Lock granularity, bubbles, shaders, texture filtering

[sthalik]

It’s no secret BMS is inefficient.

1. I’ve seen a team member say that locks have bad granularity, threads are spawned too often.
2. Bubble’s atrocious, totally breaks immersion. Changing radar range suddenly makes stuff exist.
3. Shaders could use some work. Not just HDR, even day ground texture got something too inefficient about it…
4. Anisotropy eats more GPU than it should.

So why don’t WE do something about it?

BMS has done an amazing job WRT stability and FM but there are many aspects where it’s lacking.

[Guest]

I’d love to have the bubble issue fixed. Not being able to see anything past 20-30nm is annoying, especially coming from DCS A10C. The F-16s TGP just seems like a big step backward.

However, the question is HOW can we do anything about it? I would guess that if this were a simple .cfg line tweak someone would’ve done it by now.

[sthalik]

BMS could release parts of code needing refactor to those who are able. Talking about lack of manpower is kinda’ funny. Last time I spoke with the shader wiz dude, he was willing to help. Heck, I’d even do that myself, over a larger timeframe.

There’s so much BS about how everything is ‘hard’ or ‘impossible’. It took 10 minutes for code changes needed for headless BMS. And people have been wanting that for years?!

[jinro88]

As long as we don’t extend the bubble too far. DCS’s performance in the campaign is abysmal.

[sthalik]

Profiling can tell how much bubble can be extended.

[Guest]

What do you mean “headless BMS”??

I do agree though. If they release bits that others can work on, those with the skill can hack at it, report their findings to the BMS team, and then they can decide how exactly to implement the new discoveries. However, won’t this mean that there will be “different” versions of BMS later on? ie, if they released the code to change the bubble, and we found out we can change it to 30, 40, or 50nm, but BMS decides to just extend it to 30nm, what’s to stop others from hacking it again to extend it to 50nm in their version of the sim? And won’t we have problems later on if some of these end up to be MP-critical?

In other words, are we going to be doing more harm than good by going down this path?

[Dee-Jay]

The reason why we want to remain a “small” team, working with guys that we trust and that we know is in Falcon4 history.

Yes we could have more talents in the DEV team … but above all … we want them to be “friends” first and be sure that they 100% understand BMS objectives and philosophy.

… It will take more time … yes … thats is not a problem … for BMS team, time is not money.

Team’s friendship is paramount… but everything is possible. Team member list is not “locked”.

[sthalik]

What do you mean “headless BMS”??

BMS server without a graphics card, running at 100+ FPS on ramp.

However, won’t this mean that there will be “different” versions of BMS later on?

The issue here is also the quality of Microprose code. It’s known for action-at-distance, abhorred by maintainers… Stuff can’t be separated into distinct subsystems easily, it’s full of global variables, cross-references and other trash, a mixture of ancient kindergarten C++ and old-style non-standard C that makes Fortran programmers blush.

if they released the code to change the bubble, and we found out we can change it to 30, 40, or 50nm, but BMS decides to just extend it to 30nm, what’s to stop others from hacking it again to extend it to 50nm in their version of the sim?

If someone profiled the code, found the bottleneck, then it’d get extended. BMS damn sure knows how to extend the glorified bubble.

@__DeeJay__, while a reply is appreciated, interested in getting something done. @__Dunc__, what about those shaders? You want them fixed?

[Guest]

@Dee-Jay:

The reason why we want to remain a “small” team, working with guys that we trust and that we know is in Falcon4 history.

Yes we could have more talents in the DEV team … but above all … we want them to be “friends” first and be sure that they 100% understand BMS objectives and philosophy.

… It will take more time … yes … thats is not a problem … for BMS team, time is not money.

Team’s friendship is paramount… but everything is possible. Team member list is not “locked”.

I’m sorry but the way I understand that statement is that you want to keep the team small because you only want “friends” working in the team. It sounds like elitism and just patting each other on the back….

I agree with you that time is not money, we are playing with a sim that is more than a decade old, but if allowing the community to help out means that features get added sooner rather than later, isn’t that a good thing? As for “BMS objectives and philosophy,” I see where you’re coming from that is why I mentioned concerns about “different versions” of BMS, but your post does not help at all. If someone from BMS would come up with rules that “3rd-party modders” would have to respect and follow, I’m sure those with the skill would be happy to oblige; if not, then they simply do not work on BMS.

Just like with the FLARE project, if the original modder had not opened his work to others, who knows when FLARE would be released? On the other hand, if he had opened his work earlier, maybe we’d have FLARE by now? I’m pretty sure that the guys working on FLARE now would’ve still respected the original modder’s wishes even if he opened his work ages ago. So why not do the same with some parts of BMS?

Just my 2cents.

[Guest]

@sthalik:

BMS server without a graphics card, running at 100+ FPS on ramp.

I see, thanks for explaining.

[Red Dog]

Well, if I were a deciding factor, the load of crap I read on this very forum certainly wouldn’t convince me to give access to more people in the way you describe
Doing that, it would go to hell in about the same time lapse this forum went to hell when it was opened to the public.

that’s my opinion only and does not engage the views of the BMS dev

[sthalik]

Ice:

I’m more concerned about limited human lifespan. I’d rather not get things done in 3-4 generations…

Atari (which dies in gote with 9-stones handicap and negative komi) bogeyman is a tired old song. Better to contribute back (yes, “back”, as BMS didn’t create the code in a vacuum) the flight model, stability changes et al. than use excuses.

Fragmentation? Better than monoculture…

But wouldn’t use such hostile tone as Ice did…

@__Red Dog__:

Doing that, it would go to hell in about the same time lapse this forum went to hell when it was opened to the public.

Please define “go to hell”.

No one says anything about read-write access for arbitrary people. Just review of sane incoming patches…

[Dee-Jay]

@-Ice:

I’m sorry but the way I understand that statement is that you want to keep the team small because you only want “friends” working in the team. It sounds like elitism and just patting each other on the back….

We just want to avoid code leak and troublemaker.

[Dee-Jay]

I agree with you that time is not money, we are playing with a sim that is more than a decade old, but if allowing the community to help out means that features get added sooner rather than later, isn’t that a good thing?

Adding features is a VERY easy job.

Adding features without introducing bugs in a totaly diferent part of the code and be able to coorditate all those new features to make everything working fine with all other DEV updates is a totaly diferent challenge.

Do not think that we do not like the idea of “improoving BMS” good and fast. but it is not possible … just take a look to the critisms from the community about the UP4 which has been released a bit too early …

[molnibalage]

I’d love to have the bubble issue fixed. Not being able to see anything past 20-30nm is annoying, especially coming from DCS A10C. The F-16s TGP just seems like a big step backward.

Nothing have to be fixed. This is a feature, this is the price of dynamic campaign. Buy a supercomptuer and you can forget the bubble. Most of players do not have supercompter. If annoys you BMS4 fly DCS and do not complain…

[Dee-Jay]

@sthalik:

BMS could release parts of code needing refactor to those who are able.

Forget about that idea.

If you have a part of the code, it is because you are team member.

[Guest]

Red Dog, I agree with you on that and that is one of the reasons I am thinking that there needs to be a set of rules to abide to. Plus I would guess that the BMS team would have a good idea of who can do what. I’m not saying open EVERYTHING to EVERYONE, but I’m just thinking a small addition of skilled and dedicated manpower will not be a detriment to BMS.

sthalik, like you, I am very grateful for BMS. The team has definitely given life to this old sim, so much so that I cannot return to DCS without feeling like stepping backwards a few steps. “Paying it forward” definitely is the way to go.

Dee-Jay, please note that my previous post was not meant to contain any hostility, and I am happy to see that you have not viewed it as such. I understand your points, and would gladly volunteer to be a tester should a position open itself. Regarding code leaks and trouble, again, I’m not saying release the code to everyone, but rather tapping the correct individuals on the shoulder and asking them to join the fold.

Surely not everyone in the BMS team knew each other personally when you guys started? Why not let some new and willing individual step forward and prove himself, rather than just closing the doors all the time?

I know that it’s easier to “stay as you are” and keep within the comfort zone and opening up, even a little, sounds very scary and risky. All I’m saying is that the end benefit to BMS just might be worth it, if someone is willing to stick their neck out a bit.

Again, my useless 2cents.

[Guest]

@molnibalage:

Nothing have to be fixed. This is a feature, this is the price of dynamic campaign. Buy a supercomptuer and you can forget the bubble. Most of players do not have supercompter. In annoys you BMS4 fly DCS and do not complain…

Wow, that was mature.

Why not just get rid of all the eye candy, some people play on laptops after all. Are we going by the majority or by least-common-denominator?

[Blu3wolf]

@molnibalage:

Nothing have to be fixed. This is a feature, this is the price of dynamic campaign. Buy a supercomptuer and you can forget the bubble. Most of players do not have supercompter. In annoys you BMS4 fly DCS and do not complain…

because DCS is not going to elicit complaints?

what a ridiculous statement….

with my computer, I feel limited more by the code than by performance. Id prefer to have it an option that can be fixed.

[Amraam]

I think everybody misunderstood Molni comment.
He’s not wrong about bubble. Bubble system is not so bad and have no direct issue with the problem. the problem is the capacity of sensor to deaggregated unit outside the bubble when this particular sensor have the ability to see outside this range (TGP, FCR, …). Increase the bubble will be slideshow in campaign, even probably in TE. That’s not the solution. The solution is to give the capacity to the sensor to force the unit to be deaggregated trough the FOV of the sensor, and keep it deaggregated during the TOF of standoff weapon. Then, aggregated again when sensor FOV is outside the unit, and no weapon in flight.
We have not wait public comment to work on it ;).

I love people comment like “yeaaahhh I know, it’s easy, just do that”. That’s why is better to keep the code for people have experience with Falcon code, sometimes the solution is not the more easy to have.