Jump to content

zorlac

Members
  • Posts

    411
  • Joined

  • Last visited

  • Days Won

    2

Everything posted by zorlac

  1. Incredibly hard. Its not just a simple case of patching a bit of code, it would be easier to rewrite it from scratch to support multi-core/multi-cpu. The problems that appear when you start running code on seperate cores is vastly more than just twice as difficult. I have been playing (briefly) with multiple threads in the upcoming version of TouchBuddy, and its an absolute nightmare to back-port it to use the multi-core/cpu efficiently.... so I only used it where I really needed it, and even that small routine was a horible horrible few days of coding. You have to ensure that everytime one part of the program access some shared data, that that data is tested to see if it can be locked (or wait until it can), then attempt to lock it (and wait if it cant) then read or write to the shared data, then unlock access to the shared data, so other parts of the code can access it. If you dont lock your shared data, you end up with one part writing while another is reading, and one or the other will obtain an invalid result. You also have to make sure that if you are blocked from access by a lock, you have to wait until you can, or go off and do something else and come back again. Debugging multi-core/cpu code also becomes a bloody nightmare... OK, your code just did something odd.... which part? it might not be the part you think because some of it is running separately, and it might be that part that is crashing. Was it the part that wanted to read some data, or the part that wanted to write to it? if you have more than 2 "threads" going, which one of the multiple threads was it that crashed, and why? if thread #4 crashed, which thread was it that it got upset about, or did it crash all by itself? I am hoping that Tank Killers will be written from the ground up as a multi-core/cpu capable game, but I would be very very suprised if LOBS is multi-cpu/core aware.
  2. holy cow!!! And it looks more stable than my 25T :music_whistling:
  3. Release of v1.3.0 is getting close now... just some last issues and testing of current builds, plus bug fixes. I did post the release notes for v1.3.0 here: http://www.touch-buddy.com/forums/showthread.php?t=558 but obviously these will be updated when we do release v1.3.0 I feel very confident about this release, there is a hell of a lot of new stuff in it, and the release notes for v1.3.0 are as long as the release notes for all previous version put together!! We are in "lets make videos" mode at the moment, so expect a few more to be released, showing various aspects of TouchBuddy in use with games, showing advanced features, methods for creating profiles, watch the Toolkit in action, etc. By the way, TouchBuddy will be exactly 1 year old on New Years Eve, which was when I first had that crazy idea... looking back at the stuff that I have had to learn, and what has been achieved with TouchBuddy and WBK's Toolkit, I am just amazed :)
  4. have you tried adjusting the TIR profile curve to remove the "dip" or "deadzone" in the middle of the range of movement?
  5. found the broken map (flotzilla.mis had been corrupted)... replaced it, patching server, will be back online again shortly
  6. could this be due to the problem with network play and a dual core CPU? I always run my lomac using procaff (mostly because starforce crashes the entire machine if i dont), but I know some people here have mentioned an online problem with dual cores (and some people say it doesnt affect them). Just a thought...
  7. Bob @ RC Sims (UK) mentioned that the x52 Pro had Hall sensors, so an improvement over the old pot technology. I can see it being a good replacement if your existing joystick went pop, but its pricey if you just want to upgrade from your existing and fully functional x45/x52/Cougar (or whatever)
  8. I had an X52 for about 2 hours a year ago, but it had a huge dead-zone, so I returned it, and have been happy with my springless x45 ever since. Will wait until either my x45 goes pop, or until I get my hands on a x52 pro to play with, and then make a decision.
  9. dedmac (i think) tries to move the server's screen to map view, but the VVS504 server have that disabled, so it just sits there watching an airfield as a spectator. dedmac doesnt allow you set the "server" player to join one side of the other
  10. My company will be billing you for the expense of cleaning my desk, there is now partially eaten sandwich all over my keyboard and screen :lol:
  11. 512kbit/sec is probably good enough for 8-10 clients without running into any major lag problems. On the server, tweak the "MaxFPS = xx;" option (doesnt exist by default) in graphics.cfg file so it spend less time updating the graphics, and more time running the code. Both the VVS504 servers run with "MaxFPS = 5;"... it doesnt mean the connected clients only see 5fps, it just means it updates the server's own screen less often.
  12. Hurraah!! Getting Vista to run in VMWare has been a real mission :(
  13. which link are you talking about? (Currently pulling 600Kbyte/Sec via my DSL link, and 2.5+Mbyte/Sec over my corporate site link)
  14. Another mirror, if that helps... the lockonskins link is dead, and the jabog server is going into meltdown from the VTB movie :) http://www.syn-ack.com/lomac-hosted/Project_50.wmv
  15. I added an extra Gig of memory over the weekend, and LO:FC Starforce barked at me for the CD.
  16. oh, thats right, rub it in why dont you :P :joystick:
  17. Is this "hall effect" sensors, instead of the usual pots?
  18. then just lock the joystick "twist" so it doesnt do that anymore. I am tempted by this, but will wait and see. As I have a long term hand/thumb injury that the long hours of flying tend to aggravate, I removed the spring totally from my current x45. Will need to see if the spring on the x52 is light enough... or maybe i`ll just buy it anyway and butcher it :) and the SDK... am curious about that too...
  19. yep, its configurable to show basic TEWS version or the "extras" version...
  20. TouchBuddy website seems to be up again (I dont host it). I have just got back from a 3-4 day LAN and social weekend, and am just shattered. I had some good coding sessions while I was there, but spent far too much time relaxing and chatting, and not as much coding as I was planning (I didnt manage to get around to starting to engine/fuel management active module).
  21. yes, I am fully aware of the limitations of the current LUA export system.... its the future stuff I am hoping that will be coded into BS that want to get to play with (and really woulddnt take that long to add into the core game code). If its just the same LUA data we currently have access to but in a new package, we are OH so screwed for external MFDs, screens, dials, Shkval/ABRIS, etc....
  22. looking at the lua code posted on that site, its partially incomplete: at startup, they open a tcp connection... at close, they shutdown that connection. In the "LuaExportAfterNextFrame" routine, they do obtain some data from the game... but do nothing with it (apart from Barometric Altitide -> "socket.try(c:send(string.format("Arn.Resp:15=%.0f:\n",altBar*0.032827))) ") Spitting this data out after every frame gives rapid data update, but also slows the game machine a lot. Or maybe I`ve just eaten/drunk too much this weekend and I cant read LUA code properly right now :)
×
×
  • Create New...