recyclebin

Thank you for the great input. After months of agony about upgrading, I finally bit on an i5-4670K (decent deal w/combo motherboard). Gonna keep my fanless HD6670 for now. My old E6600 system was very stable, but the upgrade itch finally got to me. I expect to do some mild overclocking (4 GHz or so I'm thinking). The Sony laptop I mentioned has recently developed worsening fan noise, so I'm wondering how long that will last. Not a huge issue, but it gave me the extra push to do the upgrade.

That's a good point. I didn't think about what graphics settings and part count were involved in my subjective evaluation of speed. The part count is minimal for the vehicles my son and I have been playing with, but I need to check on the graphics settings. I think for my PC with the HD 6670 I selected pretty high settings, but for the Sony laptop with Intel graphics I left it at default (not sure if that automatically varies based on KSP evaluation of graphics/CPU capabilities). I will try to experiment with different graphics settings on the Sony to see how much difference it makes. But in general it sounds like the current KSP will not make much use of a high-end graphics card. Thank you for the feedback. I'm addicted to KSP (within time constraints), and I admit that this is irrationally driving me to consider a PC upgrade. My ~6 year-old system (with various component upgrades over the years) is adequate for my needs, but my periodic upgrade itch has been stimulated partly because of KSP. I'm struggling with justifying an upgrade. My wife will question it, but that's not a deal-breaker. I'm mainly just cheap. My main justification for an upgrade is to pass my current system on to my son. That would be a large upgrade for him compared to the Sony laptop. But we don't let him actually use the computer much (especially for games), so I realize this is a flimsy excuse. Damn you KSP!

First post, but I've been lurking for a while and loving experimenting with KSP. The subject of graphics cards has been on my mind lately. I've been thinking of upgrading my system, and KSP requirements are part of that. I believe that KSP doesn't use any compute acceleration (for now), but what about the general 2D/3D acceleration that is part of DirectX (sorry - Windows-centric here)? Does anyone have an idea how much KSP will benefit from a more-powerful graphics card? A sub-question is which is better currently for KSP (v0.20.2 as of 10-Jul-2013) - AMD or Nvidia? I understand that integrated graphics like Intel 4000 will consume extra RAM bandwidth, thus reducing the available RAM bandwidth for CPU operations. But I assume that is pretty minimal, so it comes down to how does the fairly-poor Intel integrated graphics actually affect the performance of KSP versus a discrete graphics card? I am currently running an old dual-core E6600 system (overclocked from 2.4 GHz to 3.0 GHz) with 8 GB RAM and an AMD HD 6670 graphics card (fanless). This runs KSP well, at least with the pretty basic aircraft/rockets that I have been designing. If I upgrade I will of course go to a much more modern processor (Intel Ivy Bridge or Haswell), so the increase in CPU processing power will be assured. But what happens if I stick with my HD 6670 graphics card? This card, while somewhat old, is I believe still significantly faster than any current integrated graphics (Intel or AMD). I guess the integrated graphics issue isn't really relevant since I have no problem going with a discrete card. My real question is will upgrading from the HD 6670 to a more-modern discrete card actually provide any major benefit to the speed of KSP in its current form? My general impression is single-threaded CPU performance is king for the current version of KSP, but really how much does the graphics card matter? Nothing I've been able to find with searching has seemed to answer this question. Mostly people assume a more-powerful graphics card will be better, but there doesn't appear to be any quantification of what the KSP speed gains will be as we move up the graphics card ladder. As another data point, I have a Sony laptop with a 2.0 GHz dual-core mobile Intel T6400 processor (a "Penryn" model, one "tick" newer than my E6600, but probably not much difference other than clock speed and graphics card). This laptop uses the integrated Intel "GM45" graphics, which is I believe at least three generations below the current Haswell integrated graphics level. So the graphics performance is pretty pathetic. This laptop plays KSP poorly (stock, with no addons, and with simple airplanes/rockets that my 9 year-old son designs). The CPU clock is 2.0 GHz vs 3.0 Ghz on my desktop, but the serious reduction in KSP performance on the laptop seems greater than can be explained by the CPU clock speed difference. This seems to point to a major influence of graphics card performance, at least when comparing old Intel GM45 integrated to discrete AMD HD 6670 on a CPU that is 33% slower. I don't play many other games, so if I can be convinced that my current HD 6670 discrete graphics card isn't an issue for KSP, then I will be happy to keep it and save the money of a graphics card upgrade. Thank you for the input, and I hope this will provoke responses that provide concrete information on KSP performance vs graphics card type (understandably tainted by many other factors such as CPU model, RAM size/speed, hard drive type, etc). Now that I think of it, maybe I should start a new thread on this. But being a newbie I am hesitant to do that. BTW: this is something I just thought of - is there a "FPS" test in KSP that we can run to give an indication of performance under somewhat-controlled circumstances? I apologize if my E6600-era knowledge makes this a stupid and obvious question.