Jump to content

Is It The End For Mechwarrior: Online, Or Finally A Much Needed Fresh Start?


543 replies to this topic

#381 Bishop Steiner

    ForumWarrior

  • PipPipPipPipPipPipPipPipPipPipPipPipPipPipPip
  • The Hammer
  • The Hammer
  • 47,187 posts
  • Locationclimbing Mt Tryhard, one smoldering Meta-Mech corpse at a time

Posted 11 September 2014 - 06:10 AM

View PostTombstoner, on 11 September 2014 - 05:56 AM, said:


But when using 155mm artilery shells.... essentaly bombs.... a 5.5m "Miss" can still be effective due to the size of the blast radius.

The skud missle used by Iraq on isrial during the first gulf war had a COF of about 2+ kilometers. for a conventioal warhead thats next to usless butt he skud was designed to carry nukes.... thus a nuke hitting 2-3 kilometers away and calling it a miss is just semantics.

and all of which really have nothing to do with direct fire HEAP shells from autocannons and their accuracy being fired from moving mechs, at moving mechs.

#382 Prezimonto

    Member

  • PipPipPipPipPipPipPipPipPip
  • Survivor
  • Survivor
  • 2,017 posts
  • LocationKufstein FRR

Posted 11 September 2014 - 06:13 AM

I'm mostly with Bishop. I don't want to troll. I don't criticize the devs without a specific gripe, and when I do I try to remember to point out how the situation could be improved.

I've got such mixed reactions to the spate of recent annoucements. Happy(bought out IPG), sad(new clan pack announced and I won't be purchasing), annoyed(the idea role warfare has been reasonably implemented enough to consider it possibly finished), and bewildered(a new IP that directly competes with SC and ED followed with not using Cry Engine which they have significant experience using?) have all crossed my path in the last week.

None of that changes the fact that previous decisions and pace of development here have left me finding significantly more enjoyment following the forums than actually playing the game. I'm just toughing out the wait for CW, if it's good, I'll attempt to get back into MWO, if it's not I'm probably done with this game and PGI.

Edited by Prezimonto, 11 September 2014 - 06:13 AM.


#383 Chrithu

    Member

  • PipPipPipPipPipPipPipPip
  • Bad Company
  • 1,601 posts
  • LocationGermany

Posted 11 September 2014 - 06:49 AM

View PostJackpoint, on 11 September 2014 - 04:20 AM, said:

Cryengine is inferior to every other engine out there, cpu bound over gpu and chosen for cost only.

http://www.gamespot....n/1100-6418427/

Mwo could have been so much more without the cryEngine .

Remember this?.



Hmm ok so that link shows me CryEngine is cheaper than UE4. But I'd love to see benchmakrs or whatever that prove that CryEngine per se is CPU bound over GPU? Because my experience is that again MWO is the only CryEngine game I have played for which this is truely perceivable and thus an exception to the rule.

View PostBishop Steiner, on 11 September 2014 - 05:31 AM, said:

well, the biggest issue Cryengine has is high speed. Hence the issues with MASC, or getting mech sover 150 kph without mass rubberbanding, and attendant hit reg issues. SC threw a lot of time and money with actual crytechs to get their issues reworked and were still put considerably behind schedule. But because Roberts is better at lying and kissing babies than Russ, its OK for them to miss deadlines.


Well as far as I am aware at the scale MWO is running and considering it is using 32bit I guess the speed problem it has is something you would run into with any other 32 bit engine as well taking the same approach at scaling things.

As far as SC goes their problem was that they simply decided NOT to modify CryEngines 64 player networking code to fit their testing purposes but instead finish the work on their own MMO-ready networking backend (which was planned to be made all along) before starting the multiplayer alpha.

And as far as forgiveness about missed "deadlines" in the SC community goes: That is simply earned. For starters they were always upfront about the fact that they absolutely WILL delay stuff if they need more time to do things right. Then they usually keep people in the loop about the reasons for delays. Heck in the case of the delay of the Dog Fighting Module they even asked the community whether they are fine with the delay to get their networking backend done or if they should double spend to get CryEngine's networking ready in time just to replace it later on anyways. And the voters by a large margin chose the delay over double spending. And then there juts is a slight but important difference between saying "We plan and hope to be done in two weeks." and saying "It will be done in two weeks."

Sure SC is behind their original schedule, but firstly it was always announced to "just" be their internal roadmap and goal, secondly I repeat they also said they will delay things to do it right and not rush to hit their roadmap and thirdly: no one following the project and capable of understanding just how ambitious just the original planned feature set was and how much more ambitious the project grew to be with ever more money flowing into it would expect that they are done with everything in 2015 (their orignally planned time of release).


View PostBishop Steiner, on 11 September 2014 - 05:31 AM, said:

Probably UE4 would have been a better choice, for the engine, if just performance was the goal. But (and I again, could be wrong) am pretty sure at the time of development, Cryengine was far more affordable and accessible than UE4.

By the time they were done (and really, they still aren't) reworking the code from the ground up, IDK if it was still actually more cost effective than UE, though.


I was wondering about that, and wondered it it was through scaling that SC finally overcame those same issues? (space maps having less boulders to worry about)


And dont forget the "have to shut yourself down".


As far as I understand it SC's approach to solve the scaling and speed issues is to migrate everything to 64 bit. And I also still think that PGI would have run into just the same issues with UE 4 as they have with CryEngine.

#384 IraqiWalker

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 9,682 posts
  • LocationCalifornia

Posted 11 September 2014 - 06:56 AM

View PostJason Parker, on 11 September 2014 - 06:49 AM, said:

Hmm ok so that link shows me CryEngine is cheaper than UE4. But I'd love to see benchmakrs or whatever that prove that CryEngine per se is CPU bound over GPU? Because my experience is that again MWO is the only CryEngine game I have played for which this is truely perceivable and thus an exception to the rule.


Probably because MW:O is the only game that forces rendering over such massive distances with objects this huge. Too many moving parts + an engine that needs a LOT of optimization = more apparent draw on resources.

#385 RetroActive

    Member

  • PipPipPipPipPipPip
  • 405 posts
  • LocationFL, USA

Posted 11 September 2014 - 07:03 AM

View PostJackpoint, on 11 September 2014 - 04:20 AM, said:

Cryengine is inferior to every other engine out there, cpu bound over gpu and chosen for cost only.

http://www.gamespot....n/1100-6418427/

Mwo could have been so much more without the cryEngine .

Remember this?.


I do remember that and the mechs look horrible. I don't want to drive that cartoon Atlas.

#386 Chrithu

    Member

  • PipPipPipPipPipPipPipPip
  • Bad Company
  • 1,601 posts
  • LocationGermany

Posted 11 September 2014 - 07:08 AM

View PostIraqiWalker, on 11 September 2014 - 06:56 AM, said:


Probably because MW:O is the only game that forces rendering over such massive distances with objects this huge. Too many moving parts + an engine that needs a LOT of optimization = more apparent draw on resources.


Sure. I do not argue that. But what is to say that other engines would be able to handle this better? That really is my only question here.

We acknowledge that MWO the way it is done puts a pretty unique kind of stress on the capabilities of the engine and then we draw the conclusion that the engine was a bad choice (as in other engines would have been better) without the ability to have useful comparisons to support that argument (like a mech game done on another engine) simply based on the fact that the game is capping on the CPU and leaves the GPU's unstressed?

Am I the only one that sees that this logic might be slightly flawed??

I never played Living Legends but since it is a Mech game on CryEngine as well I wonder if it saw similar problems?

#387 Bishop Steiner

    ForumWarrior

  • PipPipPipPipPipPipPipPipPipPipPipPipPipPipPip
  • The Hammer
  • The Hammer
  • 47,187 posts
  • Locationclimbing Mt Tryhard, one smoldering Meta-Mech corpse at a time

Posted 11 September 2014 - 08:46 AM

View PostJason Parker, on 11 September 2014 - 06:49 AM, said:



Hmm ok so that link shows me CryEngine is cheaper than UE4. But I'd love to see benchmakrs or whatever that prove that CryEngine per se is CPU bound over GPU? Because my experience is that again MWO is the only CryEngine game I have played for which this is truely perceivable and thus an exception to the rule.



Well as far as I am aware at the scale MWO is running and considering it is using 32bit I guess the speed problem it has is something you would run into with any other 32 bit engine as well taking the same approach at scaling things.

As far as SC goes their problem was that they simply decided NOT to modify CryEngines 64 player networking code to fit their testing purposes but instead finish the work on their own MMO-ready networking backend (which was planned to be made all along) before starting the multiplayer alpha.

And as far as forgiveness about missed "deadlines" in the SC community goes: That is simply earned. For starters they were always upfront about the fact that they absolutely WILL delay stuff if they need more time to do things right. Then they usually keep people in the loop about the reasons for delays. Heck in the case of the delay of the Dog Fighting Module they even asked the community whether they are fine with the delay to get their networking backend done or if they should double spend to get CryEngine's networking ready in time just to replace it later on anyways. And the voters by a large margin chose the delay over double spending. And then there juts is a slight but important difference between saying "We plan and hope to be done in two weeks." and saying "It will be done in two weeks."

Sure SC is behind their original schedule, but firstly it was always announced to "just" be their internal roadmap and goal, secondly I repeat they also said they will delay things to do it right and not rush to hit their roadmap and thirdly: no one following the project and capable of understanding just how ambitious just the original planned feature set was and how much more ambitious the project grew to be with ever more money flowing into it would expect that they are done with everything in 2015 (their orignally planned time of release).




As far as I understand it SC's approach to solve the scaling and speed issues is to migrate everything to 64 bit. And I also still think that PGI would have run into just the same issues with UE 4 as they have with CryEngine.

only thing off the top of my head, and it's just one game being described, will look for more later
http://www.dsogaming...mance-analysis/

Mind you though, if MWO is having to do all these calculations, I can understand why hit registry especially at speed, would suck so hard.

View PostRetroActive, on 11 September 2014 - 07:03 AM, said:


I do remember that and the mechs look horrible. I don't want to drive that cartoon Atlas.

Cartoon atlas reminds me of the horrible "Pain and Gain" renderings of the Transformers in "Age of Extinction". (well technically vice versa since T:AoE is newer).

Either way, the mech designs were definitely horrible.

#388 Bishop Steiner

    ForumWarrior

  • PipPipPipPipPipPipPipPipPipPipPipPipPipPipPip
  • The Hammer
  • The Hammer
  • 47,187 posts
  • Locationclimbing Mt Tryhard, one smoldering Meta-Mech corpse at a time

Posted 11 September 2014 - 08:59 AM

View PostJason Parker, on 11 September 2014 - 07:08 AM, said:


Sure. I do not argue that. But what is to say that other engines would be able to handle this better? That really is my only question here.

We acknowledge that MWO the way it is done puts a pretty unique kind of stress on the capabilities of the engine and then we draw the conclusion that the engine was a bad choice (as in other engines would have been better) without the ability to have useful comparisons to support that argument (like a mech game done on another engine) simply based on the fact that the game is capping on the CPU and leaves the GPU's unstressed?

Am I the only one that sees that this logic might be slightly flawed??

I never played Living Legends but since it is a Mech game on CryEngine as well I wonder if it saw similar problems?

Probably because previous Unreal Engines handled it very well?

#389 Mad Strike

    Member

  • PipPipPipPipPipPipPipPip
  • The Privateer
  • The Privateer
  • 1,298 posts
  • LocationLima , Peru

Posted 11 September 2014 - 09:40 AM

View PostBishop Steiner, on 08 September 2014 - 08:03 AM, said:

There is much discussion, with the announcement of the New Game, Niko's Community Post, the recent spate of communication by Russ, etc. Some are angry. Some are happy, a lot are justifiably confused, and or, concerned. I'm probably a little bit of all the above.

While reading some comments and replies by other long term community forum members and contributors, it got me really thinking. Especially about focus on cleaning up the toxicity on the Forums. Some find it offensive. Some "silly". Others a great thing. But the more I think it through, I can't say I find the intent to be silliness. Or wrong. All I have to do is think of the number of innocuous threads (even things like Art posts) derailed and led into unsubstantiated, circular arguments to the point the thread was effectively derailed.

Simple truth is, there was a subset of people who posted for no reason other than to undermine the Devs, and this Community. One could argue that some have justified reasons, others have this fever dream that they can get PGI stripped of the License, and some other studio will ride in and save MechWarrior (despite no one touching it with a 10 ft pole for a decade), do these "contributions", in the future in any way benefit the Community?

In the real world, there are repercussions to undermining the order of a community. Fact is, on a Online Forum, there is ZERO "right" to free speech. These can at times, be amongst the most toxic forums. Individuals do not always provide constructive criticism, but look to tear the community down, be it because the game didn't unfold the way they thought it should, or because they think they are being a BD(i)H, or because certain individuals and groups simply like to watch things burn.

That is known as being a cancer. Cancer left alone, festers, and eventually kills a community and a game as surely as anything. To prevent that from happening to our community we need to try to support the efforts of Niko and the Mods, not fight them. If we try to minimize the unfounded negativity, and focus on constructive input, of both the good and the bad (as we perceive it), the general attitudes of the Forums, the Game and Community can change very rapidly, to a positive and welcoming experience for players, new and old. MWO has a good community, for the most part. With a few angry bad apples. But by and large, even those critical of the direction, or some decisions, are mostly positive contributions to the health of the Community.

Now.

That doesn't let PGI off the hook. In my mind, it now put the burden even more squarely on them. PGI has one less excuse for issues, for not communicating with us. The millstone of certain people rampantly attacking their every word, should be reduced or removed. So now they need to step up in a big way. Communicate. Deliver. LISTEN. And when things go sideways, as will ALWAYS happen, communicate some more. The same standards we the community are held to, the Mods, volunteer or otherwise, and Devs, need to be held to, ALSO.

Our role?

CONSTRUCTIVE Criticism. When and where there are legit issues, that can realistically expect to be balanced, fixed, etc? We need to, in concise and civil manners, bring it to PGI's attention. I have actually found considerable success just tagging Russ on Twitter, and instead of being confrontational, saying "hey, discovered this issue", or "hey, had this idea". Doesn't mean everything is handled the way I want, and doesn't mean I like having to go to twitter to get it done. Perhaps now though, Russ can actually get a word in edgewise on PGI's own forums, he might use it more regularly.

There are certainly things to complain about. Some of the Clan Balance ideas were just poorly handled. Invisible walls on 2 year old mps, etc. Yet I get the feeling that half the people doing the complaining, have never bothered to submit detailed support tickets about these things, either. And other things, we just have to accept are NOT going to change, like the Heat System. Sized Hardpoints, etc. Would they make it better? Almost certainly. And nearly 3 yrs down the rabbit hole, such a core tear down is almost certainly NOT going to happen. So stop cluttering every post opining about these things, and then decrying PGI as idiots. It just makes getting the things we can realistically expect to see fixed, harder to dig out and find.

And if that is not acceptable to folk? There's the door. Get off PGI's Boat. Go, live long, prosper. FInd a game that DOES make you happy, and go enjoy the rainbows and sunshine there. (Though I have a feeling many of the usual suspects, will be just as miserable and complaining on other forums too, as some folks are only happy when they are making things miserable for everyone, themselves included). But there is no reason to be HERE, making themselves, and everyone else miserable, and stirring up trouble.

So me? Yeah, there are issues. And I plan to keep calling PGI out when I see them. But it is past high time this town had a Sheriff, and the riff-raff got the heck out of Dodge.


As for the continued life cycle and viability of this game? Can't say my name is Nostradamus. But it would be silly for them, at this point to abandon it, now that the "hard part" is mostly over (finally, and yes, well behind schedule). PGI is a Dev Studio on the other hand, not a Mechwarrior Studio. If you expected them to spend every cent, every waking moment, from now til eternity on MWO, then you really need a reality check. They need to develop games to grow and have a chance at being stable and healthy, as a business. But as CW get's implemented, the actual staff demand for MWO will go down, freeing up programmers for other games, which are known as Revenue Streams, which keep the lights on, and all the titles running.

So if you are secretly hoping this new game fails, or that PGI loses the MWO franchise, you are really sabotaging yourself. BEcause if this new venture fails, MWO will almost certainly go dark. And since no one BUT Smith and Tinker/PGI were willing to actually pay for the license in the last decade of MWO, don't expect some major "good" studio to run in, scoop it up, and save the day. PGI fails, MechWarrior is probably dead, for a very very long time, if not permanently.

I am all for discussion here, for disagreement. But to hope and expect, as Niko apparently does, that it be productive, constructive, intelligent and CIVIL, is not in unreasonable.

I hope the Devs and Mods weigh in on this also, and speak their peace, correct misassumptions of mine, etc. THAT is part of healthy dialogue.





oh..TL;DR? Too bad. I'm not looking to give sound bites and CliffsNotes. Read and contribute, or do not.





Cheers, fellow mechwarriors. Let's see how CW unfolds, and I hope to see ALL of you, dirtside.



(this started as a reply, in a different thread, so you will likely see an edited version, elsewhere in the forums)

With the news on this post, looks like a new beginning, to me.
http://mwomercs.com/...s-igp-feedback/


It's not the end , unless we want it to be.

Every brand needs more than one product , it's normal on every company regardless of success or not.

Personally , i don't want MWO to end here because even if we have bugs and other stuff I LIKE THE GAME.

Every person on this forum that play this has it as their primary game is because they LIKE IT , it's very simple even if people put money on it (like me). I haven't seen a single F2P title out there that feels COMPLETED. Not a whitknight or doomprayer here , neither neutral.

Edited by strikebrch, 11 September 2014 - 09:43 AM.


#390 Matthew Craig

    Technical Director

  • 867 posts
  • LocationVancouver, BC

Posted 11 September 2014 - 10:08 AM

First off I'd like to thank Bishop Steiner for the tone of this thread I think it is very fair and we've never said we are perfect as devs, we're human like anyone else. This is the first thread I've seen in a while that has encouraged me to respond so again thanks for that Bishop Steiner.

The reason for my response is to help clarify some of the speculation around engine choices. Firstly allow me to clarify Russ' statement about the engine being CPU bound, it is true that our game is currently CPU bound, Is it true that it has to be that way because of the engine? no not really but there is truth in the detail allow me to explain.

One of Mechwarriors challenges is that being an FPS it is regularly compared to other FPS titles e.g. if Counterstrike can get hit detection right then why can't MWO. There are a couple of notable differences with MWO the first being that each Mech can field a wide variety of weapons, in a game like CS you don't fire 6 lasers and 2 LRM banks all at the same time with AMS going off etc.

This adds to the complexity of the simulation that the server is doing and the overall workload for host state rewind and networking this is partly what limits us to 12 vs. 12 currently and made it such a challenge in the first place.

Also Mechs are not constructed in single parts and don't behave as simple characters this meant that earlier in the project when Mechs were being setup the approach that was taken was to compose them of separate attachments to give us flexibility to animate separate components, blow them off etc. over time we realized that this isn't ideal for optimal draw call use.

Most modern engines CE3 and UE4 included use deferred renderers. This means that there are multiple passes that objects get rendered in, so any object using a notable number of draw calls gets amplified e.g. it has to be rendered in the ZPass, Shadow Pass, General Pass and potentially Post Processing Pass which can amplify the draw call usage.

Draw calls remain very expensive on the CPU side and much of the research and development currently going into Mantle and DX12 is aimed at reducing overhead for draw call usage (and API / driver overhead in general) so that games can use more draw calls and focus on creating great consumer experiences without having to get overly concerned with hitting draw call barriers.

So would another engine have had similar problems to CE3? potentially as mentioned the character editor initially controlled how we composed the Mechs and setting them up optimally for both performance and flexibility were opposing goals in CE3 where another engine may allow an easier path to attain both goals.

With respect to why CE3 or UE4 there are many factors at play, familiarity with the engine not being as large a factor as many would think as most modern engines are very similar in terms of feature sets and tools they provide though that isn't to say there aren't notable differences.

Everyone that cares to follow engine development knows that the last few years has seen major tectonic shifts in that area not only in terms of technology but also in terms of licensing and like any game studio we actively follow developments and respond accordingly. The new project has very different requirements from MWO and we evaluated from the available engines what was the best fit for that project and we have been very impressed as have many with the tool set that Epic has put together for UE4.

Does that mean we would want to do things differently for MWO, well the premise is unsound, in that the landscape as mentioned was very different when we evaluated tools for MWO and at this point in MWO development switching engines is highly unlikely for obvious reasons. As many of you can appreciate we have heavily modified CryEngine at this point to suit the needs of MWO and so it is hardly a stock version any more, also our goal is as it has been for a long time to attain a very stable and functional environment for all the players playing MWO on a daily basis; long gone are the days of Closed Beta and we appreciate that and aim every day and every release to make the game more stable and reliable.

We've only done one major engine update since Closed Beta and that was to ensure we got working DX11 support which we still hope to leverage further in the years to come, as Russ mentioned in the town hall so long as there are MWO fans that want to play we will continue to aim to improve and expand MWO.

We know we've had some major struggles in this respect and many have involved fighting with the engine technology e.g. the long standing Scaleform issue we had (disco polys); however, that can't be entirely laid at the engines feet as any new game exercises an engine in potentially new ways that the engine developer can't foresee. We feel that hopefully once we get on top of the current de sync issues bar any further issues that arise from CW we will be reaching a point of maturity with the technology for MWO where we can begin to refine the base of MWO e.g. further UI 2.0 improvements both in the form of fixing bugs and functionality.

We also know that there are many still running older hardware that would like MWO to run more efficiently than it currently does and truth is that we could likely do more with more art resources dedicated to assisting optimizing existing assets. The general request from the community though is for new assets (maps/mechs) which shifts lots of resources to new assets and doesn't allow for there to be much internal pressure to improve older assets.

Something worth noting that I would like to correct is the notion that performance for the game has degraded over time and that performance was much better back in Closed Beta. We gather telemetry on many aspects of the game including average user fps and the truth is that while there have been both jumps forwards and backwards in terms of performance the trend line is net positive since Closed Beta.

It is likely the odd regressions and the fact that it can be true for a small number of users that their particular performance may be lower than earlier in the project that leads to the conclusion that it must be that way for everyone when overall that isn't the case.

We will continue to do our best to ensure the game is stable and runs great and if you ever ask yourself if we (the devs) are aware of a particular issue or care the answer is overwhelmingly yes we do care and are painfully aware, if something takes a long time to get resolved it is generally either because it is proving to be very difficult to fix or there are other competing priorities (many of which the community don't get visibility on e.g. improving customer support tools, dealing with billing issues etc. etc.) that need fixing first.

Hopefully that helps to give some more detail to the debate.

#391 Joseph Mallan

    ForumWarrior

  • PipPipPipPipPipPipPipPipPipPipPipPipPipPipPip
  • FP Veteran - Beta 1
  • FP Veteran - Beta 1
  • 35,216 posts
  • Google+: Link
  • Facebook: Link
  • LocationMallanhold, Furillo

Posted 11 September 2014 - 10:12 AM

Thank you. It is nice seeing so much Teal text in the forum.

#392 Bishop Steiner

    ForumWarrior

  • PipPipPipPipPipPipPipPipPipPipPipPipPipPipPip
  • The Hammer
  • The Hammer
  • 47,187 posts
  • Locationclimbing Mt Tryhard, one smoldering Meta-Mech corpse at a time

Posted 11 September 2014 - 10:24 AM

View PostMatthew Craig, on 11 September 2014 - 10:08 AM, said:

Really in depth stuff.

Wow Matt, thanks for the response, and especially for the in depth break down of the Engine, in terms a layman can handle. So much of the forum rage seems to be claims about this and that and the inability if the programmers, etc, etc, etc. Yet none of those posts remotely approached the details as this, which for me puts the burden of proof with you and the rest of the "code-gnomes" who make my Stompy Robot Entertainment possible.

Would I like to see those pesky Spiders register hits more consistently? Would I like to see a little more consistency in Mech and Weapon Scales? Would I like to see other things? Sure. But this type of post also helps me as a non programmer to grasp why some things simply are no t a priority, why some things take so long to do, and where the real nature of some of the issues lie.

So again, thanks for taking the time to post such a thorough response on here. Hope to get some more future insight from other Devs, now that you and Paul have been so helpful!

Cheers!

#393 Rebas Kradd

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,969 posts

Posted 11 September 2014 - 10:37 AM

View PostMatthew Craig, on 11 September 2014 - 10:08 AM, said:

Hopefully that helps to give some more detail to the debate.


No kidding. Thanks a bunch.

So, if I understand correctly...map improvements are being held back by new map production, and UI2.0 is waiting on the same people tied up with desync/CW?

#394 Why Run

    Member

  • PipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 370 posts

Posted 11 September 2014 - 11:05 AM

View PostMatthew Craig, on 11 September 2014 - 10:08 AM, said:

First off I'd like to thank Bishop Steiner for the tone of this thread I think it is very fair and we've never said we are perfect as devs, we're human like anyone else. This is the first thread I've seen in a while that has encouraged me to respond so again thanks for that Bishop Steiner.

The reason for my response is to help clarify some of the speculation around engine choices. Firstly allow me to clarify Russ' statement about the engine being CPU bound, it is true that our game is currently CPU bound, Is it true that it has to be that way because of the engine? no not really but there is truth in the detail allow me to explain.

One of Mechwarriors challenges is that being an FPS it is regularly compared to other FPS titles e.g. if Counterstrike can get hit detection right then why can't MWO. There are a couple of notable differences with MWO the first being that each Mech can field a wide variety of weapons, in a game like CS you don't fire 6 lasers and 2 LRM banks all at the same time with AMS going off etc.

This adds to the complexity of the simulation that the server is doing and the overall workload for host state rewind and networking this is partly what limits us to 12 vs. 12 currently and made it such a challenge in the first place.

Also Mechs are not constructed in single parts and don't behave as simple characters this meant that earlier in the project when Mechs were being setup the approach that was taken was to compose them of separate attachments to give us flexibility to animate separate components, blow them off etc. over time we realized that this isn't ideal for optimal draw call use.

Most modern engines CE3 and UE4 included use deferred renderers. This means that there are multiple passes that objects get rendered in, so any object using a notable number of draw calls gets amplified e.g. it has to be rendered in the ZPass, Shadow Pass, General Pass and potentially Post Processing Pass which can amplify the draw call usage.

Draw calls remain very expensive on the CPU side and much of the research and development currently going into Mantle and DX12 is aimed at reducing overhead for draw call usage (and API / driver overhead in general) so that games can use more draw calls and focus on creating great consumer experiences without having to get overly concerned with hitting draw call barriers.

So would another engine have had similar problems to CE3? potentially as mentioned the character editor initially controlled how we composed the Mechs and setting them up optimally for both performance and flexibility were opposing goals in CE3 where another engine may allow an easier path to attain both goals.

With respect to why CE3 or UE4 there are many factors at play, familiarity with the engine not being as large a factor as many would think as most modern engines are very similar in terms of feature sets and tools they provide though that isn't to say there aren't notable differences.

Everyone that cares to follow engine development knows that the last few years has seen major tectonic shifts in that area not only in terms of technology but also in terms of licensing and like any game studio we actively follow developments and respond accordingly. The new project has very different requirements from MWO and we evaluated from the available engines what was the best fit for that project and we have been very impressed as have many with the tool set that Epic has put together for UE4.

Does that mean we would want to do things differently for MWO, well the premise is unsound, in that the landscape as mentioned was very different when we evaluated tools for MWO and at this point in MWO development switching engines is highly unlikely for obvious reasons. As many of you can appreciate we have heavily modified CryEngine at this point to suit the needs of MWO and so it is hardly a stock version any more, also our goal is as it has been for a long time to attain a very stable and functional environment for all the players playing MWO on a daily basis; long gone are the days of Closed Beta and we appreciate that and aim every day and every release to make the game more stable and reliable.

We've only done one major engine update since Closed Beta and that was to ensure we got working DX11 support which we still hope to leverage further in the years to come, as Russ mentioned in the town hall so long as there are MWO fans that want to play we will continue to aim to improve and expand MWO.

We know we've had some major struggles in this respect and many have involved fighting with the engine technology e.g. the long standing Scaleform issue we had (disco polys); however, that can't be entirely laid at the engines feet as any new game exercises an engine in potentially new ways that the engine developer can't foresee. We feel that hopefully once we get on top of the current de sync issues bar any further issues that arise from CW we will be reaching a point of maturity with the technology for MWO where we can begin to refine the base of MWO e.g. further UI 2.0 improvements both in the form of fixing bugs and functionality.

We also know that there are many still running older hardware that would like MWO to run more efficiently than it currently does and truth is that we could likely do more with more art resources dedicated to assisting optimizing existing assets. The general request from the community though is for new assets (maps/mechs) which shifts lots of resources to new assets and doesn't allow for there to be much internal pressure to improve older assets.

Something worth noting that I would like to correct is the notion that performance for the game has degraded over time and that performance was much better back in Closed Beta. We gather telemetry on many aspects of the game including average user fps and the truth is that while there have been both jumps forwards and backwards in terms of performance the trend line is net positive since Closed Beta.

It is likely the odd regressions and the fact that it can be true for a small number of users that their particular performance may be lower than earlier in the project that leads to the conclusion that it must be that way for everyone when overall that isn't the case.

We will continue to do our best to ensure the game is stable and runs great and if you ever ask yourself if we (the devs) are aware of a particular issue or care the answer is overwhelmingly yes we do care and are painfully aware, if something takes a long time to get resolved it is generally either because it is proving to be very difficult to fix or there are other competing priorities (many of which the community don't get visibility on e.g. improving customer support tools, dealing with billing issues etc. etc.) that need fixing first.

Hopefully that helps to give some more detail to the debate.



Question: Regarding the performance metrics. Did you all not reduce the overall quality of textures to increase performance. It was my understanding that the original game was more demanding than we have now. Does your analysis factor changes from your end. Also, can you reconcile that in anyway against natural progression of hardware. Obviously if the game stays static while use computers are increasingly improved, obviously you'll see positive growth. The real question is whether the positive growth exceeds what would be expected by the natural progression of technology.[/color]

I ask these questions with only the utmost respect, please do not construe otherwise.[/color]

Edited by Why Run, 11 September 2014 - 11:05 AM.


#395 Matthew Craig

    Technical Director

  • 867 posts
  • LocationVancouver, BC

Posted 11 September 2014 - 11:25 AM

View PostRebas Kradd, on 11 September 2014 - 10:37 AM, said:

No kidding. Thanks a bunch. So, if I understand correctly...map improvements are being held back by new map production, and UI2.0 is waiting on the same people tied up with desync/CW?


Map improvements are happening but at a trickle, UI 2.0 improvements are currently behind further additions to the UI to support CW phase 2.

#396 Bishop Steiner

    ForumWarrior

  • PipPipPipPipPipPipPipPipPipPipPipPipPipPipPip
  • The Hammer
  • The Hammer
  • 47,187 posts
  • Locationclimbing Mt Tryhard, one smoldering Meta-Mech corpse at a time

Posted 11 September 2014 - 11:30 AM

View PostMatthew Craig, on 11 September 2014 - 11:25 AM, said:


Map improvements are happening but at a trickle, UI 2.0 improvements are currently behind further additions to the UI to support CW phase 2.

But the real, important question is: Where on the priority flowchart is the addition of the Urbanmech? ;)

#397 Ghogiel

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • CS 2021 Gold Champ
  • CS 2021 Gold Champ
  • 6,852 posts

Posted 11 September 2014 - 12:04 PM

View PostJason Parker, on 11 September 2014 - 07:08 AM, said:


I never played Living Legends but since it is a Mech game on CryEngine as well I wonder if it saw similar problems?

I was a MWLL dev. Yeah we saw similar problems in CE2. Crysis was the defacto benchmark game for high end PCs for a few years there.

We were working under an era of different hardware. So while it was CPU heavy, it was less of a bottleneck because at the time 9800GTs were basically the standard GPU and Phenom2s could just about keep up with those. The 1gb of vram was actually a top concern during my optimization passes along with lowering DP in LODs where possible. The HUD ***** fps in MWLL so bad too, which I presume the flash was all cpu run. turn off your HUD and get like 50% fps boost. :<

Edited by Ghogiel, 11 September 2014 - 12:05 PM.


#398 Matthew Craig

    Technical Director

  • 867 posts
  • LocationVancouver, BC

Posted 11 September 2014 - 12:08 PM

View PostWhy Run, on 11 September 2014 - 11:05 AM, said:

Question: Regarding the performance metrics. Did you all not reduce the overall quality of textures to increase performance. It was my understanding that the original game was more demanding than we have now. Does your analysis factor changes from your end. Also, can you reconcile that in anyway against natural progression of hardware. Obviously if the game stays static while use computers are increasingly improved, obviously you'll see positive growth. The real question is whether the positive growth exceeds what would be expected by the natural progression of technology.[/color] I ask these questions with only the utmost respect, please do not construe otherwise.[/color]


I think what you are referring to is the default setting for many settings was dropped to medium from high at one point to help ensure that the average user got decent performance which will have included texture settings. Texture quality has always remained the same in fact we increased it from 1k to 2k textures around the time prior to clans so many of the newer mechs use 2k sheets.

We appreciate the fact that we want the game to look great but also run on a wide variety of hardware two conflicting goals we do our best to balance where we can. To this end there were some reductions in quality for low/medium settings e.g. the missile trail effect was greatly reduced, but nothing high or very high has been removed.

Its a valid point that hardware progresses therefore performance will climb over time, my point wasn't that we are perfectly happy with performance or to be interpreted as excusing the current state of performance it was primarily to counter the notion that performance has declined over time or that we've made the engine perform worse as that simply isn't the case. Hopefully that answers your question?

#399 Bishop Steiner

    ForumWarrior

  • PipPipPipPipPipPipPipPipPipPipPipPipPipPipPip
  • The Hammer
  • The Hammer
  • 47,187 posts
  • Locationclimbing Mt Tryhard, one smoldering Meta-Mech corpse at a time

Posted 11 September 2014 - 12:14 PM

View PostMatthew Craig, on 11 September 2014 - 12:08 PM, said:


I think what you are referring to is the default setting for many settings was dropped to medium from high at one point to help ensure that the average user got decent performance which will have included texture settings. Texture quality has always remained the same in fact we increased it from 1k to 2k textures around the time prior to clans so many of the newer mechs use 2k sheets.

We appreciate the fact that we want the game to look great but also run on a wide variety of hardware two conflicting goals we do our best to balance where we can. To this end there were some reductions in quality for low/medium settings e.g. the missile trail effect was greatly reduced, but nothing high or very high has been removed.

Its a valid point that hardware progresses therefore performance will climb over time, my point wasn't that we are perfectly happy with performance or to be interpreted as excusing the current state of performance it was primarily to counter the notion that performance has declined over time or that we've made the engine perform worse as that simply isn't the case. Hopefully that answers your question?

Answers mine! Though I must say the maps looking at videos seem crisper and more eye popping in Closed Beta, is that just an affect of all the particulates and smoke and such you have added over time? (as it seems many maps get little tweaks as I notice new things all the time)

Or is it the rose colored glasses of the "golden Age of Camelot" syndrome, where everything was better "back in the day"?

#400 Cimarb

    Member

  • PipPipPipPipPipPipPipPipPip
  • Caladbolg
  • Caladbolg
  • 3,912 posts
  • Twitter: Link
  • Twitch: Link
  • LocationA hop, skip and jump from Terra

Posted 11 September 2014 - 12:16 PM

Anyone that does not think this game is amazingly beautiful has never saw http://youtu.be/cZP1J1jTdfE





9 user(s) are reading this topic

0 members, 9 guests, 0 anonymous users