's little to stop this technology being widely adopted.
Talkback
I hate all the adds on this site.
and I have big feed
Firefox:
adblock.mozdev.org
Any current browser. I use IE and get no pop-ups.. Must be because I bother to do updates
One point that almost all AMD vs Intel reviews are missing is that the Intel processors require more expensive DDR2 memory while the AMD chips use DDR memory. Using the memory types and speeds listed in the article, the memory for the Intel systems cost $15.00 more than the memory for the AMD systems. Not a large difference, but is shaves the low-end price differential to $72.
Opterons on Sun are very storng.
Why do these reviews never provide numbers for the very common and performance oriented tasks associated with software development and execution? How about testing compiling java or .net applications, executing java, .net, php, etc applications, running databases (ok, we sometimes see that, but not part of any standard suite)?
I would think this would address a lot of people's interests, much more so than rendering 3D scenes, and is really a case where someone would benefit from the extra performance.
Regarding the multimedia related tests in general and more specifically the video and audio encoding tests, it would be very useful to include information about the used source media. Such as: What was the size/duration/length of the source video or audio file? What was the resolution of the video file?
This would make it possible for a user of a CPU not included in the benchmark, to compare the performance in some (inaccurate) way. Just seeing that the AMD CPU X can encode a file of unspecified length and resolution in 15 minutes doesn't say much about the general performance, only the relative performance between the processors included in the benchmark.
At the moment, with the presented review, I don't have a clue whether how useful it would be to upgrade my Athlon XP 2000+ to such a dual-core chip. Sure, video encoding would be faster... but _how much_ faster?
Second, I think it would be useful to include some kind of performance test of compilers. Building a large project with a commonly available compiler would be most interesting, as it would enable many people to compare the performance of the tested machine with their own systems. One example could be compiling the Linux kernel using GCC (with a specified version ofcourse) or compiling a large C++ project such as The ACE ORB preferably also using commonly available compilers such as the GNU compilers.
I likes the speed test thingy!
Also i lkes my steak well done and not raw like many so called chefs likes to cook it.
Help! Aren't we also comparing AMD 64 bit (Vista ready) solutions to Intel 32 bit (Vista-hobbled) solutions (except for EE)?
Thank you for this article. I am in the process of my CompTia certs, and this issue came up. I would like to say, I find that the bias toward AMD seems to be a cost issue. If you don't have the money to spend, then yeah, it's a better deal. Power consumption? You never discussed the wide array of MB makers that would have effected that heavily. Additionally, the EE can operate at 2 different FSB rates, meaning that it was dependent on MB type as to whether it would function at its full capacity. The worst part of the comparison - the lack of 64-bit software platforms taking full advantage of this architecture. Yes, it's true, the hyperthreading should significantly slow down the EE when there is no use of it (more for the data to flow through, overbreakdown of data, etc). It simply becomes overhead. Furthermore, the 955 chipset also adds needless overhead load when not in full utilization. Max power to a unit not using its resources will slow it down. I'd like to see this EXACT test re-run when we have a solid, 64-bit architecture OS and associated Benchmark software out there. Your summary really should have included the slow down on the EE being a lack of backward compatibility on processor functions with 32 bit architecture. It should also consider how different vendors will handle the different MB's. That would have made it more thorough. You can't compare two Corvette's for effeciency, heat build, etc., when one has 4 cylinders and the other 32. The difference would be a gigantic engine being drug into the ground when the max track limit speed is 100 mph.?? But, again, thankyou for this article. This is part of the reason I read ZDNet. Lots of good info about the Athlon that makes knowing about it in its current state, with 32-bit platforms still dominant, important to know before I go out and buy anything!
I have the top of the line amd and intel in front of me here now. Both on thier own brand name boards (stupid not to test it that way), with bfg 7900 gtx graphics cards, corsair memory, 10,000rpm hard drives.
Ran 3dMark06 on both
Intel has better numbers.
Sorry
I have started a poll on "Has Intel's New Chip Architecture Finally Defeated Amd?"
You can vote here: http://www.gitc.ca
Can’t wait to see the results for this!
Luke, I am your father..
by seing all the test above amd versus intel
i gues the winer is AMD
:) I luV AMD