I have to disagree. Consumers nowadays are doing the following things:

1. Video encoding - transcoding IPOD Video/Touch/iPhone, transcoding TVB dramas, PC to XBOX360 etc
2. Picture processing - there is a growing consumer base going forward to SLR photography, processing RAW using Adobe Photoshop and other consumer retail software
3. Home Video encoding - encoding/decoding recorded home video to DVD
4. New generation PC gaming - Crysis, etc

These are the applications that are multi-threaded SMP in nature. According to Techspot review of an Intel Q6600 G0 (http://www.techspot.com/review/36-intel-core2-quad-q6600/page6.html), on real world content creation applications, it has an average of 10-30% improvement (time savings and performance) over X6800. Thats close to 80% time improvement over any single core processors. So the only thing a consumer will not go for quad core is actually the price and not whether they can fully utilize a quad core processor. Dual cores are cheap as chips nowadays so again thats another sweet area to rake profits. Just look at the E-series and upcoming Celerons.

On other applications, you're right because current retail software are way behind in terms of SMP and will only be good on enterprise applications. But that trend will change very soon since Windows Vista is a platform for multi-core processing already. We can expect situation to improve. Right now buying a Quad Core is rather to secure the future. Its nice to know that there are more headrooms to discover when more software introduce SMP.

Now with regards to licensing, the licensing charges are per physical processors and not virtual processors like P4 HyperThreading and in this case of interest multi-core processors but with an exception (depending on individual policies); that per physical processor does not have more than 4 processing cores. That is one reason the industry is pushing multi-core products because you'll save a lot of value when it comes to this. Check out the following references:

http://www.vmware.com/download/eula/multicore.html
http://www.microsoft.com/licensing/highlights/multicore.mspx

I'm sure there are more software companies implementing the same licensing policy especially when Microsoft adopts this policy. I just hope you didn't fail your assignment and have included reasonable considerations for this point.

Now who say native quad-core processors will beat non-native quad-core processors? During the 2006, there has been numerous discussions that non-native multi core processors will have to depend on FSB to transfer information to each processor core. That was the Pentium D. Today, non-native quad-core is as good as native quad core. Ever heard of shared L2-cache? Thats how Core/Core2 processor architecture share information which releases a lot of overhead. The only exception is when 1 core of silicon A wants to share information with another core on silicon B on a quad core processor where it had to go through FSB.

Intel will not go native quad core is because the benefits are low at this point and the risks are very high. Yield problems is going to create havocs. One faulty core will jeopardize the whole 4 core silicon. Thats 100% lost per processor. A native quad core silicon die takes up a lot of space, thats up to 2 times any dual core silicons per silicon wafer. Imagine the loses at this point. AMD had to implement e-fuses for its faulty Phenom and introduce tri-core processors later in the future. Look at Sony's Cell processor development. It'll tell you a lot going that path of road.

By going non-native, Intel can maximize its yield and can make decision very late into the production stage whether to make a Core 2 Duo or a Core 2 Quad. That gives Intel a lot of flexibility in determining how many C2D or C2Q depending on their marketing strategy.

This post has been edited by davidmak: Jan 11 2008, 05:24 PM

Yeah, its really disappointing. I think AMD has been making strategic mistakes every step of the way. Seems like a perfect storm. The only reason Intel can be that strong is because of its manufacturing advantage it has over AMD. Thats a classic rule why strong companies with cash and manufacturing capability can outrun any competitive company and flood the market with competitive products.

AMD really woke up the sleeping giant. The element of surprise is lost and Intel shouldn't be asleep real soon until they get relaxed again.

Wah lan, serious price ler. Looks like a desperate rush to dump chips. Seriously speaking RM680 (if this price is accurate), this is cheap quad-core processing power ler. Never mind the performance issue unless it really cannot compete with an average Core 2 Duo like E4xxx or E6xxx.

This is really bad for AMD in terms of reputation. New year, new product already got problems. The worst thing is the problem was found late into the supply chain into the customer/retail distribution channel. Then decide to dump them at low cheap price. Good for consumers but never good for AMD.

Yeah Q6600 is a little pricey but can use for a longer time. Probably when multi-core processing is a norm.

Yeah agree. A good quality 500-550W real power PSU with enough amperage on the 12V rails should be good. I saw a few forummers strapping 550W PSU on their Q6600s and 8800GTSes. Seems ok to me.

True but to a certain extend. The PSU rated at 400W, 500W, 600, 700W, 800W and so on are the capacity it can spit out in DC in which it can convert from AC mains. It does not mean that a 500W PSU is drawing 500W all the time. Thats merely the capacity. If you study the power distribution, it is rated in different amperage at different voltage rails. Don't be fooled. You can't add these like math. Our PC system has different components drawing from different rails. So the load varies depending on usage and the combined ratings will be in effect.

When not playing games and using the system for regular work like surfing, it is only drawing much lesser. Probably to the tunes of 200W. Most of modern CPU, GPU and other components of the system implement idle switching for some form of power savings. When in idle mode, it will switch off or disable most of its silicon section to reduce power and go into a deeper sleep. Processor frequency, FSB, etc all reduce to a lower state to save power. Also with advancement of silicon manufacturing process, CPU and GPUs are drawing lesser power due to reduced processor voltage. It is when you push your system with OC or doing full work load where it will go near the maximum capacity. So our usage of our PC varies from time to time. If you use your PC merely to play games, then consumption will be higher than those who use their PC for general purposes. Also all those idle times will eat into consumption too because it is drawing power but doing nothing.

Again, 500W or 800W or 1kW might sound a lot of power. But that wattage are expressed with respect to 3.3V, 5V and 12V. Our electrical bills are calculated in kW/hr at 240V. So to access the consumption, you should use a power meter between the mains and PSU. Now most modern PSU has some form of power factor corrector so this is not likely a big issue because only up to 20% is lost due to switching losses. If you're drawing 300W of DC12/5/3.3V combined, it is actually taking 360W of DC12/5/3.3. Convert it to power drawn in terms of 240V AC, its likely a smaller number. In the end, safe to say, 1-2 hours of gaming on a high-end rig would end up around RM2 or slightly higher. Sorry only rough estimations because a lot of unknown variables.

Fortunately electricity comes at a fair price. Buying good modern PSU is in fact a good thing. We help save power due to switching losses, reduce feedback electrical noises to TNB, reduce emissions, etc. We're more concern on the efficiency of power generation and its effects to our environment rather than the consumption of it.