Preliminary 1.1.0.8 changelog

[Insanatrix]

I would have to say that their would certainly be a noticeable difference. Evident where if you monitor your CPU usage 1 core is at 100% where the rest are idle. That in itself would suggest its wanting more. Furthermore if you overclock your CPU the CPU remains at 100% while drastically improving your frame rate which indicates it is bottle necked at the CPU. In regards to resource management across the cores that is up to the devs to optimize. But I have no doubt that multi core support would improve performance.

 

Sure it's bottlenecked at the CPU. Basicly your cpu is queuing information to process in cache and processing it when it can which creates overhead which in turn creates a "slow down".

 

Split all the threads up now and stripe em across 4 cores. The threads still have to talk to each other. Physics need's info from input and graphics needs data from physics and sound needs data from input etc... What happens when the Physics thread calls for data from Input and Input isn't done calculating the data? or Graphics issue's a function call to the physics thread and it's working on calculating something else? Guess what, It still gets queued, you still have overhead.

 

Why do you think every game out there isn't optimized for multi-core's even though they've been around for what like 5-6 years? Even the one's that are truly don't even utilize all the core's effeciently. There's no easy way to get around it.

[Druid_]

Most of the games out there don't need to use multi-core, they are GPU ltd. Therefore even those games that are multi-core/multi-thread capable they don't need to use them efficiently. Parallel processing isn't something new, its been around a looong time and some software does a fantastic job of using multi-core processors (e.g. 3D studio), but only because they need to.

 

Think of it like this, the threads might need information off each other and even though one thread might stall briefly waiting for another threads instruction to complete, these threads in question are operating without the overheads that a single thread has, even with the stall it will complete the instructions quicker ---> improving performance.

 

Btw Graphics wont issue a call to physics, graphics gets its information when its ready. The main issue is how DCS is written, in C with lua calls and god knows what else. Having an assembler that can take it all and make it multi-thread is a huge task. Not an impossible one though.

Edited May 14, 2011 by Druid_

[OB1]

Sure it's bottlenecked at the CPU. Basicly your cpu is queuing information to process in cache and processing it when it can which creates overhead which in turn creates a "slow down".

 

Split all the threads up now and stripe em across 4 cores. The threads still have to talk to each other. Physics need's info from input and graphics needs data from physics and sound needs data from input etc... What happens when the Physics thread calls for data from Input and Input isn't done calculating the data? or Graphics issue's a function call to the physics thread and it's working on calculating something else? Guess what, It still gets queued, you still have overhead.

 

Why do you think every game out there isn't optimized for multi-core's even though they've been around for what like 5-6 years? Even the one's that are truly don't even utilize all the core's effeciently. There's no easy way to get around it.

 

Every game isn't optimized for multi core ? id have to disagree. Hi-end games such as Crysis, Arma2, Rise of flight use multi core perfectly. Rise of flight uses the latest intel 6 core CPU's evenly which has greatly improved its performance against dual, not so much quad with same clock speeds.

 

Why is it that DCS is the only game I have that uses more CPU single core percentage over GPU percentage. Games such as the ones I have listed above commonly use 95-100% GPU and 25-30% CPU across the cores where DCS thrashes the CPU with only 85%Single or 50-65%SLI GPU. (on my setup)

 

there are plenty of applications and games that use multi core CPU and use them well, why cant DCS be the same.. Even if 1 core is waiting for another, it is better than limiting the overall bandwidth of information.

 

Debates such as these can be endless. all I would like is information such as.. Are they looking into doing this or If not why is it not feasible. Is it too much workload or against the business model ? Honesty and an explanation would be great. But a vague explanation such as "We don't think it would make much of a difference" makes me feel like a mushroom.. :noexpression:

[jeffyd123]

If its so easy OB1, then why, O why havent you done it?

[OB1]

I understand it would be a major overhaul

At what point did I ever say it was easy ?

[Cibit]

I was hoping to keep that as a surprise :)

 

Nate

 

What else you keeping quiet about then Nate? Sobek is already being a tease

http://forums.eagle.ru/showpost.php?p=1197888&postcount=5 ;)

[kungfoo]

Also keep in mind that there's different types of "physics" in those different kinds of "games".

 

An action game (Crysis, Unreal Tournament, MW2) may actually get away with something that _looks_ like like believable physics.

 

In this sim however, there's the actual physics as it happens in the real world with a lot of stuff going on that is difficult to acurately model using numerical methods.

 

There's a reason, why there is one guy in the credits that did nothing else than "mathematical models of dynamics, ballistics ..."

 

Also as has already been pointed out, anytime, something depends on another thing which in turn depends on another thing, things tend to get very hard to pry from each other, so they can be processed in a truly concurrent way, which is what we're looking for, when you want true multicore processing.

 

Long story short: It is making a very complicated task even more complicated.

[nemises]

Every game isn't optimized for multi core ? id have to disagree. Hi-end games such as Crysis, Arma2, Rise of flight use multi core perfectly. Rise of flight uses the latest intel 6 core CPU's evenly which has greatly improved its performance against dual, not so much quad with same clock speeds.

 

Why is it that DCS is the only game I have that uses more CPU single core percentage over GPU percentage. Games such as the ones I have listed above commonly use 95-100% GPU and 25-30% CPU across the cores where DCS thrashes the CPU with only 85%Single or 50-65%SLI GPU. (on my setup)

 

there are plenty of applications and games that use multi core CPU and use them well, why cant DCS be the same.. Even if 1 core is waiting for another, it is better than limiting the overall bandwidth of information.

 

Debates such as these can be endless. all I would like is information such as.. Are they looking into doing this or If not why is it not feasible. Is it too much workload or against the business model ? Honesty and an explanation would be great. But a vague explanation such as "We don't think it would make much of a difference" makes me feel like a mushroom.. :noexpression:

 

I "think" the engine is beeing updated in a modular fashion.

I believe the audio system (for eg) . has been updated and is offloaded across multiple threads.

Lots of fancy shader improvements are used now, which offloads a lot to the vid card (which deals with threading natively)

But ultimately the core code in the engine was written waay before multi cores were common and optimising existing code is a lesser risk than re-writing from the ground up.

 

anyway, just my 2c , not trying to extend or prolong any mushroom growing :)

[Kreider]

Hi, I am new to DCS Flightsims Warthog and Blackshark. Wanted to thank ED for the game support and the feedback. Having read back in the forums I am amazed at the sheer level of support and interaction from ED members. Thank you. ED, bring out another title and you have my 50-60 bucks for sure.

 

At the same time, I am very happy that these forums are full of people who genuinly offer help and advice without bashing or ridiculing newcomers like myself. Refreshing. Thank you.

[OB1]

anyway, just my 2c , not trying to extend or prolong any mushroom growing :)

 

LOL!

[Prophet]

Sure is an impressive list!

[digitaljjd]

At what point did I ever say it was easy ?

 

I'd have to agree with that. I'm just Joe Blow consumer, but I'd say ED is reaching the limits of it's current entertainment side of The Fighter Collection Simulation Engine and is going to have to go through growing pains with future modules, but if I'm not mistaken, I think that was their plan? I'm sure they will handle it in the end. "In the simulation world, the only letters you need to get eveything from A to Z, is TFC and ED." How's that for a future slogan? :D -Can I get a free copy of the next module for that?

Edited May 14, 2011 by digitaljjd

[andysim]

is the tanker fixed in 0.8? ive asked in the support forum and for over a month there has been no word and its not in the changelog.

 

The tanker gets stuck and drifts off the map when an AI aircraft finnishes refueling.

[Olgerd]

is the tanker fixed in 0.8? ive asked in the support forum and for over a month there has been no word and its not in the changelog.

 

The tanker gets stuck and drifts off the map when an AI aircraft finnishes refueling.

 

Yes it is fixed.

[andysim]

whohoo!

 

/does happy dance :pilotfly::joystick:

[Druid_]

More good news on the tanker!

 

What else is not in the list? Maybe Wags can patch the list or publish the patch to the patch list?

[Nate--IRL--]

More good news on the tanker!

 

What else is not in the list? Maybe Wags can patch the list or publish the patch to the patch list?

 

It was the same for the last two patches too, there are simply too many things to list.

 

Nate

[WH_Boomer]

It was the same for the last two patches too, there are simply too many things to list.

 

Nate

 

So does this mean there is the possibility that there is more fixes for multiplayer than:

Multiplayer

• Multiplayer performance degradation due to an anomaly large SADL message traffic was fixed.

• Muliplayer users spawning at same location fixed.

• Multiplayer timeout / server timeout fixed.

 

 

Primarily all of the server side only functions, such as JTAC communications, A-A TACAN, etc. I know the demographics you guys operate on shows the multiplayer is a small part of the community, but certainly not one than can be ignored for an extended period. Stuff that worked in beta but doesn't now should be a priority I would think.

 

Just my humble opinion, thanks for all the work you guys do get done.

[159th_Viper]

So does this mean there is the possibility that there is more fixes for multiplayer than......

 

It's being worked on as I type.

[159th_Falcon]

Small Hijack

 

compatibility patch between black shark and whartogII is now confirmed by Wags i see.

 

Any definitive decision on FC2 yet?

 

//runs away

[miguelaco]

Has HDR effect been tweaked so we, ATI 69xx users can expect white boxes and flickering glitches have been resolved for .8 patch?

 

More info here: http://forums.eagle.ru/showthread.php?t=71368

[Mathiew]

I was expecting more focus on multiplayer improvements for this patch... :(

[goldfinger35]

This are my favorites:

-Wingmen Engagement logic improved when threats are in target area. Better self-preservation.

-Wingmen now better engage targets around player SPI when give "Engage With..." command.

-Tor, 2S6, and Osa search radar RWR indications fixed.

-Reported errors in campaigns have been fixed.

 

And of course JTAC and multiplayer fixes.

[Druid_]

The one bug I'd like sorted is the lag you get when you ACTIVATE units. I know it was mentioned somewhere that ED were working on it. Hope it comes sooner rather than later!

[Baggy MacDouche]

Has HDR effect been tweaked so we, ATI 69xx users can expect white boxes and flickering glitches have been resolved for .8 patch?

 

More info here: http://forums.eagle.ru/showthread.php?t=71368

+1