System Builder Marathon, Q3 2013: $650 Gaming PC

Page 4 - 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.

pauldh

Illustrious


We use other stories to compare parts straight up. SBM's are all about the best performance and value only. Each builder can pick from anything Newegg has in stock and none of us care lick about brands. But we must be able to justify our picks based on bang, stability, value, etc.

Building one machine per quarter each is a limitation as we want to represent varied price points too. And It's tough to sway our budgets each round, tackling low prices then high prices etc., as we lose meaningful comparisons between our own builds. Q1 2013 was a rare exception where we heightened the competition at low prices, and addressed the high end with a bonus build.

The FX-6300 is pretty much perfect at this budget, and we'll be using feedback to see if we should remain at these budgets or drop to $500/1000/2000 again. Don tried an FX-8350 enthusiast PC and it did poorly, so I doubt he will revisit Vishera at $1K+ again. So it's up to me, and even $650 can be quite limiting.
 

slomo4sho

Distinguished


Maybe have a silent/efficiency theme in the next marathon?
 
I'd like to see more compressed budgets, starting lower, e.g. $500, $800, $1100; or $550, $750, $950. In addition to comparing against previous cycles, you also get to address the difference another $200 can make.
Another alternative would be to do a SBM, then do a 10% or 15% of budget upgrade and retest to show what the extra buys.
Paul, I like your explanation of the mild overclocking in this build; let's just hope the winner knows that adding a $30 fan is not going to allow a higher OC without serious risks of killing the mobo.
 
Its still a bit disappointing the Toms build with intel processors by default .......but wait nearly a year before they will try building with an AMD chip .... and when they do they find its a better option than the similarly priced intels they have been using .

But thank you Paul for finally getting around to this . Its the build some of us were recommending all this time because we always knew it how superior it was .
Now all you need to do is eliminate the old poorly threaded games you use in these comparisons . They skew the results in intels favor and almost no one ever turns up on the mb asking for a build for skyrim
 

InvalidError

Titan
Moderator

But that would be the same $100 regardless of whether it is a $500 or $5000 build. People aren't particularly interested in items that are the same across the board or expected to be scavenged from what they may already have in stock such as screens, mice, keyboards, etc.
 

pauldh

Illustrious


No problem, I've been looking forward to it.:)

Please do keep in mind my above & past comments. Based on our budgets, Vishera falls in my lap. Don's not going to touch it. At $140 it was less attractive and didn't fit in a $500 budget. Last quarter was Mini-ITX.

Q1 at $600 was really my one shot to do FX-6300, and I chose i5 instead, which I'm still convinced was the smart move to win that heightened competition. I barely squeaked a win as it is. 4.5 GHz FX-6300 can't beat 3.4-3.7 GHz i5, we've established that on the test bench. It could have secured a 7870 had I done a mild OC, but that helps in such a tiny % of the weighting overall. So, up until now the timing/theme/budgets has been wrong for FX. I've never questioned it would spank i3 in apps and game well. Just took a while to show it in an SBM.

As far as Skyrim, personally I think it's a great game to test. It is still #7 most played on Steam. Like it or not it also represents how a large bulk of games still act. Canning it to favor an architecture is not doing readers a favor. I'd like to ditch single player BF3 before Skyrim. Already tried to replace it with Crysis 3, but was shot down. That's a hard/time consuming one to bench and all get consistent comparable results.
 

pauldh

Illustrious


Hmm, is that all three of us? We do try to explain the goals and thought process for each part. Thomas has the best chance to secure parts Tom's has approved. Value parts are rarely tested, so I have more digging to do. Unfortunately, it often come down to price on the given day for me, and choosing among a small handful of the best priced options.
 

Agreed. "Why" is almost always the most interesting (and most important) question to explore.


Ah, so I take it you're doing the $350 build? Excellent, I will look forward to it.

 

InvalidError

Titan
Moderator

I think that would require a completely separate set of entries and price brackets since I doubt it would make sense to have a $2500 enthusiast HTPC.
 


The issue is officially x87 is a decade dead coding language... officially neither intel nor amd are supposed to support it. That said intel has kept unofficially supporting it, where as AMD has not. this grants a massive (artificial) advantage to the intel cpus in the few programs that happen to use it. (AMD's fault for cutting it out i guess)

Unfortunately for AMD, skyrim, civ5 and sc2 ALL use it. All three are hugely popular titles. The problem isn't that THG uses skyrim, civ5 or sc2, it's the number of titles you use when doing an average performance matters greatly. if 3% of games developed in 2013 use x87 code, and you only sample 4 games with one using x87 core you heavily weight the average fps results in favor of the intel cpu. Now an argument could be made that those 3 titles are so huge that they make up 25% of the pc game sales... and if this is the case then i have no issue with using it as one of 4 games used for your average FPS numbers.

but i think we all know SC2/Skyrim/Civ5 do not make up 25% of game sales, nor does 25% of the new title market use x87 code. In the interest of statistical accuracy you guys should be adding titles to your benchmarking lists if you mean to keep an x87 representative in the mix, and the mix of games representative on the whole (so if half the games are single threaded junkers half your benching list should be as well)

I think we all know Intel has the better cpu, so in the end even if you balanced out your sample the intel cpus might be ahead by as much or more in the final analysts... the statistician in me would just like a slightly better cross section representation of the gaming market.
 

pauldh

Illustrious


Yep, I like the close/low budgets also, although it alienates those who want a true high-end system. We did that first quarter and ended up with very similar parts.

Good point. While this story was never meant to tackle that, I should have been more clear on the required upgrades for higher OCs.

This would be more like it:
http://www.tomshardware.com/reviews/piledriver-k10-cpu-overclocking,3584-2.html
 

pauldh

Illustrious
ingtar, I pushed for 6 games, it's not going to happen. These already cause some hectic scheduling to coordinate.

But by you own standards there, our suite is already too well threaded to represent averages. We'd need to include an MMO representative like LoL or Dota 2, and an old title, etc. If we made the sampling huge enough to truly represent PC gaming as it stands today, a dual-core G2120 would probably go toe to toe on averages with FX-6300. Correct?
 


I agree completely. personally i would add guild wars 2... which is modern and notoriously single threaded; especially in world battles. or you could add shogun 2 or rome 2, both of which are much more representative of modern games then skyrim or civ5.
 
I also believe this is the best of the three builds so far when it comes to bang/buck, but one must be aware of the potentially harsh limitations of its motherboard.
I appreciate the digging that the budget builder has to do. I've been on the forums long enough to have had my own ego variously bruised or stroked over my suggestions. The bruises I've mostly forgotten (except perhaps if there was a lesson in them), but what I remember as the best stroke of all was when someone looked at one of my builds and said of it that I pinched a penny until Lincoln cried.
I maintain at least one list of the absolutely cheapest build I could honestly recommend, which I revise from time to time and then edit and use to answer specific requests. I should post the unedited one(s) some time. I'm really looking forward to your $352 build for comparison.

Edit: Even though I haven't played it lately, I'd love to see GW2 in more articles and benches.
 

jbird2383

Honorable
Sep 7, 2013
25
0
10,530
No sense at all to go with a stock cooler on a 6 core chip when all types of builders, gamers, know how much a priority and important it is keeping your components cooled especially when newegg runs deals all the time for the Evo and 212+ for the 10 to 30$.

The money you save from selecting the Radeon 7950 and going with a 4 GiG ram kit vs a 8 gig kit I would put towards a aftermarket cooler like the Evo or 212+ and whatever was leftover and I would consider a beffier PSU. I like the Cooler Master 922HAF case vs the NZXT as it is more expandable and up-gradable.

Ram is an easy upgrade "down the line" and you can always catch a good deal and starting out at 4gig isn't that bad.
 

pauldh

Illustrious
Hmm, I know this MSI was the most affordable 970 when selected. And really, given I knew I wouldn't raise voltages, that worked for me. iirc there was only one other under $80 and I think it was from ECS. But I'm not positive on that as we suffered a very long delay this quarter. Second cheapest isn't in my notes, and it's not fresh anymore unfortunately.

That Biostar looks promising though in CPU support and VRM heat sinks. Onus, ugh, yeah I had a few of their old VIA KT boards come back to bite me hard in time, but they've had some real Gems too since then. I'd have no problem picking that board.
 

ojas

Distinguished
Feb 25, 2011
2,924
0
20,810

True, true. I over-simplified, i admit.

Not complaining though, i really wanted to see the FX-6300 in a proper rig rather than just a test bed.
 
Status
Not open for further replies.