OP: I felt compelled to respond and maybe try to clear the misconception. I understand your strategy of upgrades because I used to be like that myself. However, I've long discovered the flaw in it and would like to attempt to help you do the same. You mentioned a 5970 but I'll make it easier for you and do a 5870:
5870 NOW = $400 or more
5770 NOW = ~$160 (can be found cheaper on sale)
The 5770 route would have you upgrading in 2 years (because of your low res/cpu.) In 2 years, the $160 equivalent from either side of the fence would rape a 5870 with a rusty pipe. Keep that for another 2 years.
End result: Your method has you spending at least $400, not seeing any meaningful advantage for the first 2 years and seeing a disadvantage for the 2nd two years, assuming you keep it that long. Advantage means compared to getting 2 mid-range cards.
My method has you spending $320 total and at worst has you experiencing marginally worse performance due to your bottlenecks for the first 2 years, while experiencing much better performance for the 2nd two all while costing at least 20% less.
I know you're not a trade and sell type of guy but this calculation assumes you would throw the old card in the can.
What works for me is that, every time I have the upgrade itch, I throw that money aside. When the pile gets high enough, I go out and upgrade the whole system at once to whatever is the sweet spot then. I used to sell my old parts but now I mostly give them away due to ebay/paypal rapeage. I know this way is cheaper AND provides better experiences because I used to do it your way.
It's your money, so spend it how you like. I know you think you're saving money using your strategy but you're not because of how quickly high-end loses value. If you don't like tinkering with your box often or you have a hard time saving money (not judging, I've just had friends back in college who can't hold onto $20 for more than a week), then this strategy is not for you and your strat would indeed be better. But if you don't have that problem, you may want to reconsider what you plan to do.
I would like to throw in my 2c on this strategy. I used to just haphazardly buy video without much of a strategy, but I usually ended up with either too much cpu or too much gpu as well. A few years ago I got onto the "buy the best ~$150 card you can every year" strategy and it has worked unbelievably for me.
I started with an x1950xt. I had very poor timing as I bought it 6 mos before 3870 came out. ditched the x1950xt on fs/ft and lost ~$80 after 6 mos, bought a 3870 to replace it in mar/april 08 range. bought a 4850 a few months after launch, lost $50 in the swap. I would still have that one, but the annoying single slot cooler was too loud so I sold it on fs/ft and bought a gtx 260 oc, again for ~$50 net.
So I ended up going from x1950xt which was bought very late in the cycle, and now I have spent less than $200 over a 2 1/2 year period to keep generally midrange performance. I wanted to upgrade last year, but there wasn't anything respectable in the 150-200 range for me and I knew that my gtx 260 was worth only $100 so I waited. and waited. I'm still waiting, perhaps for 5830 or perhaps I will end up holding out a while longer and getting a fermi-based derivative in q4 2010 or q1 2011. Fortunately, there isn't much out there that smacks around a midrange card these days so I have options.
Both of these points are very valid. temjingold, I'd like to point out that I wouldn't consider the 5770 to be a mid-range card, and that the
second year of owning that card would suck. Under ordinary gaming market conditions, that is. These days, yes, a $150 card could last me 2 years.
to byranw: I think that your "buy a $150 card every year" method points out the true long-term value of these cards. However, today's $150 cards are faster than an Xbox 360. Hell, so is my card.
So, I guess what I'm getting out of this is that
anything will run modern games at 1440x900. If thinkgs get so dreary that I simply can't play modern games anymore on this thing, then I'll buy a $150 card...
If this thing keeps on sloggin' through today's games, though, I'll keep her for a while. I originally said "until this fall," but that's not a hard deadline. If Jumpgate Evolution comes out and she can't run it on High, then I'll rethink my options. But if
nothing comes out that I can't play (and yes, it must at least look relatively good at whatever settings I can take) this year, then I'll save my $150. And my $400.
As you guys have taken into account with your LCD suggestions, monitors hold their value better than internal components. However, the sheer number of 120hz screens launching this year could inject some competition into that market, and drive prices down. If ATI embraces a 3-d solution in their 6000 series, then the cost of the glasses will probably drop too...
If I'm not seeing much gains from the first year of upgrades, why
have a first-year upgrade? It would be nice to get onto some sort of upgrade schedule, so I'll try to pick parts (eventually) that result in a more easily upgradeable system. And do it all at once, even though I'm not convinced that said method actually saves money. Perhaps I'll mirror my last cycle -- save money by buying a card at the end of its life-cycle, after something else comes along that's the "next big thing." In any case, you've convinced me
not to upgrade this year unless things get worse.
If $150 cards can last for two years, then apparently performance shouldn't be a reason to go big or go home. I mentioned 3dVision several times; I'd really like to get a setup working. Sometime.
Conclusion/tldr: I'll just save, and do a larger (meaning price, performance boost, or both) upgrade next year. Or buy a $150 card this year, if needed.
I still don't think that I'd hold on to a $150 card for more than a year, though.
It's not really worth spending money to support AvP, Bioshock 2, etc. What games that are coming out this year will both be fun
and require a lot of muscle to play? I can't think of any. So, if I upgrade
soon, it'll just be for the 3-D. And a 9800 can run 3dVison just fine, thank you. At least, it can run Freelancer in 3-d. Probably at the full 120 fps, too. What else could you need?
At that point, however, all of the upgrade cost is going towards the 3-d setup. Might as well wait until that setup gets cheaper...
Oh ze noes, my logic is recursing!
I think I just put a tl;dr smack-dab in the middle of a post, and wrote an entire paragraph for it. Whoa.
Hopefully the OP will see the logic in your post.
Facebook
Twitter