System Builder Marathon, Q1 2013: System Value Compared

Page 5 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.
Why the same CPU and GPU? I don't get it. You're just making a case for someone to forego an SSD in return for a faster video card. Which I personally wouldn't recommend-how many of us only use our machines for gaming?
 


Heh, I was just joking about the naming thing.

Ten years? Out of your case and (perhaps) your PSU, maybe. Out of anything else? I suppose there's a vanishingly small chance your hard drives will still be relevant. Hard to say; capacity requirements are subjective and highly volatile over time. I'm sitting on several 80ish GB HDDs from years ago that technically still work (or did, last I checked), but they're effectively useless because they have too little capacity to justify a place in either of my rigs.

My Seasonic PSU from ~2005 is still going strong in one of my machines, though -- along with all the other components bought at the same time. But I expect very little from that rig these days. It's a network storage device that runs Firefox.

To some extent, I think we're talking past each other here. I would never advocate skimping on a power supply, for example -- but happily you don't have to break the bank to buy a good quality PSU. Power requirements (relative to performance) have actually declined in recent years, so if you're happy with a GPU/CPU combo today that'll run on a ~500 watt PSU today, chances are that same PSU will feed a comparable-performing GPU/CPU combo two or three generations hence. $70, or thereabouts, buys you a really nice ~500 watt PSU.

Likewise, if you drop $80-100 on a case, you can probably reuse that for two or three builds.

I'm not suggesting that you can't get long-term mileage out of certain parts; I'm saying that worrying about most components' upgradeability isn't worthwhile. You build the best thing you can, for your budget, today, and the rest should take care of itself. I see a lot of people shoot themselves in the foot because they feel shackled to the bulk of their old machine; choosing to jump through hoops to upgrade it every year or two rather than ripping off the proverbial bandaid and starting over every 4 or so.

Tech moves too fast, and prices drop too fast, to restrict yourself in that way. Upgrading individual parts is a good way to explore bottlenecks; it ain't such a great way to wring the best possible value out of your hardware.
 
Hi! It's nice to read a good review.
My remark would be that an enthousiast users spend much time "multi-tasking". That is a point I miss. The SSD in the 1000$ pc should give a good advance in that situation.
I'd suggest make an SBM including 3 hardware setups for the same price, but for different purposes : games, productivity and content creation/multitasking.
 
I'm not sure what the target audience is, but I have zero interest in the three-budget-build-showdown. You're a site for enthusiasts. I may be in a minority here, though I doubt it. I'd be much more interested to see you go back to low-middle-high-ultra levels of builds as opposed to this cheap-budget-low end build lineup.
 
[citation][nom]Fulgurant[/nom]...Ten years? Out of your case and (perhaps) your PSU, maybe...My Seasonic PSU from ~2005 is still going strong...Power requirements (relative to performance) have actually declined in recent years...I'm not suggesting that you can't get long-term mileage out of certain parts; I'm saying that worrying about most components' upgradeability isn't worthwhile...I see a lot of people shoot themselves in the foot because they feel shackled to the bulk of their old machine...Tech moves too fast, and prices drop too fast, to restrict yourself in that way. Upgrading individual parts is a good way to explore bottlenecks; it ain't such a great way to wring the best possible value out of your hardware.[/citation]
I'm thinking my SG-650 may outlast ME, as overbuilt as it was found to be. Declining power requirements, as you point out, should nicely compensate for whatever limited capacitor aging will affect it. Going from 80+ bronze to 80+ gold (or even platinum) would be penny wise and pound foolish.
As to hard drives, I've been inclined to replace them, needed or not, after ~4 years or so, figuring they're due for a failure. I've been buying mostly WD Black drives though, with 5-yr. warranties, so I'll probably use them for 6-7 years. As to the bolded statement, this has been true. I try to tone it down in the forums, but I'm inclined to believe the worst as far as the economy goes, as our elected parasites seem determined to enact policies that destroy the value of our money. That's a key reason I've tried to buy parts I won't have to replace in some kind of automatic upgrade cycle. As to exploring bottlenecks, that's one reason I enjoy reading Tom's, so I don't encounter unexpected bottlenecks. Perhaps it is also a matter of age, but I tend not to play the latest, most demanding games. Nothing I have cannot be played on "enjoyable" settings using a HD7770; oh certainly my HD7870 is nicer, but I don't feel like I'm suffering on the lesser card. In fact, the HD7870 is on a shelf at the moment, as I'm wringing out a GTX650Ti to explore nVidia vs. AMD differences.
Anyway...I signed up for the contest. Any of these machines are very similar to what I'd build. I've got a friend whose old E5200 machine died recently, so I might do a little parts shuffling, but then end up building something for him.
 


(Emphasis mine.)

We are on the same wavelength. :)
 
Many, many posts on these forums are about building ultra-budget gaming boxes. I would ask the System Builders to DROP the money amounts to present a more humble spectrum for folks of more modest means. I'm thinking $300, $400, and $500.

Now, this isn't something I'd be interested in seeing EVERY quarter, but it would be a nice change of pace.
 
I too would be very interested to see if "enjoyable" game play is possible on some low-budget builds. Or, if the budget is so low, what are realistic expectations? From recent game review articles, it looks like as little as a HD7750 can handle any game, particularly at less-than 1920x1080 resolutions, with "decent" settings. True?

[citation][nom]Fulgurant[/nom](Emphasis mine.)We are on the same wavelength.[/citation]
Yes, it did seem that way 🙂
 
[citation][nom]count_rugen[/nom]... I'm thinking $300, $400, and $500.Now, this isn't something I'd be interested in seeing EVERY quarter, but it would be a nice change of pace.[/citation]

If your budget is that low, just buy a generic box from a generic vendor, because nothing special
is going to be viable.

But if your budget is low and you do care about trying to get the most for your money, then buy
2nd-hand; thus my earlier comment about the parts I've obtained in recent months, eg. two GTX
560 Tis for 177 UKP (3DMark11 = 9600+), various SSDs (Vertex2/3), PSUs (Thermaltake
Toughpower, 700W to 1200W), mbds (Gigabyte GA-Z68XP-UD4, ASUS Maximus IV Extreme),
HSFs (more TRUEs with excellent Noctua/Scythe fans included, Phanteks PH-TC14PE), etc.

IMO it's a simple choice if you have a low budget: if you want to build the system yourself, buy
2nd-hand, because buying new will result in a very boring box IMO. If you don't want to build the
system yourself or want new parts, then get a generic Dell or whatever, because a cheap set of
modern new parts isn't going to allow for anything useful re overclocking anyway. Remember,
current non-K Intel CPUs can't be oc'd much, whereas a cheap old Clarkdale offers huge potential
(at least on that front AMD offers something more re CPUs like the 965, but the rest of the system
with a low budget would be kinda naff so having a nice budget AMD CPU doesn't help).

Ian.

 
I'm exactly that user you're dismissing when you say that someone with a big monitor isn't shopping for a $600 gaming PC. I have a $350 refurbished Auria EQ276W I picked up from Microcenter. I can't find a color calibration that I like, but it's 2560x1440 with DisplayPort, HDMI, and DL-DVI for cheap! There are other options out there too, especially for hooking up to a gaming PC, that can get you a 2560x1440 IPS panel for a LOT less than the Dell or HP screens.

When I sit down to price out gaming PCs for Mechwarrior Online, SC2 Heart of the Swarm, Skyrim, etc., your value PCs are exactly what I look at.

However, I would also always, always put at least a 32 GB SSD in any system I would build.

So I think building a $750 gaming PC with an SSD and focused on 2560x1440 gaming performance would be extremely interesting. Double bonus if you can keep it quiet and stuff it in a mini ITX case. :)
 
While the “Best CPU/GPU for the Money” articles are my favorite, I feel they fail to recommend the ideal combinations of CPU and GPU for gaming. Perhaps these “System Builder Marathon” articles could predict the CPU/GPU combo “sweet spots” followed by building 3+ distinct rigs to verify or disprove these predictions.

Initially some research with different CPU and GPU combinations would be performed and presented, probably just those models from the “best for the money” articles. Perhaps even present matrices of CPU/GPU combinations displaying frame rates and value in a variety of games. I realize building and testing each combination would take a phenomenal amount of labor, but I’d be happy with extrapolated data from a minority of known data points.

Distinct rigs would then be built with certain shared components (ram, storage, case, etc). Or at least have some rules about minimum spec requirements, especially regarding storage (minimum size, ssd required/disallowed, etc) (requiring an ssd might lead to some creative solutions like a cheap small cache ssd running intel srt). Obviously the resulting rigs would be built with a "sweet spot" in mind, rather than a fixed budget.

I’d also really like to see some APU’s thrown in, at least in the research stage just to see if there’s possibly a “sweet spot” at the very low end.

Alternatively, perhaps the ideal CPU and GPU combos could be integrated into the “best for the money” articles or occur as another monthly article.
 
1) cheaped out motherboards
2) cheap unimaginative chassies which look like aluminum stuff, may as well just buy a Gigabyte Setto and save more cash.
3) same processor used, its like lack of imagination.
4) same Graphics recyled.
5) good enough PSU but hardwired, in a cheap chassis with limited cable management must be a nightmare cabling up that system.

why dont we just recycle the same setup over and over, just throw in something a little more expensive to make up the costs. I Seasonic G series RRS 450w is modular and powers all these systems for $50 or so, take that over all these Neo Eco's particularly in those cheap chassis's. Intel systems don't benefit from DDR3 above 1600 anyways and good looks in a chassis where you cannot see them is pointless, the Crucial Ballistix Tactical Tracers are expensive modules when a simple $35 DDR3 1600 will suffice. i5 3330 and two builds with 3570k then the $1600 alternative build (lol) with another 3570K, yes I get that the 3570K represents Intels only real chip of value but what about the 3470 and possible even a i3 on the cheapo build.

This was just boring.
 


No one, to my knowledge, has ever claimed that you can't find affordable high-res monitors. The issue is that high-res monitors require correspondingly high-horsepower hardware. That's why it's impractical to game at high resolutions on a tight budget. Your GPU has to render 78% more pixels than it would if you were using a 1080p monitor.

Even if you can build a $600-$750 rig that delivers playable frame rates (with good image-quality settings) at your chosen resolution -- which, I grant you, may be possible at any given moment in time -- the rig would only perform up to that standard for a short time, relative to a similar rig hooked up to a smaller screen.

As to your request, THG usually includes your resolution in their benchmarks. You can judge for yourself whether your proposed mini-ITX, SSD laden, high-res gaming, $750 rig is feasible.
 
[citation][nom]youssef 2010[/nom]I think I'd really want to know what a $3000 PC looks like now.[/citation]

That wouldn't be very practical for a gaming build's budget, so it'd probably be built more like a very nice workstation that has great gaming performance on the side.
 

We have looked into balance along these lines, pairing multiple CPUs and GPUs. And you are right, it takes a crazy amount of lab time! And it's impossible to keep up with driver and game patch releases, that can and do change the outcome over time.

But the problem is, (and reason for my comment) there is no exact science and no set in stone right recommendation. It varies based on the games, resolutions, settings, single/multiplayer, and even each individual gamer's tolerance for framerate drops and variance. You've got to research your games, know your desires, and then grab the most value you can from parts that will for sure be good enough to meet your needs. Monthly "Best for the Money" stories will help you there as they take changing trends in hardware demands and pricing into account. One example.... dual-core Pentiums have been high value gaming CPU's, but they aren't looking too good in games like Crysis 3 right now. Hence, recommendation will adjust accordingly over time as demands change and parts appear less worthy or future-proof.

Anyway, here are some old stories below that represent the type data you are asking for.

http://www.tomshardware.com/reviews/build-balanced-platform,2469.html
http://www.tomshardware.com/reviews/balanced-gaming-pc-overclock,2699.html

Check each game, and you'll see how the balance needed shifts around a lot. But do not get hung up on the numbers, or the colors. NVidia was behind here in part 1, with more CPU overhead and thread dependency from their drivers. But hundred of hours of testing after part 4 had AMD way behind in CrossFire scaling... with higher CPU overhead, the opposite of what is seen here in Part 1 with dual-GPU cards. Yet, Catalyst driver improvements over time addressed this issue, bringing the picture back closer together. (and killing the data, so to speak). Driver maturity and the individual game is key. Remember, any one driver version is a snapshot of how the situation looks at one point in time. The situation often can, and to some degree will, change over time.


 
Status
Not open for further replies.