PC industry upgrade strategies?

[VirtualLarry]

To upgrade or not to upgrade, that is the question.

I used to only upgrade, when my PC basically kicked the bucket.

Then around the conroe/A64 era, I was buying CPUs and mobos left and right.

I tried building machines to sell, but I lack a network of customers, and I don't advertise, so that didn't go so well. (Tried electronics flea markets, only sold a couple of boxes there. Not enough to be worth it.)

But when is a good time to upgrade? I mean in terms of cycles? Intel hasn't seemingly slowed down the performance increase train, which is good, even though AMD can't really compete on performance anymore.

I don't have a lot of spare money, and newer chips are pricey. (Edit: I think the most expensive CPU I ever purchased was a Q6600 at MC, on clearance for $200+tax. Next-most expensive was an X6 1045T at $130 or $140+tax. Next-most was a Q8200 or Q9300 for $100+tax. At some point I purchased some Q9550 CPUs too, but I don't recall what I paid for them. At this point, I really don't like paying more than $200 for a CPU, and much prefer it if it costs $100 or less.)

Not to mention, Intel now charges an "overclocking tax", by charging more for their overclockable CPUs and motherboards with overclockable chipsets.

Is it best to upgrade every new CPU/mobo that comes out? Is it best to buy one machine, and sit on it until it dies?

I mean, the latter makes more sense on a limited budget, sort of, but Intel has been making great strides on power-consumption, so at some point, it works out to be cheaper to upgrade more often, just to reap the power-consumption benefits.

 

[NickelPlate]

I think the only person that can honestly answer that question is you.

I personally don't upgrade stuff just for the sake of it. I have to be dissatisfied with my current system or have a real need or desire, be it for work or play or a failure of some kind. Cost/benefit you know? But if you've got some $$ burning a hole in your pocket and want to do it than go for it!

 

[Greenlepricon]

Half the reason to build a pc (in my opinion) is because it's just fun. If you think you'll have fun upgrading and have some extra cash then go for it. Saving $30 on your power bill every year isn't worth it if you're spending more on yearly cpu upgrades for 10% increases in performance each step. I would say upgrade whenever your old system is completely outpaced or you need the extra power, but otherwise an old computer will do just as well as a new one. I don't know if I support sitting on a computer until it dies but it's not necessary to have the best of the best. Just whatever is enough to have fun with.

 

[pelov]

The rule of thumb for upgrades, back when CPUs were increasing in processing power at a very quick rate, was 100%-to-150% faster than your current processor. If you were to apply the same rule of thumb today you'd probably be waiting for a decade or longer...

Now it's just a matter of whether you want or need to upgrade. Judging by your question, -- and correct me if I'm wrong -- it sounds like you're content but you're asking out of pure curiosity?

Unless you're feeling the chip slow down or think it's inadequate for your workloads, there is no reason to upgrade.

What are you looking to get out of an upgrade? Why do you feel you need to?

 

[OBLAMA2009]

if youre an average user, you probably wont need to upgrade for the next 15 years. im posting this on a cheapo pentium i got for $50 with board and it certainly does everything i will need it to do for the near future. as long as websites are expected to run on arm phones this is going to be perfectly fine

 

[Abwx]

Same as above , although i have a C2D laptop as well,
i posted with a P4 1.6 notebook that is about 9 years old.

 

[VirtualLarry]

Well, the reason that I'm asking, is because I have been running a pair of P35 motherboards for the last N years, ever since the P35 was released. Had a pair of E2140 CPUs, OCed to 3.2 and 2.8 (for some reason, one of them was not 100% stable at 3.2, random reboots after a week or so). Then I bought a pair of E5200 CPUs, planned to OC them to 3.5 or more, and upgrade. But I never bothered, thinking it was too much of a PITA to disassemble my working system. I finally decided to pick up a pair of Q9300 CPUs, aware that I would be limited to 3.0Ghz @ 400FSB in those mobos, because they were on clearance at MC for $100. I told someone that I would use those rigs for a number of years, if I were to buy those CPUs.

So I ran those CPUs for nearly two years, but my power-consumption has been up, and according to a KAW P4460 meter, one of the rigs (my main rig, with a GTX460 1GB OC card), is projected to cost me $289 a year.

For that much money a year, I could afford to upgrade to something a bit better and lower power-consumption.

So I decided to build an E-350 rig (1.6Ghz), which seemingly ended up slower than my C-60 netbook (1.0Ghz, with Turbo to 1.33Ghz) for some reason. I spent $130 on new parts, and used the rest of the parts I already had, so in total it cost me about $290, including OS.

I saw a slimline Gateway PC at Staples for $200 + tax AC, with a Sandy Bridge 2.7Ghz dual-core. Knowing that SB has good idle power-consumption, I purchased that.

It provided the extra CPU power I needed, and at the same time, it had around the same idle power at the wall according to the KAW. (36W)

I later added a PNY NV GT430 card, with a pair of LP brackets I got out of another PNY video card I bought. The reason being that I was getting chopped video in Skype, when there was a streaming video window in Waterfox underneath that window.

Unfortunately, it didn't fix the problem completely, although it made it much less worse. I finally guessed that the problem was with clip regions in Windows 7's UI, and how it composited the screen.

Anyways, the video card boosted the idle power consumption to 41-43W.

The 26" LCD that the computer is connected to, takes 63W.

I wanted a lower-power computer to replace the Q9300 (260W at full CPU+GPU load, or 150W at full CPU load, or 92-100W idle).

Ideally, I would have gone to MC and purchased one of their bundles. In fact, I'm still eyeing a 2700K and a high-end ASrock mobo with IDE and floppy. I still want the possibility of using the old interfaces if I need to.

That combo is nearly $400 at MC, with the $50 bundle discount. I don't even want to consider what Newegg would want for that combo.

But if I had purchased that rig, instead of spending money on two low-power rigs, I could have had a better long-term system, with lower idle power too (nearly all SB desktops idle at the same low power).

 

[OBLAMA2009]

does it really cost $300 a year in electricity to power a computer??????? i never would have thought is was that high. i read an ipad costs $1 a year to power so i would have guessed maybe $40 for a computer

 

[Abwx]

i read an ipad costs $1 a year to power

Although it is not much this number s not true .

does it really cost $300 a year in electricity to power a computer???????

Yes , if it work 24/7 with about 300W consumption.

 

[KingFatty]

Are you gaming? If so, find all the old games you already bought or wanted to play, then play those old games. After you finish all of them, you can then look at the newer games that would require an upgrade, and figure out what specs you need to upgrade to.

But what are you using your computer for? Why would you need to ever upgrade it, besides saving power? Or, if your sole purpose is to save power, just get a Raspberry Pi or netbook to save a crapton of power.

 

[frozentundra123456]

Although it is not much this number s not true .



Yes , if it work 24/7 with about 300W consumption.

Wow, never thought it could be that much, although you calculations are correct. Realistically though that would be a fairly high end computer running at full load, so more likely would be in the 50-100 dollar range even if running 6 to 8 hours per day at full load.

 

[dma0991]

My rule of thumb on this matter is that a PC should at least last 5 years, which is how long that I would expect this current IB based PC that I have to last. Usually I would most often find reasons to upgrade or build a new one if the current standards of the motherboard's connectivity are no longer good enough. I don't want a bottleneck to happen to my newer SSD that I would buy years from now just because of SATA 3's bandwidth limitation.

The best time for a regular user to upgrade would be at the release of a new version of Windows, Windows 8 being that exception and hopefully the next Windows wouldn't be that bad.

As for power consumption, it doesn't bother me in the least since I rarely leave my PC turned on longer than 24 hours at any given time or leave it turned on when it is doing nothing. It has a SSD, turns on fast, turns off just as fast. I prefer having a low power consuming secondary 'computer' on constant standby like a Raspberry Pi or this which handles my basic web browsing needs should I find turning on the PC unnecessary.

 

[Idontcare]

Wow, never thought it could be that much, although you calculations are correct. Realistically though that would be a fairly high end computer running at full load, so more likely would be in the 50-100 dollar range even if running 6 to 8 hours per day at full load.

Checkout this handy online calculator for annual electricity expenses.

300W @ $0.10/kWhr @ 24hrs/day = $262.08/year

Here in PA for my specific electric plan (100% wind power) my electric bill works out to be $0.155/kWHr when all the taxes/fees/distribution charges are factored in.

So for me 300W over the course of a year is costing me $406.

 

[Abwx]

Wow, never thought it could be that much, although you calculations are correct. Realistically though that would be a fairly high end computer running at full load, so more likely would be in the 50-100 dollar range even if running 6 to 8 hours per day at full load.

Who would have thought , mind you...

If it was only the PCs...

Assuming they are switched on 24/7 , comsumption for the following
items are :

A 5.1 power amp , with lowish sound volume eats roughly 50$.

A PC that is shut off but not galvanicaly eats 9 to 20$.

A microwave oven that is SWITCHED OFF , eat about 20$.

A switched off monitor eats no less than 2.5$.

And so on...

 

[smangular]

Checkout this handy online calculator for annual electricity expenses.

300W @ $0.10/kWhr @ 24hrs/day = $262.08/year

Here in PA for my specific electric plan (100% wind power) my electric bill works out to be $0.155/kWHr when all the taxes/fees/distribution charges are factored in.

So for me 300W over the course of a year is costing me $406.

Ouch is that about a 40% green fee your paying?

 

[Idontcare]

Ouch is that about a 40% green fee your paying?

Yeah, that's about the sum of it.

I pay $0.10986/kWHr for the 100% wind option. PPL tells me they would charge me only $0.07907/kWHr for 100% fossil-fuel electricity.

So that is a 39% green fee premium (which amounts to an extra $0.03079/kWHr).

Mind you PPL doesn't care whether I pay them for fossil electricity or wind electricity, regardless my selection they tack on an additional $0.03359/kWHr "distribution charge"...representing a 42% "monopoly" premium because here in PA we have absolutely no choice but to pay PPL for power distribution regardless who you actually buy your electricity from. (my electricity comes from Washington Gas Energy Services)

So really my choices are (1) pay PPL $0.11266/kWHr for 100% fossil-fuel power, or (2) pay PPL $0.14345/kWHr for 100% wind-based power (PPL collects on behalf of WGES and pays them later)...which at that point means it is just a 27% green fee premium.

But wait, there's more! PPL also charges a flat-fee of $8.75 per month for being nothing other than a customer. No joke, it is even called the "Customer Charge" on the line item. Yeah monopolies

And then of course the state has to get their share, pile on the taxes.

So, for August my power bill was $284. Of which $201.37 actually went to the electricity company who sold me my 1833 kWHr of wind-power.

Were I to take the 100% fossil-fuel power option then my $284 power bill would have been $224.23 ($59.85 cheaper), which means at the end of the day the green fee amounts to 26.7%. For the extra $60/month it is costing me to go 100% wind-power I get to feel just a smidgen better about my carbon-footprint. Not a bad tradeoff considering every time I take the family to Red Robin I drop about $50 on burgers and fries.

At any rate, I figure I make up for the premium of wind-power by saving gas driving a Prius (bought it used to avoid the lot premium of course). Cut my ~$300/month gas bill to ~$100/month right there.

 

[blastingcap]

Yeah, that's about the sum of it.

I pay $0.10986/kWHr for the 100% wind option. PPL tells me they would charge me only $0.07907/kWHr for 100% fossil-fuel electricity.

So that is a 39% green fee premium (which amounts to an extra $0.03079/kWHr).

Mind you PPL doesn't care whether I pay them for fossil electricity or wind electricity, regardless my selection they tack on an additional $0.03359/kWHr "distribution charge"...representing a 42% "monopoly" premium because here in PA we have absolutely no choice but to pay PPL for power distribution regardless who you actually buy your electricity from. (my electricity comes from Washington Gas Energy Services)

So really my choices are (1) pay PPL $0.11266/kWHr for 100% fossil-fuel power, or (2) pay PPL $0.14345/kWHr for 100% wind-based power (PPL collects on behalf of WGES and pays them later)...which at that point means it is just a 27% green fee premium.

But wait, there's more! PPL also charges a flat-fee of $8.75 per month for being nothing other than a customer. No joke, it is even called the "Customer Charge" on the line item. Yeah monopolies

And then of course the state has to get their share, pile on the taxes.

So, for August my power bill was $284. Of which $201.37 actually went to the electricity company who sold me my 1833 kWHr of wind-power.

Were I to take the 100% fossil-fuel power option then my $284 power bill would have been $224.23 ($59.85 cheaper), which means at the end of the day the green fee amounts to 26.7%. For the extra $60/month it is costing me to go 100% wind-power I get to feel just a smidgen better about my carbon-footprint. Not a bad tradeoff considering every time I take the family to Red Robin I drop about $50 on burgers and fries.

At any rate, I figure I make up for the premium of wind-power by saving gas driving a Prius (bought it used to avoid the lot premium of course). Cut my ~$300/month gas bill to ~$100/month right there.

I suspect the customer charge is just a flattened rate they give for a typical residential customer--the equivalent of a "demand charge." http://www.northwesternenergy.com/documents/E+Programs/E+demandcharges.pdf There needs to be things like spinning reserves to step in for the windmills when winds blow more slowly; and paying for the maintenance of distribution and transmission wires carrying power from the plant to your home (as well as tree-trimming and other utility worker wages to keep those wires up); control center staff; etc. Those costs are there whether you use 0.00000001 kWh per month or 10000 kWh per month.

It's probably not economic to do so yet, even with state/federal subsidies, but consider getting solar panels on your roof. Those come with their own problems, though, including thieves who pry them off your roof, much in the same way thieves steel copper cables from construction sites--and even live wires, though those stupider thieves rarely live to tell the tale.

 

[sequoia464]

It's probably not economic to do so yet, even with state/federal subsidies, but consider getting solar panels on your roof.

This can be a great way to go for some. I'm in California - one of my friends that lives in the central valley (hot and lots of air conditioning) just had this done - killer savings for him. I live on the coast so not really economically viable for me yet.

Edit: unless I do the install myself that is, which is actually - except for the power inverter - pretty straightforward.

 

[Abwx]

So, for August my power bill was $284. Of which $201.37 actually went to the electricity company who sold me my 1833 kWHr of wind-power.

2.5KWh permanent comsumption is huge for a private person
unless one has some kind of rendering farm at home...

 

[Idontcare]

2.5KWh permanent comsumption is huge for a private person
unless one has some kind of rendering farm at home...

My house sits on what is essentially a lake as far as the water table is concerned. This was unknown to me when I bought it. But the result of that is my basement runs 2 sumps pumps off and on all day long plus a dehumidifier that runs 24x7.

It is ridiculous but my only other choice is to let the basement turn into a swimming pool. Which judging by the one time I lost power for 4hrs and the sump pumps weren't working I would guess it would only take about 3-4 days before basement would have around 1ft of standing water.

Think of my house as a big leaky boat with expensive bilge pumps I hate it and would love to sell it if only it weren't for the fact that it is currently ~$60k underwater (financially, not literally).

I am moving to Taiwan in the spring though, so I'll have a chance to get away from the problem and move on. Haven't decided if I will sell the leaky barge for a loss or if I will rent it and try to ride out the housing market before putting it up for sale.

 

[Abwx]

Would be interesting to know the nature of the terrain.

A basic 40 litres/minute pump comsume about 100 to 130W , so unless
you re litteraly sitting on a river flow there s no way for such pumps wich have
about 8 metres elevating capacity to be inadequate , the hourly moved volume
being 2400 litres , two such pumps would almost reach 5m3/hour , so it s
likely that your current system is largely oversized.

 

[Xpage]

I actually saw a Holmes on homes episode where there was a house with a similar situation as good. They coated the basement walls with something thick, a waterproof plastic then setup a wicking system in the basement after some digging, it was a closed system, and worked well. House was nollonger humid, no mold. You may want to look that episode up

 

[Puppies04]

I actually saw a Holmes on homes episode where there was a house with a similar situation as good. They coated the basement walls with something thick, a waterproof plastic then setup a wicking system in the basement after some digging, it was a closed system, and worked well. House was nollonger humid, no mold. You may want to look that episode up

Or you could turn the pumps off and breed cave fish in the basement. Their sigthless eyes are worth a pretty penny on the chinese black market.

 

[nyker96]

if money is concern, then only upgrade the parts you need or died. if you need more HD room, get a HD, if more cpu speed get a new cpu etc.

and don't upgrade to save energy bill cause most likely it won't be worth it.

 

[VirtualLarry]

and don't upgrade to save energy bill cause most likely it won't be worth it.

Hmm. KAW says my new lower-power G630 rig will cost me $52/year, instead of $289/year for the Q9300 @ 3.0 rig. Since it cost me $200+tax, plus a $65 120GB SSD, plus a $50 +tax graphics card, it should pay itself off in around a year and something.

By then, Haswell will be out. I'm looking forward to Broadwell too. Will Haswell mobos accept Broadwell chips, and will they finally release consumer CPUs with more than four cores with Broadwell?

Edit: Now that I think about it more, I guess if I'm only breaking even in a year and something, then perhaps that wasn't the wisest choice. Either way it works out the same, cost-wise, after a year, and then I will likely upgrade to Haswell. However, I will then have the extra PC around to re-sell, so I suppose it's a win.