92 replies to this topic

[Imperius]

MauttyKoray, on 28 July 2015 - 09:22 PM, said:

Cause unfortunately we have less enthusiasts interested in the simulation style of the game, and instead more try hards who wants their 'skill' to count (cause pinpoint alpha-lol-strike is so much skill...).

But yeah, DX12 would just be far too much work for the benefit and take time out of other, more important projects to update the game.

Again you say more pressing issues and don't list what those are.

[Imperius]

Mister Blastman, on 28 July 2015 - 09:32 PM, said:

Umm. No. People hate CW because it is simply DOTA Mecha. It is pretty ********. People didn't ask for DOTA Mecha.

Look, you list that DX 12 will get this, this and that--but guess what, it is ALL speculation!

What is NOT speculation is how little of an improvement DX 11 was over 9. In fact... 9 was WORSE than 11 in many cases. But I guess you forgot about that part?

Issues to fix:

Balance (through convergence)
- guess what, .xml fixes, band-****, NOTHING has fixed the balance issues. The core problem is convergence.

Maps (we lack them)

Game modes (they're stale--ticket based respawn gameplay like MW:LL or Battlefield please)

Hit Reg (it still sucks!)

Collisions and knockdowns (ya know, we were promised these what, two, three years ago?)

You know what, I'm not going to bother. Why? Because nothing ever changes. Nothing gets fixed. It's just more of the same. And by shoveling an API upgrade onto their plate, you're helping nothing get done even more!

You only listed two issues that require the use of the engineer. Collisions got added does need some work, seems like los cryengine code issue, knockdown are probably never being added and Hit Reg is being improved and is almost done, the rest of the issues is art team and other staff. Adding a game mode like battlefield won't happen. CW is the closest you'll get to that.

DX 12 performance gains is not speculation... Please watch the videos or read the articals at least.

Edited by Imperius, 28 July 2015 - 09:46 PM.

[sycocys]

Imperius, on 28 July 2015 - 09:13 PM, said:

Ok, what should they be working on? By the way adding DX12 API isn't a whole team task. Two people did it for a DX12 featured game and it only took 6 weeks.

Side note: the way MWO uses CPU and draw calls, it would be a significant gain I've provided all the facts to back up that claim, though videos and articals.

I'm not even talking about improving visuals although I'd love that I have two GTX980's in SLi

Unfortunately PGI doesn't have programmers that would take less than six months to a year to make the change if they dedicated 100% of their time to it. The people they have currently do a decent job, but they do it very slowly.

What these guys should be working on is developing the core features to make CW an actually interesting game mode rather than standard drops on dota style maps and starting to lay the groundwork for Solaris. MWo is about to open up to a much broader audience with the inevitable steam release, and this game is going to have a tough enough go at it without diverting the game progress to mild graphical improvements.

[Mister Blastman]

Imperius, on 28 July 2015 - 09:41 PM, said:

You only listed two issues that require the use of the engineer. Collisions got added does need some work, seems like los cryengine code issue, knockdown are probably never being added and Hit Reg is being improved and is almost done, the rest of the issues is art team and other staff. Adding a game mode like battlefield won't happen. CW is the closest you'll get to that.

DX 12 performance gains is not speculation... Please watch the videos or read the articals at least.

Once again, speculation. You don't work for PGI. You don't work on this game. The only thing that isn't speculation is past results.

[Deathlike]

Kiiyor, on 28 July 2015 - 06:15 PM, said:

I remember the post. It's why inverse kinematics was nuked.

Pretty much, anything that involved more CPU was nuked.

You'd think PGI would provide a toggle for these things, instead of globally disabling everything that they can't figure out on their own.

[Kiiyor]

Deathlike, on 28 July 2015 - 10:06 PM, said:

Pretty much, anything that involved more CPU was nuked.

You'd think PGI would provide a toggle for these things, instead of globally disabling everything that they can't figure out on their own.

The trouble with toggling IK though, is that legs might be in different positions on different clients, so that probably wouldn't work. The damage textures however, should absolutely be on a toggle - though I suspect the new damage texture system is easier to apply to new mechs than creating a whole separate texture damage channel the way the old system did. I miss it though. Seeing actuators and myomer exposed was far more immersive than the 'rusted car in a redneck's backyard' damage system we have now.

I do wish there would be an optimization pass though - and I have high hopes that the newer maps will provide this somewhat.

The new River City is far more intense (visually) than older maps, yet runs better on my potato than just about any other map.

Maybe, when our older maps have all had the crap knocked out of them, performance will have increased enough globally that IK and better damage textures can make a return.

DX12 though... man, some of that stuff is awesome. Things like shadows and particle effects are apparently an order of magnitude more efficient than they are in DX9&11. It apparently makes huge strides in freeing up CPU cycles in CPU intensive games - some of the test systems have reported significant lowering in operating temperatures and CPU load in games running on DX12, compared to the same games in DX11. I've a friend in game development who has seen some of the tech and specs, who is ludicrously excited, and regularly nerd talks the crap out of me about it.

Here's a shot of the CPU reduction in one of their demos before and after switching from DX11 to 12:

And one of the potential CPU to GPU offloading available:

You know what else it brings? Integration between dedicated and integrated graphics (or just mixed GPU's in general). Basically, every freaking laptop owner on the planet should be salivating at the thought of that. Ripped from one of the tech blogs:

man, it sounds like i'm trying to sell it. At the end of the day, we can't tell how much of this was engineered specifically to do well in tests, and we won't have any idea of real world performance gains until WIN10 is more common, but it's still interesting.

Here, some reading:

http://blogs.msdn.co...g-progress.aspx

[Imperius]

Nice post

[Imperius]

Installing Windows 10 and DX 12 I let you know how much my PC is supposed to crap out. I bet I have no issues here we go.

[MrPlow254]

I'd be able to use an Nvidia card with my AMD card at the same time if we get dx 12. I'd rather have AMD as my daily driver since well AMD drivers are way better. My Nvidia cards i have had suck because the drivers suck.

[Pinscher]

There is no sense in wasting resources to migrate to dx12. Win10 requires you to give up your personal information. They even go so far to brass ball right in the EULA to collect credit card information. Directx has for a long time been the only thing keeping the gamer market on windows platform. Im one of those gamers, that when I am forced out of my win7 pro( that i paid $200 at a retail store for) I will be done with windows. To be honest I am considering switching sooner. I will lose MWO but i will retain civ5, since it has a linux release.

I am not alone with this near future choice. What is going to worth PGI time and effort is going to be Vulkan support. More specifically linux and opengl support. Vulkan alone could be a factor for people leave Windows products.But when coupled with the absolute dastardly way Micrsoft has been tricking, or forcing win 10 upgrades, people are going to make the shift. And any game that refuses will be left behind. i realize there are options to get windows programs to run on linux. But I am doing my work and effort to get away from Spydows. Effort needs to be done on the games end as well.

Technically since dx9 is PGI's API of choice, they only need opengl support and not Vulkan. Open gl is similar to dx9 limitations, and vulkan is the next step much like dx 12.

Windows, seriously.. wtf? Most Graphical linux distros come is around 512 megs of ram usage and about 10 gigs at the absolute most for drive space. My win 7 use a bit less than 2 gigs of ram. and is bloated out to well over 20 gigs. Linux is a better platform for gamers anyways. It uses way less resources, and if it is still too much you can remake it in your own image, should you choose to do so.

Now to answer the mystery of why the one gentleman COULD NOT upgrade to windows 10. There is a ton of software out there in the industrial and commercial world. They pay thousands to millions per year for systems hardware and software. Most of the devices and software are locked to a specific OS. I personally have seen alot of machinery that the software runs win NT in the background. We are talking like large transfer machines and CnC equipment. There are more things to consider than just home pc market. If ms ever comes out and says upgrade now to win nt..... industry will most likely shut down.

[LordNothing]

id rather they just port everything to vulkan. no way in hell im using windows 10.

[LordNothing]

Imperius, on 28 July 2015 - 02:39 PM, said:

You do realize that windows 10 is free for all users of windows 7 and 8 correct?

it only costs you one soul. yours. its in the eula.

[SpiralFace]

Deathlike, on 28 July 2015 - 10:06 PM, said:

Pretty much, anything that involved more CPU was nuked.

You'd think PGI would provide a toggle for these things, instead of globally disabling everything that they can't figure out on their own.

Its not that simple.

IK animations means that leg bone positioning would have to be server side authoritative to ensure that the hit boxes synced up with the visuals 1 for 1 on the screen to ensure that whatever you shot was really "what was there" on the server.

You make it toggleable on the client side, you are going to see a LOT of shots that look like they should hit the leg sail through the leg because you flipped on IK animations on your side, but the server / potato user's clients don't have it.

IK animations are dependant on two things:

- Min-spec machines can handle the additional load to make it a universal thing. (almost 9 year old hardware guaranteed through min-specs at this point.)
- Making the hitboxes Server Authoritative WITH the IK chains being calculated both on the client and the server to allow for hitboxes to connect.

If you can't get those two things, your going to see breaks in the hitboxes for everything. (As IK animation will also affect mech torso positioning to see the legs connect on the hills.) And so when it comes to graphics vs gameplay, you can pretty much bet that gameplay will always win out. Especially if they are still advertising support for nearly 9 year old rigs.