Jump to content
Bohemia Interactive Forums

Beugnen

Member
 View Profile  See their activity

  • Content Count

    40

  • Joined

  • Last visited

  • Medals

Posts posted by Beugnen

  3. Posted

    You have to remember that pathfinding AI is horrendously difficult.

    Agreed, but it has been well established since the early 90s by the fine work of Mr Reynolds of OpenSteer, Boids, SGI, Sony and EA fame, not to mention an open library is available. [1]

    [1] http://www.red3d.com/cwr/steer/

    http://www.red3d.com/cwr/

    http://opensteer.sourceforge.net/

    ---------- Post added at 07:13 AM ---------- Previous post was at 07:09 AM ----------

    I disagree that doing pathfinding AI that works is hard.

    The hard part is making it fast and resouce friendly.

    But all the programming in the world wont help you if you screw up your node placements ( or Nav mesh ) on the map.

    perhaps, but thats generally not the way to do it since it makes things behave like a AFX model racecar track.[1]

    [1] http://www.red3d.com/cwr/steer/

  9. Posted · Edited by Beugnen

    sigh. though I'm not qualified to do open heart surgery i can relay information from experts that it is dangerous. do you understand the metaphor yet? if you must know I've been building systems longer than your brief stint in the armed forces. and yet claims such as these are rather pointless since i'm fairly sure no university offers degrees in computer assembly the likes of which that occur at the shop-front level. the likes of dell or apple are a different story of course. i doubt you would be let in the door without recognised and proven qualifications.

    finally, urinating competitions are tiresome and that is the final word ill say on this matter.

  10. Posted

    not a bad article.

    it is no surprise that the more powerful cpu attains the highest benchmark in this cpu-bound application. it does not really prove anything. it is much like finding a cpu for a ray tracing application[1]

    i fear that A2 is going to be used by some as a CPU benchmark application when in fact it is perhaps just an example of poorly written code.

    [1] ray tracers do not use GPU

  11. Posted

    I disagree with both of your points. How many water-cooled systems have you and he built? His point was about water-cooling of which I happen to have a decent level of expertise in.

    If you do not like what I have posted then get a moderator or administrator to deal with it.

    what a ridiculous and irrelevant thing to say.

    just because i may or may not have installed water-cooling systems, i do know for a fact that open heart surgery is rather risky.

    so by your logic as an 'expert' water-cooler installer, a surgeon will recommend DIY heart surgery to be performed by anyone? i think not.

    finally on spelling, everyone makes a mistake. i'm quite sure if i google the planet for all your posts it is highly probable i will find an incident of shall we say accidental spelling too.

    oh and i don't need a moderator to fight such trivial fights, i am more than capable of winning any war of words, logic and basic common sense with you anytime my friend.

  13. Posted

    A +/- 5% or less margin for error is usually acceptable when testing hardware. And regardless of that, wouldn't you expect an improvement with SLI enabled, rather than a decrease, even if it is only 9%?

    Having said that, it appears that on my system the ARMAIImark benchmark is CPU limited, as performance online feels better while playing with SLI enabled.

    agreed, normally SLI should offer improvement.[1]

    how many test runs did you run though when you got the 9% improvement? is it reproducible? do you still get the same result after quitting and re-running the game? finally, try the test again after rebooting. this will reduce the error margins i suppose but more importantly whether the improvement was a curious one-off.

    let us know how you go

    -------------------------

    [1] there is a disturbing trend of games that are immune to SLI/crossfire. initially i was reasonably sure that we were told such things was in the driver level and that games did not have to be programed against it. there are conflicting reports about this today.

    ---------- Post added at 08:22 AM ---------- Previous post was at 08:11 AM ----------

    The game is very CPU limited.... even at 3.8ghz on a dual core (i presume E8400 or above).

    SLI when enabled makes the CPU work harder as the CPU has to do some of the SLI communication work... therefore there are less CPU cycles for the game.

    This maybe why you get higher marks with SLI disabled... however in the game you will notice better smoother fps with SLI as the game is not CPU limited in non town areas.

    I think the game needs some serious optimisation as there is not that much happening for such fast mhz CPU's to be bottoming out.

    Yapa

    interesting, but im not sure that is entirely true. in SLI cards are configured as master and slave. data is sent to the master and relayed to the slave via the sli bridge thus excluding the need for cpu or motherboard intervention. even in systems without the bridge, the motherboard is able to send data between cards in pcie.

    i couldnt quite understand what you meant by cpu, so ill just recap. it is widely known and proven that the game is cpu-bound. gpu is making insignificant impact to game performance.

  14. Posted

    don't believe everything you read on sites such as slashdot. for one thing they failed to observe the difference between the following core ecosystems in the game:



    1. performance in the so-called A2 in-game 'benchmark'
    2. performance in the sandbox modes
    3. performance during normal in-game

    1. - never appears to change regardless of settings or rig

    2. - performance sky rockets on the same rigs. therefor the game is CPU bound in-game

    3. is woefully slow considering what is being displayed and the little number of units on the map regardless

  15. Posted

    BS . . . .know what you are doing and you will be fine. Also, please use a spell-checker. 'mailfunction' is what Outlook is for.

    the OP is making a legitimate point about water cooling. in fact it could be applied to any DIY work not necessarily restricted to the computing domain.[1]

    finally, don't waste forum space being the spelling nazi. if it were an allowed cause then the vast majority of posts would be subject to this rule. this is a forum, not someones thesis. be the overlord elsewhere if you have nothing of significance to say.

    ----------------

    [1] DIY house improvement was in the top 10 causes for home injuries in the past 5 years

  16. Posted · Edited by Beugnen

    I have big problem.

    When i start game... Screen is black, but i can hear sound.

    I installed patch 1.02 and it didnt fix it.

    What can i Do?

    this happened to me. the game was displaying a screen resolution with an Out of range Vertical Refresh Rate[1]. i.e. attempting to display say 90Hz on a monitor that can only support say 60Hz for the given screen resolution.

    to fix (in my example below i use 60Hz for my 1680x1050 rez. check your monitor for recommended settings:

    .\My Documents\ArmA 2\ArmA2.cfg

    ------------------------------------

    language="English";

    .

    .

    .

    Resolution_W=1680;

    Resolution_H=1050;

    refresh=60;

    .

    .

    .

    --------------------

    [1] not to be confused with VSYNC which is merely displaying the frames inline with the next vertical interrupt. it is quite feasible to have 50FPS, have it VSYNC'd so as to avoid tearing and yet the refresh rate is 60Hz. thats why i laugh at silly people who insist they are getting 100 FPS and yet their LCD is only capable of physically displaying 60 frames per second anyway (60Hz) regardless what they set VSYNC=on/off to.

  17. Posted

    I'm am constantly Crashing to desktop when ever i try to play a multiplayer game (havnt tried singleplayer, but i'm sure it's be the same issue).

    I updated my graphics driver and my DirectX (altough i had already latest on that), but it still is crashing every 2-10 min on average, i think the longest i went was 15 min. The Demo didnt do this, it ran abit slowly but it did not crash, also i have graphics set to the lowest on everything.

    Here is the link for the 3 files as asked on the Sticky concerning crashes and such. i hope this can get fixed since i only bought this 2 days ago and would love to get playing this properly.

    http://rapidshare.de/files/47757429/ArmA2_crash.rar.html

    it is perhaps wrong to assume.

    we need to rule out if whether it is the multiplayer component or stand-alone component since the game obviously activates different functions in this modes. since you haven't tried singleplayer it is perhaps premature to point the finger at multiplayer.

    try playing the game single player, campaign or scenario makes no difference so long as no LAN or online play is involved. if all goes well for repeated testing sessions then we can rule out your system at least for running stand-alone play. [1]

    if then you try multiplayer and barfs each time then it would be prudent to suggest there is something wrong in online play. perhaps AV or firewall?

    let us know how your fair:cool:

    --------------

    [1] since this will obviously test you CPU, GPU, RAM, MB, OS etc

  21. Posted · Edited by Beugnen

    all its saying is that MS surpassed the 5 year span that was commonplace. so you're right they are longer now, but in reality even longer than you say.

    good consoles have a span of a decade these days (see PS2) - and if you keep up with reading about the current gen that is the exact plan. support for up to 10 years.

    agreed, the ps2 was a milestone in console history i think. yes i stand corrected there, it seems only MS releases every 3 years. ps2 --> ps3 was 5 years. :).

    ---------- Post added at 09:14 AM ---------- Previous post was at 09:10 AM ----------

    I don't suffer from "performance drops" under XP, I did under Win 7 and Vista. I'm sure that will be fixed before long.

    OFP/A1/A2 own the realism IP for "all encompassing Military sims". The only potential challenger is OFP-DR and I have serious doubts about that game.

    There is presently no game like A2 to challenge it.

    BS is far more "realistic" than A2, I'll grant you that, and as the DCS evolves, it may well become a challenger, but as it stands, it's one helo (but almost perfectly executed).

    Regards,

    Eth

    agreed my friend. A2 has wonderful potential, im hoping for a fix.

  25. Posted

    Your comparisons are not scientific tbh. A2 is not really comparable to BS or FSX for several reasons not the least of which being the ground detail (as I mentioned before) is not at the same level as A1/A2.

    im not sure what the emphasis is with ground detail. if you just wish to focus on one aspect (an inadvisable thing to do) then Fallout 3 and Oblivion have superior GD and present much better. additionally, when A2 is loaded in either of the two 'sandpit' modes, the ground detail is the same as 'in-game' and yet the FPS triples. ground detail in A2 is not the weakest link so the performance drop is therefore caused by something else. [1]

    your statement of FSX lack of ground detail is without verboseness to form a complete argument. unless A2 sports 15cm pixel textures, 10m ground mesh elevation data, 6,000 trees/square KM not to mention visibility at 100s of km then A2 ground detail is perhaps irrelevant.

    i hear reasons from others that 'A2 runs slow because of all the AI and units'. i say take a look at X2/X3. it models 1000s of ships and presents better FPS. others say 'its the physx', i say look at BS. finally others cheer 'its because of realism'. i say they have never used a PMDG product.

    there is a vast universe of software out there especially simulators, A1/A2 does not own the IP of Realism and some others actually do it better.:cool:

    ------------------

    [1] cpu-bound

×