blkspade

You're almost guaranteed negative scaling with crossfire in DCS currently.

Having just made the switch to an i7 4790k, it did help to stay around the 30s-40s around the airfield most of the time. I too wanted to wait for Edge and happen to play ARMA, plus personally objecting to supporting Intel. It didn't fix all performance issues at stock settings, but I haven't overclocked yet. That said if your current AMD board has support for the FX-9590, then it should mean your VRM's can stand the load for a 4.7-4.8ghz overclock on the 8350 with better CPU cooling. An decent aftermarket cooler would cost $80-$120, and could transfer to an eventual future upgrade. I did this for a while, but the pump failed on my closed loop cooler and I had to switch back to my older aftermarket air cooler. I also had to disable a module while on water because my board couldn't handle the load of 8 cores @ 4.7 on water. Something like this http://www.newegg.com/Product/Product.aspx?Item=N82E16835608018 will likely carry you a long way on air though.

I think the idea is that even if you're instinct is to dive, its not usually with negative G. You'd roll and pull positive Gs. Doing so still puts you in the plane of travel of the flare in the case of the Su27, since its not "up" but from the top of the aircraft.

Switching to BVR mode has never automatically activated the radar. You can manually turn the radar on and off with its default key, which I think is 'I', regardless of HUD mode. Vertical scan is a radar mode itself, which is why it turns on the radar.

I wasn't referencing the AMD chipset, but the USB 3.0 host controller that was in use on my AMD board. ASMedia I do believe.

That has more to do with a USB chipset flaw than the actual devices. I had this issue on my AMD board and a USB Audio interface. It wasn't a problem on an Intel system i tried it on.

Well aftermarket coolers are generally universally supported on all current platforms, so can be moved across builds. If dealing with the pain of current DCS performance is an option anyhow, holding off other upgrades until DCS 2.0 might be the most economic approach.

As you may have missed, he's clearly operating in a budget of about $200 max. A GTX970 offers him next to nothing on his current hardware and monitor resolution. Moving to Intel would limit him to the maybe highest clocked i3 with the lowest end motherboard. That becomes a wash, since that i3 is only a quantifiable improvement in the generally poorly coded simulation genre. It would only be good if DCS is literally the only concern period. That budget leaves him between a rock and a hard place. Looking further into making the best out of that motherboard, would start with an aftermarket cooler and an overclock. The max typical OC for that chip is 4.5ghz. That wouldn't offer as much as the newer vishera at the same speed, plus the potential of actually reaching 4.8ghz. The catch is using a 6 core vishera for both reduced cost and lower load on the MB VRM. The 4 core would be better with respect to temp and vrm handling, but that steps deeper into side-grade territory like shifting to the i3.

While the piledriver CPUs are a bit better performing than the bulldozers (IPC), its not that substantial. The only way to get appreciable performance out of DCS with either is with a steep overclock. Now that in itself is only going to happen if you buy a decent aftermarket cooling solution. That comes in the form of the top Noctua cooler for air, or a closed loop liquid cooler all in the $100+ price range. Single player campaign crushes the AMD CPUs at stock settings in DCS. Considering you need ~4.8GHZ+ to get maybe in the 20s, it may not even be possible with your motherboard, not with an 8 core anyway. My FX8350 with an R9-290 GPU, would get something like 16FPS stock in the F-15 campaign. I could OC the hell out of it on water to the point my motherboard VRMs would get too hot and throttle anyway. I had a higher end board than what you're using.

It still seems logical and possible that the statement "DCS 2.0 won't be multi-core" is completely independent from the DX11 API itself actually running the multi-threaded rendering path, unless that is being explicitly excluded as well. This is somewhat of an important distinction, that still deserves some clarity 2 months later. It ultimately means threads would have an effect within the API limitation, and difference between someone thinking a dual core is enough where quad or higher would actually be meaningful.

I know. It just seems to me certain statements and expectations based on those statements are worthy of some sort of context. Its a been like a whole 2 months since I last brought it up, and its a long thread. :megalol:

That still irks me. Better performance from DX11 is primarily granted by the existence of 2 features, Multi-threaded rendering and Direct compute. If we assumed everything slow about DCS was the physics, then offloading that to the the GPU via Direct Compute will be great. It is however universally true there is a performance hit flying low level over cities, or pretty much just sitting on the ground compared to thousands of feet in the air away from cities. That performance hit is from rendering objects, which does relate to draw calls, which does gain from MT rendering. Without using MT rendering, DX11 is somewhat of a prettier DX9. The campaign missions also have performance issues unrelated to being close to cities. Excluding a bug, this could only be due to either the sheer number of AI calculations, the rendering of said AI, the physics of the AI, or the combination all in the single thread. I wouldn't be surprised if the physics is a major part that might contribute AI using the SFM. All this talk about "Using more GPU" sounds largely like they'll be using Direct Compute. Which, in practice, only offers post processing effects (pretty) and physics offloading. I swear the only chance there was no benefit in multicore, if they did indeed test it, is if the test machines were all high clocked Intels.

The likely lean toward blaming ED might come from DCS being the only title not working on AMD cards.

Not sure you've got that right. A scanned version of the 1975 -1 is really easy to come across. I don't think I saw anything about intentional single engine operation. As mentioned the issue is completely related to low temperature, which as a by product increases air density. It does make aerobraking somewhat impossible in winter missions, as the plane will never slow down to the point where the nose would fall on its own. I also have a (bad?) habit of starting my engines before running to the bathroom, which would lead to disaster in winter missions. Edit: Even the 1986 version is easy to find actually.

Really? I'm sure a recall a time where Aim9s would lock on to ground targets. Perhaps its been changed.

Its a football that can alter its glide path dynamically, but can only go as far as the arm can throw it. Supersonic jets are much bigger arms.

They work great in an ambush, though that's the best case scenario for all the Russian fighters, if they engage and miss they're pretty well screwed against everything else. Had one almost get me the other in my Eagle as I was more immediately concerned with some inbound flankers. He missed, my Aim-7 missed, dog fight ensued. My guns didn't miss.

This is often stated, and plausible but I find it insane with respect to fighter aircraft at least. If we think of the first DCS modules separate from FC, they were attack aircraft. With attack aircraft it doesn't largely matter (save for stupid AI wingmen) since the opposition is mainly AI ground objects (pre-CA). Even then at some point being offline you really just learn to game the game. There are a number of thing you wouldn't do or be expected to in an A-10C, and largely can never fail if you don't do anything willfully stupid. With AI fighters you have an opponent that could see everything, have an unrealistically low set of limitations, and still manage to be way too easy once you've figured out how to game the AI. How long have the SP campaigns been broken on some form or another. Anyone that is capable of MP and not using it in DCS is doing themselves a huge disservice.

Even though they don't do this, I think it goes without saying that every flyable fighter aircraft should have a contemporary human flyable aggressor. It is literally the only way to get the most out of the product. Given there is a large desire fly and employ these aircraft for what they were designed to do, doing that against only broken AI gets stale fast. The F4 is probably the most logical opponent to the Mig-21. Any single dissimilar same Gen fighter would do. There are obvious limitations to time frame such releases could occur, but I would personally not by a fighter who's contemporary wasn't planned for release by some Dev.

Yep, thats why I mentioned it.

It should also be noted that fuel flows from the externals only when the gear is up. The AAR hatch being open also will cause fuel to only feed from the main fuselage tank.

Knowing what we know about the AI, leaves me to wonder if anyone would even be able to enjoy the back seat without a human pilot. The AI has a seemingly perfect level of sensor manipulation, but there is probably some big reason AI aircraft use the SFM. The AI does a number of stupid things as pilot (particular in jet aircraft), I can't foresee wanting to be RIO with an AI driver. That in my mind slightly diminishes the value of the rear seat in the SIM, if you're probably going to always want/need a human pilot. An AI RIO though is likely to be so OP, it might be more reliable not having a person there.

Perhaps the better question is, how much do we have to pay for a good RIO?

A workaround I realized before just deciding to change the key assignments in TIR, is that double tapping and holding the key quickly will get it it registered in DCS.

Well DCS mission files are basically archives comprised of various files. So apparently all this time it was possible to insert and trigger an exe file, which is like a major security problem.