System Builder Marathon, Q4 2012: $500 Gaming PC

[Fulgurant]

[citation][nom]mikenygmail[/nom]Yes it is, and all of this silliness is the result of not including the cost of Windows.Not including the cost kind of makes the entire thing seem silly, since Windows is not free.The article is great otherwise, although an unlocked, overclocked CPU would obviously be a better choice.[/citation]
Why stop at Windows? Why not toss in the price of a monitor, a keyboard, a mouse, a mouse pad? Hell, desks and chairs aren't free either. We should include them too!

I sympathize with your point of view; I really do. For a long time I would have agreed with you, but I just don't see the point anymore. It's obvious that the SBMs are mere guidelines; it's obvious that they don't include Windows. Anyone looking to build a new rig ought to read through the article carefully enough to see those things.

Likewise, anyone looking to build a new rig, unless they're heavily pressed for time, shouldn't mindlessly follow the SBM's recommendations -- especially not the $500 SBM's recommendations, because harsher budget constraints require more care. As the saying goes, you can spend money or time.

Incidentally, Anandtech does include the price of Windows in their System Builder guides, but Anandtech uses different budget guidelines, so the point is moot. At the end of the day, their approach is no more helpful to anyone willing to read the whole of each site's articles.

 

[ceh4702]

So let me ask a question. Does this processor slow down the video card? Can the processor really keep up with the video card? Is the throughput high enough?

 

[mikenygmail]

[citation][nom]fulgurant[/nom]Why stop at Windows? Why not toss in the price of a monitor, a keyboard, a mouse, a mouse pad? Hell, desks and chairs aren't free either. We should include them too!I sympathize with your point of view; I really do. For a long time I would have agreed with you, but I just don't see the point anymore. It's obvious that the SBMs are mere guidelines; it's obvious that they don't include Windows. Anyone looking to build a new rig ought to read through the article carefully enough to see those things.Likewise, anyone looking to build a new rig, unless they're heavily pressed for time, shouldn't mindlessly follow the SBM's recommendations -- especially not the $500 SBM's recommendations, because harsher budget constraints require more care. As the saying goes, you can spend money or time.Incidentally, Anandtech does include the price of Windows in their System Builder guides, but Anandtech uses different budget guidelines, so the point is moot. At the end of the day, their approach is no more helpful to anyone willing to read the whole of each site's articles.[/citation]

Very funny but wrong. We're building a box and Windows is stored on the hard drive INSIDE the box.
Those things you listed are not in the box, and your comments are entirely "outside the box."

Windows is needed to run the games benchmarked in the article.
Without it, benchmarks should only be run using totally free, simple to use software.

Thanks for the sympathy though, I think. I was not aware of anandtech's article including Windows but that would only serve to reinforce my point.

You type well, which is greatly appreciated.

 

[Onus]

[citation][nom]pauldh[/nom]Getting back to earlier comments regarding games, many of you have noticed the drop to three...I'm hoping to generate some data on a 4th before we box these rigs up, and you'd see that data comparison next quarter. We’d have to move on this quickly, and are open to suggestions: Assassin’s Creed III, Guild Wars II, Far Cry 3, Hitman: Absolution, any others?[/citation]
I'd like to see GW2. Not only is it a game I play, but it is a different type (MMORPG) from the shooters often tested, and supposedly more CPU dependent than a lot of games. Getting a consistent benchmark could be difficult, but I might suggest using a personal storyline instance in which there are no other players.
To expand on that, when testing CPUs, it would be interesting if there were repeatable performance differences between a Necromancer, using a number of computer-controlled minions, and a Warrior, going in solo just swinging his oversized mallet.

 

[mikenygmail]

From years back...

SCSI = Super Costly Sucker Interface
IDE = Inexpensive Direct Easy

Anyone care to add an acronym for SATA?

 

[Fulgurant]

[citation][nom]mikenygmail[/nom]Very funny but wrong. We're building a box and Windows is stored on the hard drive INSIDE the box.Those things you listed are not in the box, and your comments are entirely "outside the box." Windows is needed to run the games benchmarked in the article.[/citation]

A keyboard, mouse, and monitor are needed as well. Depending on who you talk to, any number of other incidental accessories could be required for any given build. The point is that the line has to be drawn somewhere. You're welcome to disagree about where Tom's draws that line, but it's not a big deal, and it certainly isn't a cut-and-dried matter of right or wrong.

The guidelines of the SBM's are written clearly. As long as the consumer knows what he might or might not need to add on top of the costs listed here, there's no reason to complain.

[citation][nom]mikenygmail[/nom]Without it, benchmarks should only be run using totally free, simple to use software.[/citation]

Fair enough -- and without including the cost of a keyboard/mouse, Tom's SBM's should only run benchmarks that users can control with their minds.

 

[rwpritchett]

Windows is needed to run the games benchmarked in the article.
Without it, benchmarks should only be run using totally free, simple to use software.

So then... I guess Tom's needs to include the price of the software they are using in the benchmarks then too, eh? Battlefield 2, Skyrim, F1, they all cost money. Perhaps free demos could be used, but even the synthetic benchmark software isn't free. 3DMark costs money to run the tests that are listed in this article. Sandra and PCMark are the same. Tom's isn't using free Lite versions of the benchmark software.

Plus, one of the parts of the SBM is comparing how much computer you get for your money with SBMs from the past. If you start adding software costs then you break that.

 

[blazorthon]

So I built this basic build 10/28/2011...... I used the G850, and a GTX 480 substituted for the HD 7850. It actually exceeded the performance of this build because the GPU ended up overclocking by 30%.

I remember back then everyone was like "You want the I3 for games not the G850 because _____" In reality the money saved going to a better graphics card made the same perfect sense back then that it does now.

Seriously disappointing that they are just now starting to come out with build truly optimized for gamers.

A GTX 480 with a 30% overclock would perform almost exactly the same as this build's overclocked Radeon 7850.

 

[blazorthon]

So let me ask a question. Does this processor slow down the video card? Can the processor really keep up with the video card? Is the throughput high enough?

In many situations, yes, the CPU is a bottle-neck, but it's generally fast enough to provide a good experience anyway.

 

[bak0n]

[citation][nom]Crashman[/nom]Run Linux, this is a hardware shootout.[/citation]
Then benchmark using Linux not windows.

 

[Azn Cracker]

a 1 tb drive is only $5 more....

 

[mohit9206]

a 1 tb drive is only $5 more....

thats not necessarily how it works in other countries

 

[army_ant7]

As I was going through the comments, I was planning on replying to certain posts, but as I went further, I saw a lot of people who shared my thoughts on different matters, already spoke up, so I just +1'ed them. One notable one is fulgurant. I don't find your logic flawed at all, plus, as mikenygmail has pointed out (I think), you do seem to have good typing habits. :lol:

As for you mikenygmail, even though I won't make any accusations, what you're doing comes off as trolling to me already (sorry to say), and only just because fulgurant and others before him/her have already pointed out undeniable facts against your claims, yet you still push on. Not that I want to quell your drive to share your thoughts, but please do give more thought to what their saying. For one thing, again, as fulgurant and others have pointed out, the issue your presenting is one of a subjective nature, and I agree with what he/she said, as long as it is made known to the readers what is accounted for with the given price ($500), it's still useful info. Benchmark data with specific settings on a specific hardware system is good as long as the details (what those settings are and the hardware parts are) are made known. From there, the reader can choose what to make of that data.

The style of killingchicken (or whatever his/her name was) came off even more as trolling (to me), just because of the way he/she made his/her comments.


Anyway, that aside, I would've really been happy to see what the Athlon II X2 750K could do, had it only been available in the US. Interestingly though, it could be available in other countries already as I learned from mousseng's post somewhere "above," about a polish website having benchmarks of it. Though I didn't bother checking the website, I too would want to see TH benchmark it, basically because I trust TH more (not necessarily the most rational basis for such a choice, but hey, preferences are preferences ). Maybe they're supplying them first in countries of less wealth first since there might be a bigger market there for budget CPU's.

...getting Windows for free legally is easy if you care to do it. Dreamspark has many free versions available to college students and most people know at least one, even if by proxy. Even in the unlikelihood of not knowing any, there's still the eval copies that MS gives away for free on their own website.

I'm just curious blaz... Would it still be "legal" (by the EULA or anything) for some college student to take advantage of that program (I have heard of it before.) and install it on someone else's PC when it wouldn't clearly be used for the college student's benefit in relation to college life?
Though I understand you mentioning evaluation copies to support your point, they might not be as reliable in the long-term, since I'd think they'd expire eventually, aside from the fact that it would be dishonest to use an eval. copy if (and just if) you aren't considering buying the retail copy, IMO.

 

[Guest]

I would like to see one more catagory added into the system builders marathon: Hackintosh. It is difficult to find out what the best/current products to choose for a good Hackintosh that is up to date, and a regular build of the most problem free components with a writeup on the results would be very helpful, expecially from a trusted source like Tom's.

 

[mousseng]

I would like to see one more catagory added into the system builders marathon: Hackintosh. It is difficult to find out what the best/current products to choose for a good Hackintosh that is up to date, and a regular build of the most problem free components with a writeup on the results would be very helpful, expecially from a trusted source like Tom's.

While there are a lot of great resources for Hackintoshes already (Lifehacker, Hackintosh.com, tonymacx86, etc), I think having Tom's do something with the idea would be wonderful. I don't think it would mesh too well with the SBM, given the specific benchmark suite and its incompatibilities, but it would certainly make a great article (or article series, if it were popular enough). There's definitely plenty of content there for them to work with.

 

[Guest]

[citation][nom]killerchickens[/nom]$501 Plus $100 for a copy of windows 7.[/citation]
who buys an operating system anymore?

 

[army_ant7]

who buys an operating system anymore?

Non-pirating system builders and people who just want to change their OS. Who else?

 

[Guest]

Add $100 US to the system price since the OS wasn't included in the cost. And note, video card requires a minimum 500watt power supply and they used a 450watt.

 

[Crashman]

Add $100 US to the system price since the OS wasn't included in the cost. And note, video card requires a minimum 500watt power supply and they used a 450watt.

Read the article, then slap your forhead...repeat...delete.

 

[lunyone]

Add $100 US to the system price since the OS wasn't included in the cost. And note, video card requires a minimum 500watt power supply and they used a 450watt.

Well you must not be up on your PSU's, so here's a couple of examples (one good and one bad). All PSU's are not made equally.

*** Both reviews were performed by JonnyGuru.com, so you will have similar testing methods (in order to minimize variables).



*** Conclusion and Scoring:

Page 5 - Scoring
Performance (40% of the final score) - let's come up with some scoring now, shall we? This shouldn't be too tricky. We have that better than 1% average regulation, so no points come off there. I'm passing it for Gold, so no points off there. Where I do have to make a deduction is ripple suppression. I have no problem with the minor rails, but the 12V rail isn't quite as good as I want to see from the really elite units. At the 50mV mark, give or take a few millivolts, I usually remove half a point. So, I do believe we're going to see a 9.5 here.

Functionality (20% of the final score) - here, we start off with an automatic full point deduction due to the fully hardwired cables. No exceptions, I'm doing that for everyone these days. Normally, I'd be removing some more pointage for the PCI-E cabling as well. We have only one six pin on this unit, while the unit's capacity is right on the edge of being able to handle the slightly more power demanding cards. That said, I'm going to let it slide, this time. Like I said, it's right on the edge and you really should be looking at the 400W and up range to power one video card needing the kind of power that requires the 8 pin connectors. So, I guess I really have no more complaints, here. The cables aren't too long, there aren't too many of them, and I really like the shallow depth of the unit's housing. Let's do a 9 here.

Value (20% of the final score) - thanks a lot, Seasonic. I was taking a drink of Coke when I saw Newegg's $59.99 price tag, and now my nose is on fire from the inside out and my keyboard really needs cleaning. Good grief... sixty bucks for this kind of performance? You've got to be joking. Yeah, you can find much more powerful units there for the same money. But look at those units... many of them are much cheaper designs that don't have a hope in hell of matching the numbers this wee beastie can manage. That said, the price is just a tad on the high side. But only a tad, and the performance more than makes up for a lot of it. 9.5.

Build Quality (20% of the final score) - Seasonic's upped their game when it comes to build quality recently, and it shows in this unit. I can't fault anything about it. 10.

Performance

9.5

Functionality

9

Value

9.5

Build Quality

10

Total Score

9.5



Summary

Seasonic has really been getting things done lately. First, the X series came and impressed. Then, the X series broke the 80 Plus Platinum barrier and impressed again. Performance was tweaked, resulting in huge 1250W units with near perfect regulation and ripple suppression. Now, they've taken everything awesome about those bigger units and downscaled them into smaller and more wallet friendly units. Is there room for improvement? There always is, yes. Does this unit need improvement? Nope. It's already a better supply than anything else at this power level I can think of right now.

The Good:

Gold efficiency
outstanding voltage stability
shallow housing
very good ripple suppression
The Bad:

I wanted to complain about something, but Seasonic didn't leave anything to really complain about
The Mediocre:

is there any real competition for this unit?



*** Conclusion and Scoring:

Performance (40% of the final score) - on one hand, there's really not much to get excited about here. The performance of this unit was exceedingly bland by my standards. Regulation was just ok. Efficiency was just ok. Overshoot transient testing showed merely average turn on spike suppression. But there are two things that really stand out at me. First, this unit could use some more ripple suppression. No matter which of my two review samples we talk about, ripple on the 3.3V rail went out of spec, and I have to consider that in the score. Since my original sample has been repaired and is again outperforming the second, I'll give Seasonic and Arctic Cooling the benefit of the doubt and go by the first unit's scope shots here. Out of spec is still out of spec, however. The second outstanding factor I want to consider is how well that first unit performed with only one 12V output part soldered in. Can't ignore that either, because it means that the 550W peak rating should be attainable. I wouldn't do it intentionally, because the wiring is a little thin gauge for that, but it's possible.

So, since this unit seems quite average in performance, I'll start out with a middle of the road 5. One point comes off for ripple. One point goes on for good design. What's that spell? F-I-V-E. Fair enough? I think so.

Functionality (20% of the final score) - there are some rather questionable decisions made by the Arctic Cooling crew here. First, we have cabling so short you will likely need a small case to house this unit in. Second, why an ATX12V but no facility for EPS12V? Many boards need such a connector these days. Third, the PCI-E cables are all differing lengths. And all the SATA connectors are on only one big long chain. What for? And what's up with the wiring to the fan? There has to be a better way to get power to it than that long cord that runs down to the main cable exit and into the unit. That said, there are some cool things going on here. We have a couple of fan only connectors, which is unusual in units these days. We have that F8 fan, which really is a quiet son of a gun. And we have enough PCI-E connectors to power two cards, even if the unit itself may not be able to take the most power hungry models. I think I'll go average again here. 5.

Value (30% of the final score) - since I cannot find a vendor in North America for this unit, I'll go by the MSRP for it which is $93.90 US. Yeah... ninety-four bucks for average 500W performance? I don't think so. This might start to compete at twenty bucks cheaper. 4.

Aesthetics (10% of the final score) - this unit is quite good looking to me, aside from a few obvious flaws. First, that blister packaging is not doing the unit any favors. Arctic Cooling would do well to toss the units into a bag before cramming them into a hard plastic finish marring shell. Second, I don't care for the kludgey look of the fan's power wires. Just drill a hole next to the fan and run the wires into the housing there, please. It'll look a lot neater and cleaner. 8.

Performance

5

Functionality

5

Value

4

Aesthetics

8

Total Score

5

Summary

Average power supply is average.

The Good:

looks good
silent
handles hot boxes well
decent regulation
lots of PCI-E connectors for the power level
has two fan control connectors (use low power fans on these)
The Bad:

not a 550W unit as implied by model number
thin gauge cables
all SATA connectors share one long chain
high MSRP
fan power cabling
The Mediocre:

3.3V ripple suppression needs a fair bit of work


So to put it lightly, do your research before you make accusations/statements about PSU requirements. I'm not an expert, but I try and make informed decisions about the parts that I buy. I have been bitten by cheap (not well made) PSU's in the past. You can have issues with better made PSU's too, but they will be warrantied better and will have a smaller tendency to fail than lesser quality ones.

 

[lunyone]

Links didn't go through, so here they are:

 

[lunyone]

Wow, fail x3, sorry!
Seasonic 360w
http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story5&reid=313

Arctic Cooling 550w
http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story5&reid=209

 

[army_ant7]

@lunyone
It's good that you shared the info, but it might've been better just to paraphrase or pick out specific parts from the links instead of copying huge blocks. Whoever may be interested could just go to the links you provided themselves.

Just some friendly advice, and again, it's good you're being helpful.

 

[lunyone]

@arm_ant7
I know that it was a bit much, but a lot of people are too lazy to click on links and then try and find the final score/results. I wasn't expecting all of the spaces in there, but was just trying to get to the point fairly quickly.

 

[sentack]

I just want to verify that, if I use the same model of graphics card but get the 2GB version, the same overclocking values should be about as good? Of course I need to test everything but it sounds like it should be fine, really.