System Builder Marathon Q4 2015: $912 AMD LAN Box PC

[Onus]

Normally I come down pretty hard on builds that don't follow the SBM rules, but in this case I think it is relevant to the discussion of heat [from an AMD system] in a small case.
A problem you will definitely have though is risks to the VRMs. A liquid cooler on the CPU neglects those, so airflow must be exceptional. Furthermore the issue is exacerbated by the desire to get a good OC, common to all of the SBM builds where overclocking is an option.

 

[RedJaron]

sorry to dispoint, at least you know how to put a person down. should come in handy one day, i dont expect on this forum though

Don't play the victim. No one is attacking you. Why is it ok for you to tell others they're doing it wrong and you can do it better, but then it's rude for us to point out the problems in your arguments? Warmon covered the rest. I don't understand what that 8350 mATX part list has to do with this quarter's SBM bonus build or FM2 ITX builds in general.

 

[Onus]

sorry to dispoint, at least you know how to put a person down. should come in handy one day, i dont expect on this forum though

Don't play the victim. No one is attacking you. Why is it ok for you to tell others they're doing it wrong and you can do it better, but then it's rude for us to point out the problems in your arguments? Warmon covered the rest. I don't understand what that 8350 mATX part list has to do with this quarter's SBM bonus build or FM2 ITX builds in general.

It doesn't, directly, and as a Mod I've often deleted such remarks as irrelevant in past SBM threads. In this case, there appears to have been some initial confusion about the form factor (since cleared up), but most importantly, I thought it actually added to the discussion regarding potential heat issues, so I left it.
At this point though, I think we can all recognize that your SBM build was bound to the ITX form factor, so we can now leave mATX (and AM3+) out of it.

 

[Crashman]

I don't understand what that 8350 mATX part list has to do with this quarter's SBM bonus build or FM2 ITX builds in general.

I think we can all recognize that your SBM build was bound to the ITX form factor, so we can now leave mATX (and AM3+) out of it.

Great, cake and ice cream for everyone!

 

[RedJaron]

I think we can recognize it was a big disappointment as well ( at least it was to me ). I was a little dismayed until I found one useful data point in it: the uncertainty of overclocking and the extra time and effort to get parts that will work well together should not be dismissed when adding up the cost of a "budget overclocking" build.

Oh, and I thought the cake was a lie.

 

[Crashman]

I think we can recognize it was a big disappointment as well ( at least it was to me ). I was a little dismayed until I found one useful data point in it: the uncertainty of overclocking and the extra time and effort to get parts that will work well together should not be dismissed when adding up the cost of a "budget overclocking" build.

Oh, and I thought the cake was a lie.

You never came to get it. Maybe when I'm visiting the HQ you can come up for an expensed meal, cake optional at your discretion.

 

[alextheblue]

I actually have to disagree with both of you about these FM2(+) cpu's/apu not well suit for an ITX system. They're very capable in such system.

I said it's not WELL suited. I'll try to clarify. When you're talking about a budget overclocked FM2+ rig, you're going to get better results at a lower cost with a larger build. They're not trash in an ITX build, but they are less competitive with Intel - especially when combined with discrete graphics. That's more my point, and I think that line of thinking is indirectly supported by the review results. Oh and your thermal margin numbers don't mean a ton unless you run sustained whole-system loads with a discrete graphics card (which dumps more heat into your chassis) and graph your clockspeeds over time. I think you'll find you occasionally get throttled (or miss boost clocks more often if you use turbo settings) once you start pushing voltage up even if your reported margin looks OK on the surface. Your case is also as you said not the smallest. So it's hard to compare your results to Eric's for a couple of reasons, primarily integrated graphics and the chassis.

Within the confines of the selected components, I think Eric did OK. But yeah I think the 860K would only do really well in a lower price bracket entirely, a cheaper mATX/ATX build, sub-$200 graphics, Evo 212 cooler, EVGA or Corsair 80 Plus Bronze 600W, etc. The higher your total system price, the less competitive you get. That doesn't make it horrible, just not AS good.

 

[RedJaron]

The GPU only dumps heat into your case if you don't use a blower fan. I did, so that largely wasn't the problem. However, I'm really against the idea of using an aftermarket cooler on the 860K. The chip plus a $30 cooler is nearly the price of an i3, so why not go that route? If you want a cheap $10 cooler to help thermals and noise a little, I don't think that's too bad.

 

[cdrkf]

The GPU only dumps heat into your case if you don't use a blower fan. I did, so that largely wasn't the problem. However, I'm really against the idea of using an aftermarket cooler on the 860K. The chip plus a $30 cooler is nearly the price of an i3, so why not go that route? If you want a cheap $10 cooler to help thermals and noise a little, I don't think that's too bad.

The new cooler AMD have just announced sounds like it should remove the need for an after market as it's 125W rated with a heat pipe and a quiet fan. It's a shame it's taken them until 2016 to figure out how bad their old cooler was though...

 

[Onus]

I think we can recognize it was a big disappointment as well ( at least it was to me ). I was a little dismayed until I found one useful data point in it: the uncertainty of overclocking and the extra time and effort to get parts that will work well together should not be dismissed when adding up the cost of a "budget overclocking" build.

Oh, and I thought the cake was a lie.

I read build articles like the SBM looking for data points. The Munchkin build(s) provided some good data, so I'm not at all disappointed. Great job; what disappoints me is same-old/same-old, which has no lessons in it.

The cake was a lie, but crumbling up [waste-free] vanilla wafers into the ice cream more than makes up for it.

 

[RedJaron]

Only the last one felt like a disappointment to me. The first one, though weird, was pretty surprising when looking at the gaming benchmarks, especially the ultra detail ones. The second (Intel ) was fun because I never imagined I could beat the first by 20% value. But this third one . . . grrrr. Useful data and frustration aren't mutually exclusive.