• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

A new rule I follow when upgrading

Permabanned
Joined
13 Nov 2006
Posts
5,798
I have just upgraded from a Q6600 @ 3.2ghz to an i5 SB @ 4.6ghz.

The upgrade in most applications is 2x as quick as my Q6600. Pi used to be calculated in around 16 seconds and it is now 8.

The next time I will upgrade is when pi can be done in 4 seconds and most applications are benchmarked twice as quick as my i5. Probably in a couple of generations :)
 
But will you only consider single threaded performance? Pi might never be done in 4 seconds. Perhaps the coding over the next few years will come to take advantage of more and more cores so that processors will tend towards that over single thread performance...?
 
But will you only consider single threaded performance? Pi might never be done in 4 seconds. Perhaps the coding over the next few years will come to take advantage of more and more cores so that processors will tend towards that over single thread performance...?

good point.
 
But will you only consider single threaded performance? Pi might never be done in 4 seconds. Perhaps the coding over the next few years will come to take advantage of more and more cores so that processors will tend towards that over single thread performance...?

With any luck, games and other applications will take advantage of more and more cores so his test will remain useful.
 
But will you only consider single threaded performance? Pi might never be done in 4 seconds. Perhaps the coding over the next few years will come to take advantage of more and more cores so that processors will tend towards that over single thread performance...?


Yeh good point. Most applications Compression / Video / .NET etc take advantage of more cores.

If these applications are twice as quick benchmarking then it would be time.
 
Just seems pointless keep upgrading when theres not much benefit.

Say I'm video encoding for 5 hours and with a new rig i could do it in 4.

Hardly worth the difference. My Q6600 has served me well until now.
 
My personal approach has been to at least a 50% increase in speed, anything less and generally it is hard to notice the increase when you're not running a benchmark. When i've broken this approach i've normally felt disappointed that i spent a fair amount of money with nothing to show for it.
 
My personal approach has been to at least a 50% increase in speed, anything less and generally it is hard to notice the increase when you're not running a benchmark. When i've broken this approach i've normally felt disappointed that i spent a fair amount of money with nothing to show for it.

Yet if you wait till you see that big a performance jump, the new chip costs the same, but the old chip is next to worthless. Buy a Q6600 for £190 maybe on launch, then buy a £170 sandybridge, how much is the Q6600 that uses more power but is half the speed worth, £50 maybe.

Buy a penryn for £200 though inbetween and sell the Q6600 for £140, then buy an i5 750 for £170 and sell the older chip for £100-120, then buy a 2500k for £170 and sell the 750 for £120.

It ends up costing you marginally more, and I really do mean marginally, for always having the fastest computer around and not being stuck with a half speed power sucking chip for 4 years inbetween.


Big upgrading once in a while is the LEAST economic way to buy computer parts. Not only has the chip become almost worthless, so has the memory(well maybe not actually, got a great price on my ddr2 mem maybe 6-8 months ago, I'd get half as much now), the mobo will be worth almost nothing, etc, etc.

Depreciating of old parts means selling when they are all but obsolete gets you very very little. With upgrading it can be a juggling act and getting lucky with a great offer, getting memory at the exact right time, selling older memory at the right time.

We had 4gb ddr2 packs for £35 a couple years ago, ddr3 was maybe £55, then it went up and was £80-100 for a 4gb pack, and we're back down to £40 a pack.

The best value chips hold their value the best, a Q6600 was for intel insanely well priced, didn't see another as good value chip from Intel till the i5 750. If you got a QX6800 you'd have seen 60-70% of its value wiped out in a couple years, a Q6600 kept its value far better. Same goes for mobo's, buy the fancy daft £180-250 mobo, 2-3 years later its worth £40 while the cheap £100 mobo you could have bought is now worth £35.

Buy the ultimate "value" parts, depreciating hits you less hard, upgrading maybe once a year becomes a very small cost.


Essentially the choice is £50-100 total cost to upgrade after selling old parts, once a year for three years while constantly having the latest cpu and gpu, or wait and spend £300-500 in one go 3-4 years later, with no value from your old kit, and spend 50% of that timeframe with a fairly slow and power hungry computer....... similar cost but far better value from the consistantly upgrading route.
 
Last edited:
Your advice makes sound economic sense but the problem I have with is for me what you suggest just is not doable due to financial restraints (I am currently unemployed :() as a result last year I got a big upgrade since I went from a Core 2 Duo @2.13GHz to a i7 930 @ 3.80GHz (OC'd) I have since bought a GTX480 on the cheap due to their current prices and it is blisteringly quick! I will jump to Hex core if I can when the prices come down a lot more since I am not paying £450 for a CPU :P

Stoner81.

PS - I do agree with what you said my point is that it is not always doable ;)
 
I follow the rule of only buy stuff when the price isn't going to drop much further in the next 12-18 months. :D (I buy most of my stuff 2nd hand). Couple of examples:

Q6600: 2nd hand price has been £70-80 for the last 18 months
Q9650: ~£150 for the last 18 months
4GB branded DDR2: £40 (since the price hike ~16-18 months ago)

How can you tell if something is going to drop much in the next 12-18 months? Well it's partly educated guesswork based on past trends. For instance you can now get an i7 920 for £110-120 2nd hand if you look around. Unless anything staggeringly good value comes out in the next 12-18 then I doubt the price will drop much below £90. I also picked up a Corsair TX750W for £55 last summer, I could probably resell that next year for at least £50. :D
 
Yet if you wait till you see that big a performance jump, the new chip costs the same, but the old chip is next to worthless. Buy a Q6600 for £190 maybe on launch, then buy a £170 sandybridge, how much is the Q6600 that uses more power but is half the speed worth, £50 maybe.

Buy a penryn for £200 though inbetween and sell the Q6600 for £140, then buy an i5 750 for £170 and sell the older chip for £100-120, then buy a 2500k for £170 and sell the 750 for £120.

It ends up costing you marginally more, and I really do mean marginally, for always having the fastest computer around and not being stuck with a half speed power sucking chip for 4 years inbetween.
You need to factor in the cost and hassle of changing your motherboard everytime you swap your CPU...

Also I see Q6600 still being sold for around £70, not £50, but I agree with your point that if you upgrade often then you can do it quite cheaply if you go for the cheaper bang for buck parts.
 
Back
Top Bottom