Building the Stepper Gaming System
By Edward “Computer Ed Crisler
When it comes to building a gaming computer a lot of people are turned off by it due to the cost. They think they need to spend a lot of money to enjoy the PC gaming experience. I wanted to prove them wrong and so last week I set about to build a low cost gaming rig. Since I love building in SFF I did this using the Node 304 case and a Gigabyte F2A85XN motherboard. To ensure the best results I opted for the high end of the APU lineup and used an A10 6800K to which I paired 8 gig of Kingston DDR3 2133 RAM. I topped off the build with a Samsung 120 gig base 840 SSD and ended up with a smoking quick little computer.
For testing I made a few presumptions. After all this was a budget build so that meant little money. I presumed I had thrown all my money into the computer so I started firing up F2P games. I chose Neverwinter, Star Wars, DC Universe, League of Legends and then added a beta FPS game I love, Loadout.
My goal in this testing was to explore the play experience. While I did use benchmarks during the testing they where for base data, not the end data results like so many sites use. Benchmarks are all well and good but actually using the computer is all that matters at the end of the day and I was striving for a smooth playback with decent graphics. I am a bit pickier than most so I figure my judgement on this will give more casual gamers great results.
I kept my testing realistic and started at the high end, 1080 resolutions with the games set to medium detail level within game sliders being used. I made use of the default settings to keep this simply. The experience was really solid with all of the game delivering good playback. I did see some stutter in super congested areas within the MMO’s but during the combat or in areas where there was risk of combat the playback was very smooth. League of Legends and Loadout where nice and smooth the whole time and gave good playback.
Next I dropped to 720 resolution and cranked the detail to high, this to me seemed a more realistic setting for this type of setup. At these settings the games had outstanding play with only the barest noticeable stutter in super congested common areas in the MMOs, nothing however to really harm the play experience.
With a total cost for the build of under $600 I think I achieved what I set out to do. A term used a lot today is a Steam Box and this build qualifies. It was super compact and could fit on an entertainment center easily, making use of the TV as a monitor. The use of 720 resolution might be snubbed by many enthusiasts but it does give a solid gaming experience.
Now I could have stopped here and been happy with the results but I wanted to see if there was more. The 6800K is a fully unlocked APU so it can be overclocked. Pushing the system to 4.6GHz with the iGPU at 1 GHz I got a nice boost in my frame rates but the overall experience stayed about the same. However I was curious at how much kick this little brute had, so this started me thinking. The scenario is that you have built this $600 PC and have been gaming on it a bit but you realize you have been bitten by the PC gaming bug and want more. You want 1080 gaming with high detail and you want some of the more advanced games.
Since there is a PCIe slot on the board we are using we do have the option of putting in a high end video card. However I am always hearing people in enthusiast forums talking about how a high end card on an APU is a waste, it will bottleneck the card. Well that is our first test, will it bottleneck the card?
To test this fully I decided to go way over the top. I put a 7970 GHz edition with 6 gigs of memory on this APU based system. Now this is a lot more GPU than this type of build will likely ever get but I wanted to see if the APU was bottlenecking the high end discrete GPU. To do this I needed a baseline and so I used a Haswell i7 4770K clocked at 4.3GHz with 16 gig of Kingston DDR3 1600 memory and the same video card and drivers.
For testing I used 3DMark Firestorm and was watching the GPU score, if the APU was bottleneck the GPU this score would reflect it. This test is pushing 1080 resolutions so I felt it was a perfect pairing of test and GPU to see if the APU was bottlenecking.
Now the scores that rely on the CPU went to the Haswell as expected. However my first surprise came when I saw the overall 3DMark. The Haswell, in overall score was only 12% faster. I ran these tests three times to verify. 12% overall is NOT a huge difference, especially when we are talking about a 236% price difference for the chips.
As if the overall scores where not surprise enough the graphic scores where nearly flabbergasting, the APU stayed within 0.8%, that’s right less than a single point of percentage. In fact in one of three tests it outscored the Haswell by 0.1%. This is pretty clear that from a raw power point of view at 1080 the APU is NOT bottlenecking a high end video card. If my testing was not enough for you how about AnandTech? In an article they posted on 1440 gaming they found that for the majority of games the A10 APU was the best buy for single video card game play at 1440. They did this BTW on a 5800K so they used a slowed base speed APU at stock speeds vs my overclock.
What about actual game play though? Well with the bottleneck issue proven to be a myth I dropped down a little in video card, I wanted a more reasonable build price. With 7950 cards now being seen near $225 that seemed a good point and still gives a higher end card. Next I started firing up some games, this time though I did some stuff with a bit more kick. I used a modded Skyrim, Crysis 3, Far Cry 3, Borderlands 2, Bioshock Infinite, Civ V and our choices from the F2P crowd as well.
Now again the actual game play was our goal and in all but Crysis 3 I was able to run at High default settings and get great game play at 1080. For Crysis 3 I did step back down to medium but even in medium Crysis 3 looks AMAZING! Now in fairness the game ran good at high but the playback was better at medium and so that was what I would suggest. In all the other games on high the playback was smooth and nice, enjoyable game play with no noticeable stuttering.
The results are actually pretty amazing. With a system costing less than $825 I was able to achieve the same gaming experience are systems costing hundreds of dollars more. What is especially cool is that I could build the system at a lower cost, have fun with it and then just put in a video card to kick it over the top. None of the parts I bought where replaced to do this, just a GPU added.
This testing proves to me that the AMD APUs are a lot more than just a budget based chip. Sure they can build a budget system and in my opinion if you are building a system that is using the integrate GPU the APU is the ONLY way worth going. However they are not done there, this budget build can give a solid platform to take your gaming experience to the next level. The high end enthusiast might look at other options for their build, but for the majority of users, the APU is the place to start.
BTW if you want see more than a discussion on this, head over to the LANOC event on September 7th. I will be there along with a slightly modified APU build as discussed here. Come join us and see it in action.
A special thank you to all the companies that provided the parts used in this article.
Show Segment Aired Weekend of August 24th