sc_neo

I don't care if it takes until late next year...just give'em time to do a proper job so we will have the most futureproof implementation they can come up with.

@Johnny_Rico True. Just rechecked prices and the 970 evo nvme 2tb is only 130-150€ more expensive than its SATA 3 counterpart. Worth a thought if one wants to invest in a 2tb drive right now. I just went with a 500gb since that one satisfies my needs right now and when prices come down even more...my ''loss'' won't be that big of a deal.

why did you not go with the 970 pro nvme 2gb version? Even more TBW...though i reckon its totally inconsequential for a gaming rig anyway...the total bytes written i mean. Still, i am not sure why you would buy multiple still comaratively expensive nvme drives if no here is able to produce any measurements that seem to justify this for gaming in general nor DCS in particular. Yeah, but when you building a system right now and if you're not strapped for chash and want to invest into something that will last you many years, maybe going its wise to go the nvme route even if they will probably drop in price quite a bit more over the next year.

Hrm, that SATA3 vs PCIe for a new SSD and gaming question is something i have not yet resolved properly for myself as well. Thing is, all benchmarks i have seen over the last 2-3 years make this pretty much negliable at the moment. I just got myself a 500gb Samsung 860 Evo which should be pretty much at the limit of the Sata3 specifications. And i have looked a bunch at the I/Os and read speeds while loading DCS and loading different levels. And i have never seen it jump beyond 400mbps. I have only looked at it with sysinternals processexplorer, there is probably better and more accurate ways to record i/os and sequential reads during loading or copying operations. But when i take this observation combined with all those SATA vs NVMe benchmarks together; then i conclude that right now and probably a couple more years down the road, for gaming at least, Sata3 is the way to go if you don't have the cash lying around in heaps. NVMes are still about 2x the price. Better invest that money into a better CPU, GPU or Hotas and stuff. I think the NVMe investment makes sense as soon as more game engines are designed in a multi-threaded way so that you actually see that game process make your reading operations jump beyond 500mbps. And with that same flawed/dirty quick look at the disk read operations during DCS missions, after the map has loaded, i see read operations in the low double digits. I mean somewhere between 2-20mbps, always interspersed with couple of seconds with no reading at all. All Sata 3 ssd should be able to handle this easily, and its more question how well ones specific setup of cpu, gpu, ram and drive are all running fast enough combined. Honestly, i believe its more important that game devs integrate proper loading/streaming technologies in there games so that the cpu has ample time to prep everything and load that stuff into memory before a stutter can even occur. But, my system is by no means beefy anymore, so if someone whos got an Ryzen 2700x or Intel 8700k overclocked with fast memory and a sata and nvme ssd could give this a go with identical DCS folders, i'd be sure interested to see whether this might actually make a difference afterall.

''So leaving an object behind outside your X radius, doesnt mean it gets throw own out of memory. Might be queued as the first to go, when memory is full?'' Probably, but i feel this is the standard way Windows memory management works. Keeping all already loaded stuff in memory even if not used anymore and only designating it to be overwritten if free memory space decreases. What i have seen yesterday was that DCS's memory working set was almost always double the size than DCS's memory private bytes. I wonder what DCS is loading into memory but not actively using. And that happend even when i have loaded one map only. And loading a mission on another map takes quite a bit longer than reloading the same map, what makes me think DCS does not keep all maps in mememory.

Are you sure this is how it works? What i mean is; the size of the preload radius (and i have no idea how this works in DCS) must not necessarily translate into less or more loading operations. It could equally well load the same tiny amount of 10mbps except that the outer edge is pushed out way further with max preload radius. Hrm, i just did a bad, quick and dirty test repeatedly loading some instant action missions with mix or max preload radius and i did not see any difference in loading speeds nor private bytes/working set DCS memory usage. I don't know, maybe this starts making a difference when actually flying a over terrain for a greater distance. I guess one would have to fly the same long mission twice with rebooting in between to really find out. The DCS World manual states on page 19; ''PRELOAD RADIUS. Radius of the preloaded terrain around user position in meters. A large setting can decrease hard disk swapping but it requires more system RAM. '' But, the min. value is 100...if it really is a metric value, 100m preload radius makes no real sense to me. Kinda don't understand why ED staff never took the time to make a detailed explanation about this setting so we woulnd not need to play the testing and guessing game.

looks promising, thank you! Time to get back into the Viggen :)

Cool, thx for keeping us in the loop. My printer is ready and waiting :)

So i have searched for all threads with preload radius in the title, its only about 8 or so. But its still not really well explained how the preload mechanic works and how your drive might influence it. But, we are in the P-51 forum section, not the most prominent place to discuss this.

btw: how would preload radius and SSD vs HDD relate? What i am getting at; if i have DCS running from a SSD instead of a 7200rpm HDD, would i not get away with a smaller preload radius since streaming from the SSD should be more rapid and hopefully fast enough to prevent DCS from stuttering? Less preload radius might in my layman understanding lead to less memory use and faster level/mission loading times?

Yep, but then i would hope that this dedicated server tech would be implemented for single player in the normal DCS World client so folks don't need to do that server workaround to achieve best performance. But i guess this would be the next step after the dedicated server has been released.

But.....isnt't DCS's main problem the sheer amount of power needed to do all the calculations for hundreds of units in the air, on the ground and water moving about, firing projectiles and missiles, shining light and hunking their horns? Vulkan can help to display this mass of objects at a great refreshrate. But the core-tech that makes all units, triggers and stuff spread over multiple threads but still remain in perfect sync is probably the big one that ED needs to tackle. Is there any info on what is being done about spreading the simulation part over multiple threads?

I'd rather wait and have the full campaign in one piece...mainly because; as things progress and are being worked on, often workflows develop and the way things are being done and implemented imrpove as well. But i am sure you two have given this a lot of thought and have made the best decision for yourself and the project. And yeah...once the first half is out it you can take a breather! Though, i am not sure whether people lay off after that cause once they have tasted from the cookie jar they sure as hell will be hungry for more. Maybe it will be kinda like the whining that often goes on if a module stays in early access for extended periods of time.

Any timeframe when the Viggen manual will be considered complete?

Maybe a good time for some to look into windows's shadow copy system or a third party alternative that gives you easy access to all your stuff on c after the fact. I use Paragon Backup and Recovery 16 (free edition) with a 3 day incremental backup plan. Rounds about to 80-100gb a month depending on how much is being changed on my c drive. Would be way smaller if had not installed a huge game on it as well. So the most i might loose is a little less than 3 days of work and changes. Simply mount the most recent or any of the increment containers (i keep about 4-5 months worth of backups on my data hdd) with a drive letter and i can copy back any file. Btw, depending on your settings in windows, the OS might even store previous versions of files on c, even if you have the shadow copy system disabled. Simply right click on any file, properties, and the right hand tab reads 'previous versions'.

hrm...thanks to both of you for sharing your experience. Damn, i won't be able to clear sufficient romm on my 120gb OS SSD to copy 2.5 OB over to see whether an SSD upgrade more or less soley for DCS is really needed or worht it at the moment. Anyway, more testimony is welcome...keep it coming :)

No offense guys, but I asked for first hand data, not for general advice. I can't test myself at the moment, i don't have access to my gamin rig. Anyway, if someone got a version both on a HDD and SSD, please share your observations.

hey guys, who got some first hand loading times comparison data for DCS between a standard i.e. 7200rpm HDD and a good cosumer grade SSD? Just would like to know how much of a difference it makes specifically for DCS and whether i should upgrade now, or wait until sometime next year when prices probably have dropped quite a bit more.

ja, sollte nicht negativ rüberkommen. Es ist halt hart...wenn man zig Wochen DCS nur über youtube und das Forum genießen kann :)

so ein Pech...war dieses WE nicht im Lande. Hätte zu gern die Persian Golf map getest, um zu sehen, ob sie vergleichbar perfromant auf meinem angestaubten System mit 7870 2gb vram läuft oder nicht. Mir wäre ein system wie bie ubisoft oder ausnahmsweise mal EA lieber, dass einen einen Titel für 10-15 Stunden uneingeschränkt probe spielen lässt. Ähnlicher Effekt, aber inklusiver. Es könnte ja zweigleisig gehen: jeder kann das freie Wochenende mitnehmen, oder das Paket 20h lang auch im nachhinein testen. Dann hätte es den werbewirksamen Eventcharakters eines free weekends, aber alle die verhindert sind oder was besseres zu tun haben als zu der Zeit zu Zocken, kommen trotzdem zum Zug. Naja, ein Andermal vielleicht.

So, after some patches and further development on the Persion Gulf map has happend since its initial release: how does the performance stack up against NTTR nowadays? I am asking since i have an AMD 7870 with 2gb of vram. NTTR runs fairly alright with very high settings when flying through open terrain, but the Vegas area is a bit taxing naturally. The ''new'' Caucasus on the other hand is pretty much unplayable when flying helicopters at lower altitudes and only great when in a jet at high altitudes, if i don't drastically lower my graphics settings in dcs, that is.

@borchi 2b and Damcopter I appreciate you guys getting together to look into things that might be improved or changed. I am sure any piece of software in general, or dcs module in particular, can always be tweaked and made better independently of the maturity of the product at the time. And the devs being open and welcoming to a new perspective that can shed a different light on their product is commendable. Keep it up!

Yep i have noticed this as well some time back and asked in a pertaining thread about it. I am not entirelysure whether any of the changes listed below have been added to the beta version already. I don't think so at least, i don't see any Viggen related updates in the DCS Open Beta changelog after the pertaining post by Cobra (https://forums.eagle.ru/showthread.php?t=218990). But those DCS changelogs are often incomplete, the BK-90 is in the game and was not mentioned in the changelogs also. "Other Changes The current, non-exhaustive (this will grow) list of changes due for that patch is as follows; Exterior mirrors now reflect the environment Fixed various clipping and shine-through issues in exterior (esp. Cockpit) Sculpted new normalmap for exhaust nozzle and area Sculpted new normalmap for exterior exhaust shell Added RoughMet textures for the exhaust shell Added turbine texture for better deep engine detail Metal elements on wings now have appropriate material definitions Animated engine exhaust fan New! Significantly improved visual fidelity of afterburner Afterburner now cast lights inside engine exhaust pipe Afterburner now cast lights on surrounding objects and terrain Corrected afterburner z-fighting Afterburner planes do not render at high viewing angles. Corrected afterburner self-illumination Fixed glass in LoD (distant) models turning opaque and bright. New! Added significantly better quality uncompressed normal maps for entire exterior Light domes are now appropriately shiny Fixed wheel/tire animation values so they spin at correct speeds." Maybe on of these changes fixes that sunshine issue.

Naja, hätte auch direkt von der Bundeswehr produziert worden sein. Keinerlei Reflektion oder kritisches Betrachten des weshalb und warum. Aber wenigstens nicht so reisserich wie manch US doku.

+1 Any non combat technical scenario would be very welcome, including firefighting or sea rescue operations. Imagine...bad weather, and you have to fly out to sea and hover over a sinking ship lifting out seamen or picking up emergency pods or such. And I believe those scenarios would not be that heavy on development resources either, but would add much to the existing mission profiles.