More excited for this than Half-Life 3

More excited for this than Half-Life 3

I'll this to my cool things to find.

 

XP10.3 can use up a lot of RAM and one's FPS will be <5 long before XP10.3 would consume my 32GB RAM ... but I would MUCH rather have an FPS problem than a VAS problem of 4GB. I can solve FPS problems in time with hardware, 4GB I can't do anything about ... so one path is "no choice" the other is "a choice". I'll vote for having a choice every time.

 

This ^^^

 

  Well said, Rob!

I don't care about backwards compatibility and wouldn't mind starting fresh if that means a much better simulator.

 

This is how I feel, despite thousands invested in FSX and P3D compatible addons. Fortunately, the 3PD's I have invested in heavily, fully support P3D, and I am confident they will do what's necessary to make the transition to 64-bit. I have no problem paying to upgrade said addons.

 

I'm not going to dwell on this however. I just jumped in to P3D about 2 weeks ago, and want to focus on enjoying the current version, even if not "perfect." It's getting dang close though.

 

 

To the best of my knowledge there is nothing later - unless someone has found it. And this quote was after Saul's post  in the same Prepa3d thread..

 

 

 

Saul is actually just another customer that was selected to be a moderator of their forums. To my knowledge he is not nor has he been an employee of Lockheed-Martin.

 

It's on the AVSIM front page now siting the same source I did but I don't see you two nitpicking in the comments section... 

I never bother reading the front page and anyway the comments I are made here are unchanged.

 

 

The only thing 64 bit is going to fix is OOM  - 64 bit code for the most part runs a few % slower.

 

Doesn't having all that unused address space give the opportunity to code the engine differently to take advantage of that?  I can imagine how the texture loading engine for example has to be designed to move in and out of ram sufficiently to stay under the 32-bit limit, whereas that would be much less the case w/ 64-bit.  I can see where if you didn't change the engine at all other than to compile it to run in 64-bit this statement you make is valid, but much less so if you redesign the engine to exploit large amounts of physical ram.

Make no mistake, every redesign brings new bugs. If P3D x64 is a new sim as opposed to a new version, we will all start at square one, customers and developers alike.

As 64bit processors and supporting chipsets are now the norm and have been for some time on the desktop front, the architecture is such that more recent processors will process a 64bit instructions as fast as a 32bit instruction and in some cases 64bit will execute considerably faster.  Specifically 64bit has support for more general purpose registers so it can process small code loops MUCH faster than 32bit because there is no need to fetch data from cache or main memory.

 

The only real disadvantage is that the same data will occupy more physical memory space ... cache implications here.  But like I said, newer CPUs (and their cache) and chipsets work better with 64bit instructions.

 

64bit address space should also reduce I/O activity as more data can be loaded ... this also means that the GC process doesn't need to happen immediately.  Many people don't seem to understand that releasing resources (the correct resources) takes CPU cycles.  The less "releasing" of resource the faster the code will execute ... so less efficient memory use actually mean fast running code (more FPS).

 

So it is very likely a 64bit P3D will actually run faster than a 32bit P3D even if it's just a straight translation without any additional optimizations applied.

 

Personally I think 64bit is long long overdue ... heck, even my iPhone 6+ is 64bit.

 

Cheers, Rob.

I like your "glass half full" look at things Rob!  Very refreshing.

 

I've been through all the changes since the FlightSim franchise was part of SubLogic on an old TRS80 Radio Shack color computer...my copy was on a cassette tape...tape drive folks...home use disk drives weren't even available or affordable yet...I can't remember which! 

 

Every change came with better and better visuals and options.  Yes, things will break in the 64-bit world.  Yes, there will be bugs (what software 32 or 64 bit has been bug free?).  Nothing new there...it happened when FS when from FSII to FS3.1 to FS4 to FS98 to FS2000 to FS2002 to FS2004 to FSX!!  At least there is active development to sort out the bugs (so far!!  LM, don't make me eat my foot!!  lol).

 

Developers who stayed in business changed with the times.  Anyone remember LAGO?  Exactly! 

 

I'd rather have a square one to start from with a new platform that have the last crumbling square that FSX is.  Don't get me wrong, I've had many, many years of enjoyment from FSX.  I'm not a huge tubeliner pilot, I like the low and slow, but I do dabble with the big iron and I like scenery.  But the current 32-bit FSX is nearly a 10 year old platform...it is 10 yrs if you count the years of development.  A DECADE!  That is a long life in software terms, so I'm not knocking it!

 

But Rob is right...with today's computer architecture, it's a no-brainer to start using what you paid for, don't you think?  Everyone spends huge dollars and buys the biggest and baddest 64-bit processors, 64-bit OSs, oodles and oodles of the fastest RAM man (or women) can make, but yet they resist upgrading software to 64-bit that can use everything their machine can offer it?? 

 

Count me in the "willing to give it a shot" crowd, bugs and all.

 

If you build it, they will come! 

Hopefully P3D 2.x is a test environment for a bigger overhaul done in 64-bit.   I've always had the sense the various releases of 2.x are almost trial balloons to test various approaches to different tasks, like how to code for cloud shadows so that they don't kill performance as they can now, etc.  Hopefully LM will announce where they are going on no uncertain terms so that 3PD's can plan ahead and count on the new platform being around long enough to invest in something new.

It's worth remembering that Microsoft 64-bit Windows uses the LLP64 data model.

 

This means only pointers expand to 64 bits and all other basic data types (integer and long) remain 32 bits in length. This means there is no change in the size of data on the disk, data shared over a network, or data shared through memory-mapped files. The consequence is that the size of this data doswn't change.

Well I don't think the entire product needs an overhaul (I used to think a from scratch would be better until I got caught up on the quadtree approach) ... P3D is simply accomplishing (or trying to) A LOT and is for the most part pretty successful.  It's the view distance that is the killer of performance ... but flight simulators have to be this way.

 

 A from scratch re-write could take many many years.  But a code migration to 64bit with a few adjustments along the way would take considerably less time.  I don't see the "need" to dump the quad tree approach to the virtual world ... which is the heart of the product.  XP10 uses the same quadtree approach and even Outerra is quadtree based ... it's a well known and solid solution to working a virtual world.

 

As far as LLP64 ... this is actually a good thing, especially for C programmers ... however, I would expect that much of the quadtree is pointer based so there will be a memory increase going from 32bit to 64bit.  I doubt there will be a need for "long long" change in data types ... but just conjecture.

 

Cheers, Rob.

 

 

 I don't see the "need" to dump the quad tree approach to the virtual world ... which is the heart of the product.  XP10 uses the same quadtree approach and even Outerra is quadtree based ... it's a well known and solid solution to working a virtual world.

 

Very interesting post Rob, can you explain (In lay man's terms) what the quad tree approach is? I've always wondered what the logic/approach behind trying to create a "world simulator" is.

 

I'm an architect and I do a lot of photo real renderings so I'm really aware of the missing pieces in the visual aspects of the sim, things like ambient occlusion, caustics from refracting through different materials, sun flare and light flare in relationship to the camera lens, subsurface scattering, etc. etc. And so from the eye candy side of things I'm always yearning for more but I recognize how difficult this is given the absolute complexity of the needs of a flightsim has.

 

I think the mistake many make when they ask "why can't our flight simulator look & perform better than...battlefield, skyrim, etc. etc" is they Don't realize that most of the scenarios taking place in those games are from 1 perspective/point of view but FLightsim has the unlike requirement of needing to have the fidelity of nearly a first person scale all the way up to the point of needing to see the curvature of the earth. With battlefield they have beautiful attention to detail and have all the elements of the area represented beautifully but what most don't realize is that the backdrop to all the maps is created with a spherical .HDR which is just a 2D image projcted to a dome surrounding the map and used as a light emitter. thi slooks incredibly realistic because it's based on a real world image but is only usable from a very small point of view.  

 

 

Very interesting post Rob, can you explain (In lay man's terms) what the quad tree approach is? I've always wondered what the logic/approach behind trying to create a "world simulator" is.

 

Yes Rob,  please explain (in lay man's term's) your last 3000 posts,  for me anyway     Your videos and posts are fantastic keep up the great work.

 

32 Bit or 64 Bit,  ill worry about tomorrow in 2 days time.  I'm having a ball anyway, just purchased A2A C182 and ill be neck deep in Orbtx, Rex and Clouds by Sunday.  I love X-Plane but i am really enjoying the different fruits that Prepar3d has to offer.  I'm finding it very steady no complaints.