Storm on Multi-Core Support

Source: Wotleaks RU community

Hello everyone,

the following has been apparently posted on RU supertester forums. Storm mentions here the multicore support and its effect and… well, I’ll just translate it.

RLQ8_X-ix88

Storm: “Multicore support – that’s wishful thinking and a myth about great ingame performance, spread by players. God willing, we’ll get 20-30 percent increase out of it. It was not done earlier because it requires an overhaul of practically the entire engine.

66 thoughts on “Storm on Multi-Core Support

  1. *Looks at Battlefield 4, Metro Last Light and Crysis 3*
    *Looks at settings at which I am playing World of Tanks*

    Yeah. Implement the damned multi core. It is embarassing.
    All those 3 games I mention look a generation apart from WoT, get higher frame rates, and even run on higher settings (though their LOWEST settings >>>>>. WoT on Ultra).
    Just… Do… something…

    You cant get your ass handed to you by modders playing with a 14 year old engine. Fucking STALKER Lost Alpha looks better and plays better…insane…

    Still, good to hear its still progressing.

    • Yeah it is, but supposedly when BigWorld hits 3.0 (Currently 2.42), multicore support will come. The guys at Bigworld (Owned by WG) are supposed to be working on this. If multicore support comes, it will be in 1 to 2 years.

  2. All i read is “mimimimimi, our Engine is old and bad but i wont admit it” Man up Storm, your game could be so much better with an advanced engine and not this bigworld crap.

    • I seriously don’t get the point of WG : they know something that could increase up to 30% performances and they call it a “myth” ?

      Seriously, working in computing for 2 decades now I can assure that 30% boost in just HUGE, if something you are creating can be 30% faster / better, you HAVE to do it, unless costs are way too high.

      • hey,
        first time they even admit that multicore might actually improve something!
        So give em time (LOL, yeah even more) they will get there, i see this max 30% increase as a first step to actually do it!
        mfg eXterminus

    • yes, a 30% performance boost would mean everyone who currently plays with 20 FPS would suddenly get a good framerate.

      • yeah optimizing the engine should be top priority and not adding new tanks imo. But I guess it costs too much or they think it will take too long to do it. BigWorld engine is very poor. Had a western company made the game I think it would have run better with a different engine of course.

        • Do i really have to remind u how WoT would be like if, let’s say EA made it….
          im quiet happy with WG on my part, wot still by far bestest game out there for me!
          Have you played anything as long and intesive as wot?
          Remember your beginning, can’t remeber any game getting me so exited and addicted like wot from my first match onwards..
          mfg eXterminus
          PS: Not saying that there are a lot of things that could be way better, just saying that wot is still fun as hell!

        • Do you realize that guys who make tanks, guys who make maps, guys who change interface, guys who make physics, guys who work on graphics and guys who work at overhauling engine for multi-core support are different people and they can do that sort of simultaneously?

  3. I am afraid that 2014 is their last chance to solve the ancient engine situation.

    If they do it in 2015 or even later, their competitors will be already miles away with more advanced and already optimalised engines and it will be impossible for WG to catch up.

    • If your target customer is average daddy, you have to focus on optimalisation.

      In long run, you can have:
      a) poor visuals but highly optimalised engine – everyone can play.
      b) great visuals but demanding engine – everyone wants to play (and is willing to upgrade his PC for it)

      Currently there are poor visuals with demanding engine without multicore support

  4. Their biggest player base is Russian/ Eastern Europe. Why would they spend programmers time and costs for small market share of good computers when their biggest market share is running “Potato” powered machines? They can use programmers time to make more lovely Soviet vehicles, and more premium tanks!. Yes, they maximise profit! Return on investment, return on capital!

    The only reason you are seeing HD models now, is the threat that was posed by War Thunder. They won’y change anything unless they see their revenue stream decline. What is needed is Half life 3, Fallout 4 or another potentially epic title being released that will empty their servers in the EU/NA. Then they can have all those lovely “free to play” players, putting nothing into the game and their pockets!

    • The very maximum your epic/hyped titles do is put a small dent into the player numbers for a short period of time – as long as it takes to play through any singleplayer title, maybe even more than once, sooner or later people are done and come back to wot.

    • Well AW is made by Obsidian of Fallout fame. This could give them the competition that will hit their bottom line in Europe and the rest of the world. We can but hope.

      • My responce Mail.ru. Ever play Warface, watch as everyone is stuck in Pattons while a person who payed something like the cost of 1 year of prem for a Special Premium Prototype T-XX with a giant Russian flag that fires special rounds and has uber armor. And to to all those that didn’t play Warface, there is no health regen and your stuck with a rusty M16, but when you pay you get special crown armor with health regen, headshot protection, and an AK-12 “Spetznaz Edition” that fires “Special” bullets.

  5. multicore support, will it solve the tomatos and bots problem? nop.
    so who cares….

    it means more retards will have the chance to better fuck up the game as never seen before….

  6. Jesus christ just either implement multicore or just go ahead and overhaul the game. The bigworld engine makes my computer look like a calculator and it also has worse graphics than the lowest settings on some of the newer games.

  7. Storm being full of shit, as always, I guess. I agree in so far, as that +20-30% are actually reasonable numbers for multicoring versus single coring performance, but what he misses is, that every increase in CPU performance for a game like World of Tanks is immesurably useful for the end-user, because the CPU is pretty much the single biggest chokepoint (hardware-wise) for the game as it stands (well, that and HDD use). It simply doesn’t matter how much they increase the game’s performance in other regards, because that chokepoint will still be the bottleneck for the vast majority of the users.

    Also: FIXING YOUR FUCKING ENGINE SHOULD HAVE BEEN TOP PRIORITY LONG AGO! This Blog had a nice series detailing, what the client does and what it should better do instead. Why wasn’t that done yet? Why wasn’t any single one of those glaring issues adressed earlier?! Why is it, that your fucking software has to drop on your fucking feet first, before you guys move your asses from denial, to vague acknowledgement, to acceptance, to planning to address it at some point, to finally fixing it (but only somewhat – hello Ghost Shells) time and time again?!

  8. To be honest, he is right – if we are talking about multicore support in general only. But this game needs complete overhaul anyway and even without multicore it could benefit greatly. It looks like it’s not even unoptimised but disoptimised on purpose, the level of bad programming was even proven here on ftr…

    • I see you are falling for the “care balancing” = “multi-core support” lie.

      If it were “true” multi-core support then BOTH cores would be maxed out instead of both running at HALF of what a single core does.

      You should see “core balancing” on my six-core AMD… each core is like 20% usage = NO increase in FPS.

      • lol? Why should they be maxed out? This game isnt so power hungry that it needs full potential of two haswell cores :))
        my friend plays this game on haswell pentium hehe

        multicore support doesnt mean that game will utilize automatically 100% of performance of core …. it will use it, when it needs to and this game – atleast on intel sandy/ivy bridge and haswell – doesnt really need more than two cores at 3+ghz
        when there will be 4 threads for WoT, i expect each of my cores running at 30%

        about AMD to put it simple, youve bought a cpu, which architecture is ahead of common game development in terms of thread count and few details… most of the games today are made for 2 to 4 threads … when the day comes, and a 90% of games will use 6-8 threads your Cpu will be properly used, but by that time it will be outdated …. youve simply bought something, that is not properly used in todays gaming and you pay the price …. FX is good investment in gaming for modern titles like Crysis or Battlefield, where those engines can make use of all their many weaker cores

        • They should be “maxed out” because the CPU is the limiting factor in performance. I have desktop gauges on my 2nd monitor that display, in real-time, the performance and load of all my hardware (CPU, GFX, HD, RAM, etc). What it shows is that everything is running at less than 50% while the CPU is pushing 100% on a SINGLE core (or the fake 50% x2 cores “support).

          • We got quite different cpus and mine can roflstomp yours in terms of perf per core… it doesnt add, that FX doesnt even have proper core ….

            Based on my experience and data ill guess, that for FX 6300 to unbottleneck the WoT, you will need four thread optimalization aka four cores of yours to do the job with some reserve like i have …. thats where your problem is …

            that doesnt mean i need four cores too …

            anyway, i was simply putting out, that WoT is running atleast on two threads now … its nothing new, but i just cant read about one core bulls..t, from 9.0 its a two thread

            about the graphics.. for my 60fps cap, my Radeon 270x never stepped above of 70-75% of GPU usage

            to play this game on full detail right now on Full HD its enough to have something like Haswell Pentium 3ghz+ and R270x/660oc NV

            its not a problem, if you know what game needs

            • Everyone except the deniers at WG knows that this game does not “really” support multi-core. The entire Render Engine runs on one core. It might get improve with DX11 support which handles multi-core better, but I doubt it will help much.

              btw, I have WoT running on Cores 3 & 5 thereby ignoring the “shared” core resources and have my FX-6100 OC’d to 4.3ghz. I have a GTX 760 which should have the same FPS on the highest settings that it does on the lowest while the card never goes above 50%. However, something about the way the game engine runs causes the CPU to do things the GFX card should handle.

              Currently the entire rendering engine runs on a single core while the chat and audio runs on a separate one. There are MANY other components of the rendering engine that could be run from a separate core with little impact on synchronization. Items like most mods, vegetation rendering, environmental rendering (buildings, clouds, lighting, etc), could all be moved to another core.

              Long-story-short, the game engine in ancient and VERY CPU dependent for performance.

            • SpudmanWP is right, your’re wrong. 2 cores, I laugh. That’s only true in a technical sense. One core is maxed out while a second core doesn’t break a sweat with how little it’s used. Meanwhile, I can’t even stably hit 120fps and max out my monitors refresh rate and that’s running a 2600k@5.0ghz. The game is so ridiculously CPU bound, you can straight leave the GPU out of the equation on a good rig, (mine runs at 25-30% whilst the GPU fan never goes past low speed.) If “it’s a not a problem, if you know what the game needs”, I’d be able to hit a stable 120 fps. I know what the game needs. Either a CPU I can overclock to 5.5ghz+ or multicore support. You’re making the mistake of thinking that if it’s good enough for you, it’s good enough for everybody. You’re also ignoring the fact that, if “[my CPU] can roflstomp yours”, then real multicore support would help people with lesser rigs as well as higher end systems, (no matter who is the CPU maker.)

  9. Exactly what I’m trying to tell people all the time. There is no simple magic switch, and this is not a cure-all. Of course you only get downvoted for this.

  10. It would be nice if they were working with a new engine on the side, adding tiers 11+ or something, and just update the old version until the new engine is ready.

  11. If it requires changing the entire engine, that just goes to show how outdated and archaic the original engine is.

    Frostbite and Crytek have been doing far more for far longer.

  12. Storm kinda sounds bitter that they have to add multicore support. In his mind I am sure hes going, “Fuck you twats and your high end computers. Why don’t you still have a P4 Prescott like me?” And since when is 20-30% increase in performance some trivial thing? Wargaming.net has never increased the performance across the board by even 10%. Any improvement that has been seen has come directly from the GPU manufacturers driver updates…

    • Come on….. Did you check the engine analysis that was posted here? WG could easily fix many of the zips-within-zips and frequent reading of data from the disks and constant packings and unpackings that the engine performs…. If the game stopped relying on disk dives less and more on RAM, we’d see a much larger improvement.

  13. The GENERAL OVERHAUL would be possible but it would require WG to trust their players somewhat:

    Imagine them announcing:
    “We are going to focus solely on the complete and total transfer of WoT into Havok and new serverside engine. Because of that there will be no new content for next 6 months after which the transfer will be done. The new system will include better graphics, larger maps, faster server side calculations, better physics, and multicore support up to 4 cores. We hope the community will support us in our decision”
    Everyone would go fucking bonkers. But they think that players are like cats chasing laser pointer light and as soon as something new and shiny vanishes they will just stop playing and abandon the game, which is of course false.
    So they stick with the old, working shit. Because why try, why care.

    The fact that company employs some 10000 PR people and community managers and only 12 programmes, also does not help.

    • You are making the assumption that content improvement and engine improvement (ie new/rebuilt engine) are mutually exclusive actions.

      They can and are done completely separate. In a way, WG has already started down this path. Recently they announced that new HD models are being developed to a SPEC that is many times higher than WoT can support. When the model is done, they run a script that translates it to the current HD spec. Later, when the spec is raised, the already-built model is re-converted to the new spec thereby alleviating the need to remake the model. Every aspect of content can be done in a similar manner.

      There is also the issue that content programmers are not engine programmers so their work load would not overlap. In other words, your content programmer does not need to stop content work in order to work on the Engine.

      • Well, Wargaming has actually split development of engine and content.
        Updates 8.x saw a lot of new content, including new nations, new tanks, new maps.
        Updates 9.x were announced to deal mostly with upgrades in mechanics, graphics, sounds, and other engine improvements.
        So, it’s not us that want Wargaming to choose between upgrading their engine or launching new tanks and branches and nations. No, we want BOTH. It’s WG that is taking that decision.

  14. Adding multicore support for 20-30 percent beter perfomance?
    TOTALLY WORTH IT!

    Having to rebuild the old engine from the ground up to enable it?
    TOTALLY WORTH IT!

    WoT is one of the companies that get more revenue per player. I don’t see what’s stopping them. Their competence is years ahead in graphics, performance and compatibility. Stoppied due to a lack of will means commiting corporate suicide.

  15. Whoever jumped on the “20-30%??? FUCKING DO IT!” bandwagon is outright retarded and is the reason why WG does not like talking about this. People don’t understand how the game is coded and what time elements lead to better performance.

    Fixing the graphical processing will probably yield a lot more than 20-30% efficiency and would require only a fraction of the time it would take to implement multicore. If they do implement multicore, they’ll most likely end up breaking every single part of the game and it would take around 3-6 months of testing just to get it back to it’s original state, and what you’ll see is more problems, not improvements. Multicore support in AAA titles like battlefield 4, etc. have had problems related to syncing and SMP processing, so they aren’t perfect either. Multicore is NOT the answer to fix performance. Just because you can’t use all 8 cores of your 2000 dollar processor doesn’t mean WoT isn’t ‘utilizing all your resources’.

    • Are you kidding? Even your stupid phone is multi-core now; future computers will be multi-core; Hell they can even squeeze longer life on WoT Xbox.

      Think about it, 20~30% performance increase also means the game can look nicer on cheaper computer, the very thing that they wanted to keep a large player base.

      It really comes down to the fact that they will need programer to work on the engine, or pay for multicore Big World that they make so much butt hurt about.