tl;dr: PCIe 3.0 boards are faster and cooler than PCIe 2.0-only boards. 3.0 boards need EATX cases, 2.0 ones don't necessarily.
tl;dr: PS2 emulation is something I want to be able to do (due to scalability in resolution and other visual effects), but from what I understand, it requires a heavy-duty system, which is what I'm after.
tl;dr: Why the Storm Trooper?
callsignvega tested a four-GPU rig on the Asus Rampage IV Extreme, and found a 50-frame difference between PCIe 2.0 and 3.0. http://www.evga.com/forums/tm.aspx?m=15 ... print=true
jihadjoe had similar results, albeit ones with a smaller margin of about 10 frames. http://www.techpowerup.com/forums/showt ... 077&page=3
It appears that the difference is more significant with multi-gpu rigs at high resolutions, and less so with single gpus that don't require moving data generated by more cards across the bus.
The only quad x16 board: http://www.anandtech.com/show/6228/asro ... -sassata/9
This article tells me that it might be the dual PLX PCIe 3.0 setup that causes the lower speeds (compared to the other 2.0 and x8 boards) in the one quad x16 3.0 board on the market (as opposed to quad x16 3.0 support on a single chip. It also shows that speeds in 3.0 boards vary from model to model, and might not be completely influenced by 3.0 support.
This thread talks more about vega's conclusions, and comparisons between the plx and nf200 PCIe controllers: http://www.xtremesystems.org/forums/sho ... OUR-SLOTS!
Based on the responses, I want to try to stay away from the nf200 chip and go with a board that has a PLX chip, even if I'm not going with a multi-gpu rig. It sounds like the nf200 is hotter and has fewer lanes.
we'll gladly go on Newegg and pick the most expensive parts for you.
If you're not familiar with the high-end market parts, you could advise me with components from previous-gen hardware as a fall-back option. From there I can pick out a case that will match those parts. Expensive does not always equal performance; I'm ignoring price for now in favor of figuring out what the baseline for performance is. I can always work backwards from performance, but I can't do the same thing with money due to diminishing returns. There are nine computers in the house right now, most of which I'm trying to get rid of for less than a hundred dollars since they are outdated by about a decade. So far, the best components I've seen aren't the most expensive (eBay sold listings at lowest price + shipping):
Rampage IV Extreme: $300
Corsair Dominator Platinum 16GB cl10 2666: $484 (Apacer's 78.AAGDW.9KD0C 2GB sticks are faster at stock, but I haven't seen any manufacturer retailers that list prices for the modules, or anyone who has tried to overclock them, so that's a bit of an unknown) (multipled by up to 2, since the Rampage IV Extreme only has eight slots)
Seasonic X-1250: $217.39
Magnum TX10-D: $923.08 (includes shipping)
GTX 680: $700 (multiplied by up to 4)
Phase change cooler: est. $950
Parts for phase change cooler for GPUs: $200
That comes out to $6873.47 at most, assuming I just shifted over my disk drives and didn't buy an SSD. I have about $300 worth of computer components right now, and the only reason I'm upgrading now instead of when the GTX 790 and Socket H3 releases is because my MSI P45 Platinum is apparently causing random reboots, and because no one will tell me what the best 775 board ever made was (which I think was the P45, but apparently its durability was not as good as other boards, if it is the cause of the BSODs and POST fails). I'm guessing that most 1366, 1156, 1155, and 2011 boards would be better than the 775 I have right now, and the older it gets, the cheaper it is to find the best components of their generation.
The problem here is that you need to define what you want.
I only have one game that will not play at all without better components, S.T.A.L.K.E.R.: Shadow of Chernobyl with the 2012 mod. It gets fps dives and lag spikes even at the lowest settings. It was made in 2007
. According to this article
, it looks like it is graphics-intensive, and a GTX 680 can't keep up with an lcd-standard 60fps on highest settings, even when overclocked. A 690 can meet the 60fps minimum, but if I were to get a 120Hz monitor, it still would lag.
I want to be able to simultaneously run Elder Scrolls IV with whatever the latest HD texture pack and render mods are, screen-record at 60fps with FRAPS, audio-record with Audacity, and stream in 1920x1080 at 60fps with OBS (haven't figured out how to get it working yet, partly due to weak system specs and bad internet; VLC or any other implementation of ffmpeg might be a better choice to merge the screen capture and streaming). I then want to be able to transcode the FRAPs codec to a lossless 48-bit color, i.e. YUV 4:4:4 so that it can be edited and merged with the audio in Premiere 5.5, then output to mpeg4 for YouTube use. So I've got three programs running at once, all highly CPU, RAM, and GPU intensive. It's going to take a lot of power and a lot of cooling to make it happen. If I can get that with a 1366 board with an i3, 4GB of DDR2, and a GTX 260, great. I'd prefer cheaper over expensive if
it will do the things I need it to do.
stick with your PS2
Since I don't own any consoles aside from an N64 and a Gamecube (not counting the Wii in the house), I assume you are referring to my Compaq 269513-006 PS/2 keyboard, which I use in place of my Saitek Eclipse since the desk I'm using isn't big enough for two full-sized gaming keyboards, and can only accommodate the Razer Black Widow and the standard-sized Compaq. I can assure you I will stick with the Compaq for now, or at least until we get a bigger desk. If I get the Magnum TX10D, I suppose I could turn it on its side and use it as the desk, or take out one of the pedestals, climb on top of the 6' case, and use that as the desk.
@Airheadq: Is there a reason why you prefer the CM Storm Trooper over the other cases? Is there some advantage in terms of cooling, adaptability, weight, or features that makes it better than the Cosmos II, or anything from CaseLabs? If you've done comparison research that led you to that conclusion, I'll read it like I did with the others'.