Always read the wiki when in doubt, the one for RAID is quite good:
http://en.wikipedia.org/wiki/RAID

Note that RAID0 isn't really RAID at all, because it's not redundant. It should be AID0. Just know that if you use RAID0, you must backup your data well and often, if any ONE of the drives fail then all data is lost on ALL drives ... a scary thought that must not be forgotten. I personally will never use RAID0, because it is an insane idea, I mean HDDs are not reliable as it is ... why make them more unreliable just to get a bit more performance ? Maybe only if you care nothing for reliability.

Mostly in response to Electro, and additional comments...

Seagate makes very good quality HDDs, I've never had any problems with them. Hitachi also makes good drives. However, Western Digital, no so much, I've had some of their drives that were faulty by manufacture ... not good, their quality isn't all that great either. Just my opinion, of course.

Corsair makes very good RAM, fast and reliable. I'm sure there are better companies, but this one is usually the best for the price, cost/benefit ratio is high :) I do NOT recommend Kingston RAM, it is crap, low quality, low reliability.

For sound cards, that's a difficult issue. I know I've had problems with my Creative cards on both Window$ and Linux, so I don't think highly of these cards even tho they are extremely popular. I also had problems with EAX, games tend to crash or the audio is abnormal with it is enabled. I replaced them with ones from C-Media which work much better and are much much cheaper.

I agree with Electro that Marvel/Yukon and JMicron are names to avoid.

Guys, thank you all very much for your input :D It has aided me greatly. And yeah, sorry for steering the thread in an off topic direction - hope you didn't mind too much.

But now back to the thread's original topic: the CPU comparison. I have now gone a little more in-depth into each of the processors. Note that I am from now on comparing Q6600 with E8400 *NOT* E6850 since it seems to be both cheaper and have a larger L2 cache (6 mb instead of 4mb) while all other specs remain the same.

What I am looking at is the additional features each CPU comes with. Here is a quick run down:

BOTH have:
In addition Intel Core 2 Dual E8400 has:
In addition Intel Core 2 Quad Q6600 has:
My questions are:

1) Which of these are the most important ones when comparing these CPUs? (I have tried to look at what each of them does, but after reading a lot of articles in the style of "Intel Smart Memory Access (...) has two major components: memory disambiguation and an instruction pointer–based (IP) prefetcher.", which tell me /absolutely/ nothing I gave up *grin*)

2) Which are supported by the kernel? (Or may be some of them do not require any specific kernel support?)

Do not base your decision on the features that each processor has. Base your decision the best suits you. Again since you run several programs at once, it will be best to pick a processor that has a lot of processors.

I would not say Seagate makes very good quality hard drives because in previous years their SATA drives were not to spec and had issues enabling DMA in Linux. Also they tend to get hot. Hot temperatures of the hard drive means bad for magnetic medium storage and for the motor. If you compared quality based on warranty, you need to look further.

Corsair memory is very, very expensive to what you call a cost to benefit ratio. Corsair memory did not work well in K8 processors back then, but other brands such as OCZ works just fine. Sure go with Corsair, but their cost to benefit ratio is 2 to 1.

That is kind of my problem. I do not know how many programs are many enough to get a quad core. I have already listed the ones that I am used to keep running at all times, though I am still unsure whether it's enough to benefit from a quad core CPU.

My current CPU (Athlon XP 5200+ 2.2GHz) is pretty much always 5-10% busy with occasional spikes to 60-80% (thanks to amarok which tends to eat up cpu each time it starts playing a new song). If I add compiz and some flash plugin usage to the equation it's often up at about 30-50% all the time.

Does this indicate that I need a quad or a dual core?

Who will tell? I can only say that as I'm writing this I have five applications open and I am running Compiz:
CPU1: 2-5%
CPU2: 0-4%
RAM: 542MB
That is with a C2D 6600 (1st generation Core 2 Duo).

All right, thanks. All in all I am now leaning more towards the Quad than the Duo.

Since I am planning on overclocking, I would have very much like to know whether there is any difference between these two processors when it comes to OC capacity.

From what I have heard C2D can go up to 3.6GHz while C2Q up to 3.1GHz with stock cooling (core temperature 30-40 degrees celsius). Can anyone confirm this?

Lower power consumption processors usually have better over clocking ability than processors that are already consuming a lot. Sometimes the smaller fabrication process helps to push the processor to higher clocks, but only if the manufacture used the right combination and have perfected this smaller level of making the chip.

Not to confuse you more, but you may want to check the Q9300. Its price is a little higher and it has an odd multiplier than the Q6600. It performs as well as E8500. If you planning of over clocking with the Q9300, be picky of the motherboard. It requires a higher FSB than expected for over clocking, so some modifications may need to be done to cool the motherboard chip set off better.

http://www.xbitlabs.com/articles/cpu...0_4.html#sect0

For what you are doing on a daily basis, you do not need to do any over clocking. You do not need any faster than a 2 GHz processor for Linux. More processors and RAM you give it, the better it will perform.

I have T7300 (2x2 GHz with 4 MB of L2 cache shared) and it runs fine with the programs that I run. I have a GeForce8 8400M GS and does UT2004 just fine for me at 1440x900 at 16 bit color depth. Though Flash videos on hulu.com and youtube.com is sometimes sluggish. Flash videos requires a fast processor such as E8500 which has a good performance rating to handle this data.

Does thermal design power give a clue about how well a processor handles OC? In that case C2D E8400 (TDP = 65W) beats both C2Q Q6600 / Q9300 (which both have TDP=95W)


Thanks. And, wow, seems like Q6600 was surpassed in all tests by the Q9300. I might really consider it (it costs only 10% more than Q6600 anyways).

Which one would you suggest?

But why not? I mean, sure it may consume more energy, but OCing doesn't do any harm to the processor itself (or does it?) - so not overclocking is just not using the processor to its fullest capacity.


Hmm... That's strange. Considering that I personally have an Athlon XP 3200+ @2.2GHz and haven't had any problems with youtube videos (even when using compiz). Of course, you can't do too much cpu-intensive work at the same time as you watch videos, but the videos themselves (+ firefox) take up only somewhere around 35-45% of my cpu.

Overclocking WILL most definitely reduce the lifetime of your processor, if you didn't know. Some goes when you add a turbo booster to your car, or whenever you push things beyond their normal settings.

That processor is plenty powerful as it is, there absolutely no need for overclocking in this case.

Note that Flash Player for Linux is horribly inefficient for some reason.

Well, yes. But the problem is that I will most definately buy a new computer in 3 years or so, while processors are designed to last much longer. So even if OCing compromises CPU lifetime chances are that it will not matter. (The unfortunate aspect of this situation is that the guarantee is probably voided by OCing - now *that* worries me)

But you are probably right, so I will start out with default clock frequency and see how fast the processor is and whether I would like to do something with it.

Yup, unfortunately. And isn't the flash player plugin 32-bit only? Heard something about that some time ago - don't know whether Adobe fixed it yet though.

Any processor in the C2D series is plenty powerful. However, there is someone who roams these forums that has in his tag line "If it ain't broke, tweak it." I believe that saying proves true here.

The C2D processors have very low power consumption, run very cool and are very efficient. I've seen tests done where they took off the heat sink and ran it for hours without fail. Remember two things when it comes to overclocking, the clock speed a specific processor is given is a safe, stable speed and heat is the biggest enemy for any piece of hardware.

Generally you can get a little more speed out of a processor without sacrificing any longevity. However, the higher you go, the more likely you are to lose life. With the C2D, you could overclock the hell out of it, put a water cooling kit on it and it should last you a long time. Longer than you plan on keeping it.

Brandon

You may want to have a look here and check what you are after.There are benchmarks of a quad and a slightly faster dual core and the dual core is faster in a bunch of them.

I would suggest MSI P35 Platinum because it can handle near 500 MHz of FSB during overclocking (what . Personally, I do not like MSI, but this one I can make an exception. This board comes with an Marvel SATA controller that may not work or will not work well in Linux. The rest of the hardware should work fine in Linux with minimal problems. The awkward heat sink design for the motherboard chip set might be hard to choose a CPU heat sink and you may not like the design. Though there are different versions of this motherboard. One that contains only DDR2 memory (MSI P35 Platinum), another that contains both DDR2 and DDR3 (MSI P35 Platinum Combo), and the last that contains only DDR3 memory (MSI P35D3 Platinum).

If you are buying OEM processors, I suggest Thermaltake Big Typhoon for a good starter heatsink. Just make sure it fits the MSI board.


BallsOfSteel, water cooling is not as efficient as a heatsink and fan cooling system. Also motherboards are designed to be cooled by the processor fan. If the fan is taken away, the motherboard will overheat and fail before reaching the maximum limits of the mainboard and processor. An active heatsink is a lot better choice for overclocking than water cooling.

Electro... have you ever used a water cooled unit before? If you had, I don't think you'd be saying air cooling is the best method for cooling a computer. I have many friends who can vouch for the performance increase and lower temperatures on their boxes due to water cooling when overclocking.

You can even google it and find people prefer water cooling over air cooling and have much better results for it. If you can afford it and have the know-how, you simply do it when you want the best performance.

You may think liquid cooling is more efficient, but the amount of disadvantages that it has compared to active air cooling methods is a lot. Liquid cooling leaks, very expensive compared to active heat sinks, and does not allow motherboard cooling (capacitors, inductors, MOSFET, and other components).

I already Google it and compared the cons to the pros. It has too many cons to work with my setups. I need cooling methods to be reliable with out a leak or stall in its operation. Active heatsinks are the best because they still work even if the fan failed. Also active heatsinks are portable, cheap, and easier to work with. With an invention of Heat Pipes, active heatsinks can be made cheaper, more efficient, and take up less space. Even if Heat Pipes leak, the heat sink is still cooling and the liquid is evaporated fast enough or turned into gas before it even touches the traces on the board.