Batstat

According to steam hardware survey august 2009 17% steam users have 64bit os installed, and over 42% of the users have 3GB or more ram installed. Both factors are rising each month. http://store.steampowered.com/hwsurvey/

At least BIS could post a answer explaining why they cant do it. Enabling /LAA and then call GlobalMemoryStatusEx to find the systems real memory cap seems to be a more valid strategi with more benefits versus hardcoding a cap at 2047MB? Can some BIS dev enlight us here?

I did notice that the cpu under max load newer used more than 40%. Normaly between 30 and 40%. Disabling HT on the server seems to be vice. Sorry for my off topic remark Proton, but your mission is huge, and it will tax the server to it's limits. I hope you continue to develop the mission :D

I did som testing, with my 8GB Q9550 server win 7rc64. If you have enough memory, using the -maxmem=2047 on the server make a huge difference. A game that normaly go down after 1,5 hours could using the default settings, could last for 4 - 6 hours using -maxmem=2047 on the server (I don't know what the default settings are) I tested it using a ideling approch, once in a while I did use fast travel to engange the enemy, only to check that they did respond. When the server used close to 1,8GB with maxmem=2047 it started to become sluggish. Maybe change the garbage (salvage) routine. Wrecks that are to old, should be removed, or the salvage operation radius made bigger. Remove the need to drive close to the salvage wreck target, if that is possible. Same for ammo if thats a problem. Fasttravel did work, I guess one or more playing the game did use one of the provided beta patches.

I have 8GB ram, how do disabling the swap file tribute to Arma 2 ability to use more memory?

Epic :627:

Rerun 3 The server newer went down, but it become partly unresponsive after approx 6 - 7 hours. Server FPS was between 14 - 17, and some Ai controlled units was stuck others was not. The total amount of memory the server at this moment did use was close to 1,8GB, VS was 1176 736K. Seems as the /LAA switch don't make a difference. The amount of max memory the server can use seems to be hardcoded. BIS has newer said anything else, but I like to see that forcoming rev of the server do utlize the /LAA switch. The test also showed that it's hard to pinpoint the exact moment when you run in to server problemes. Better missions garbage routines could help as a solutions until BIS do make use of the /LAA switch. I do not belive making the test without /LAA switch will prove anything new, but for the sake of testing I will do it, next weekend. BIS could you please start using the /LAA switch? :bounce3:

Did some testing yesterday, not done yet. My server os win7 64 bit cpu Q9550 ram 8GB 1.03 patch Steam Testing Hosting Warfare Proton Mods at the dedicated server http://forums.bistudio.com/showthread.php?t=80346 This map will normally kill your server, also when you are idle and only let the AI fight. Using machin nr 2 as client, only log in, choose a slot, start #monitor 60 and waiting Using process explorer fromhttp://technet.microsoft.com/en-us/sysinternals/bb896653.aspx to monitor the server memory usage 1. run was with no parameters to the dedicated server Min - RAM -Virtual Size 010 - 544MB - 863MB Virtual Size (VS) 020 - 540MB - 863MB 060 - 554MB - 928MB I was not there to observ it, but atfer one hour the server had stopped to respond. The AI under controll of the server did not die etc. Did not check the rpt file, no error message on the console. 2. run This time I hade applied the /LAA switch to the server exe. More or less the same result as run 1. Use of the /LAA switch did not improve anything, but this time the error msg Out of memory (request 1KB). Reserved 292912KB. Total free 1KB, Free blocks, Max free size 1 KB did come on the server. 3. Run with /LAA switch on server and -maxmem=2047 on the server First thing I noticed was that the Virtual size after mission startup was close to 1GB or 1019MB. This is strange, why didn't the server use maxmem 2047 as default when there is 8GB to use? After one hour the game was still running 060 - 587MB - 1084MB VS I was not able to test more, Need to rerun test 3 and observ when the server goes down. Then rerun test 3, but then without /LAA Biggest impact in this test was to enforce -maxmem=2047 at the server, in my opninion this is strange.

Scratch the beta error report. This is a known beta bug.You can't log in as admin, when using beta against dedicated server not running the beta. Somebody told me yesterday.

Hi Shark my feedbacks (played tonight with Noraf). 1. Most players did die before landing and after they opened the parachute. All died at the same moment? 2. Our server did crash before we did finish it. Look at the rpt ending: Exception code: C0000005 ACCESS_VIOLATION at 0088501D Version 1.03.58659 Fault address: 0088501D 01:0048401D D:\ArmA2\ArmA2Server.exe file: co20_Eod_revive_v5 (__cur_mp) world: Chernarus Prev. code bytes: 45 CC 8B 75 FC 83 C0 24 8B 00 C1 E6 02 8B 3C 06 Fault code bytes: 8B 07 8B CF FF 50 08 84 C0 0F 84 C2 00 00 00 8B And this from the crash dump file: Dump Summary ------------ Dump File: D:\ArmaDownload\serverRPT\ArmA2Server.mdmp Last Write Time: 27.08.2009 23:41:44 Process Name: D:\ArmA2\ArmA2Server.exe Process Architecture: x86 Exception Code: 0xC0000005 Exception Information: The thread tried to read from or write to a virtual address for which it does not have the appropriate access. Heap Information: Not Present This exception is explained on MSDN with: The following tables describes the exception codes that are likely to occur due to common programming errors. Go figure? ---------- Post added at 11:20 PM ---------- Previous post was at 11:11 PM ---------- Forgot to mention, after putting up mobile respawn, we was not able to reviev at least one guy and he did "jip" in after we put up the mobile respawn. Spawned at the mobile respawn. Don't know if this are helping or not. The errors could be engine related also.

Hi thx for a excelent mission Proton. Tested it last night on our dedicated server. My inital comments: For new players you could consider to 1. In the readme, give info to server admins where to put your key and bisign files. And maybe how to activate it. 2. In the readme, link to a guide where game mechanics are explained, and how to access various menues, as player and commander. Optional link to a guide that have most in common with your mission. Two thing I noticed: Playing the mission with Arma II beta client against a plain dedicated server did hang my UI in the lobby, when I choosed none as game parameters. Consider to make a variabel lenght intro, where userinput are disabled while waiting for server to settel all the init script. Give the game to the players when the server is done with running startup script. This will lower the bad side effects of yellow and red chains, witch todoay is bound to happen after the mission start. Edit: The time for server to init, givs away server and to some extent network performance. Witch could be used to hint about what settings the mission best can be played at, if this is known numbers?

:bounce3: Wow I pray that I did not misguide the effort. If no one complain we will play it on the NoPryl server first thursday to come 21:00 cet. Thx

Played the mission yesterday - noticed two bugs? 30 - 50 seconds after we started the mission a meesage told us that the airport was under attack - and we should give priority to defend it. There was no hostile attack so we continued with our main goal to take the first town. From time to time this false message appear. After we wiped the first town and hq, then the message come again, this time it was real. The wait and preparation for attack is maybe to long. The engineer could set up defence when the teams are working on then main objectives? bug 2: The airfield was overrun and we respawned to a second location. This respawn point need to be adjusted, because many respawn inside the ammo crates, and some then die. Easy fix I believe. Not a bug but a feature: There is to much fog, and in my opinion the view distance is to short. Most of all get done with the fog - witch are present all to much and all to long, in my opinion. If this was real life, I also believe that the enemy to some degree would have chosen to place heavy tanks more on the outside of towns. I believe they start outside towns, but they are not patrolling the area, they drive inside town when we attack. When inside town the tanks is sitting ducks, and cant use the speed and big guns in any efficient way. But that's just my opinion. In short: Fix the AI attacking airstrip trigger/message warning Shorten the amount of time waiting for the AI counter attack Fix respawning inside ammo crates in respawn location 2 Reduce the amount of fog (or turn it permanent off), raise the viewdistance? Maybe adjust AI tank behaviour - and avoid driving them inside town to much. Also ensure that they always are moving. Good mission I believe, but we did not end it so I'm not 100% qualified to tell the tale.

I get your message Big Dawg KS. Maybe I should find another approch. The objective is to generate a huge amount of objects, witch do not tax the cpu or gpu to much. I want to test out what 64bit, huge amount of memory and use of /LAA count for versus 32bit XP. Running mission until the game freeze (crash) for out of virtual adress space. Hopefully "running" to the crash and not crawling. If the result could been presentet in a fraps video on youtube, then the result would stand more out, but thats not very important. For the readout of the carsh I would needet to supply the video with fotage form my digital camera. Firing enough gatlings gun at the same time up in the air would do. Anyway I need to find somthing that tax virtual adress space, that may not be bullets at all :). Could be my request was slightly premature. Requesting before thinking and reading hmm...

How can I set (config) one of the gatling's guns bullet with tracers to infinite bullet time. With that I want to be able to locate the tracer light (ergo the bullet) after the bullet has came to it's ballistic rest, and possibly after the same bullet ricocheting one or several times. I have newer done any config editing so this is total unknown territory for me. :bounce3:

Thank you very much Apocal I will explore your advise. If any know about tracer burnout time I will like your advice too. Inifinite or 6000 sec bullet time, don't give any meaning if I'm not able to visualize the result. :bounce3:

The /LARGEADDRESSAWARE flag don't rise the the amount of physical memory your app can use, but the amount of virtual adress space the app can use. In a 32 bit win OS the app can use 2GB of virtual address space and 2GB is reserved for the kernel. Rising the amount the app can use in a 32bit win os, will lower the amount the kernel can use, and this can lead to lower performance (more swapping) and crash. In a 64bit win os, the flag can raise your amount of virtual memory up to 4GB. The problem: 5 of 6 or what ever I don't know is running 32bit os, and most ppl has less than 4GB installed in the computer. BIS need to ensure that the game are runing fine for all of the varoius computers (and they need to foucus on the computers the most of us are buying). But as I understand it they can enable the flag under the folowing conditions. 1. Don't assume that the high-bit is never set in a 32-bit pointer 2. Thirty-two-bit applications that are large-address-aware can determine at run time how much total virtual address space is available to them with the current OS configuration by calling GlobalMemoryStatusEx. The ullTotalVirtual result will range from 2147352576 bytes (2 GB) to 4294836224 bytes (4 GB). Values that are larger than 3221094400 (3 GB) can only be obtained on 64-bit versions of Windows. For example, if IncreaseUserVa has a value of 2560, the result is ullTotalVirtual with a value of 2684223488 bytes. Reference: http://msdn.microsoft.com/en-us/library/bb147385(VS.85).aspx ---------- Post added at 01:48 AM ---------- Previous post was at 01:24 AM ---------- Forgot to say, you need windows 7 or tools form Process Explorer from Sysinternals to monitor how much virtual adress space you are using. Could be perfmon under XP and vista also could do the jobb, not sure (can't check because i'm conneced with my 32 bit win 7 laptop) Edit: As stated in the link, the devs can specify large-address-aware when building 32-bit applications, by using the linker flag /LARGEADDRESSAWARE, even if the application is not intended for a 64-bit platform, because of the advantages that are gained at no cost. So they of us that do spend money to buy more memory, going for windows 7 64 bit can utilize some of our computer resoureces.

Another good mission Shark Played it yesterday for the first time, Client and server 1.03 Mission refuse to load because there is bad link to statick object, open mission in editor and save. We did (or NoBrainer did) and the mission worket. When we was done with port or the cranes on the port, we headed toward the observation point. When we engaged enemy at the obervation point, we recived message about that we where spottet, and enemy reinforcment would arrive at the port. It did not scare the hell out of us, since we was done there. Don't know if this was the right sequence with wrong text? One more thing, and this is not 100% verified. I belive I opted for normal play = 10 lives. One of the member of the team do think he was out of the game after 5 deaths. And he spectated in the spectate camera script to enemy AI. Hands up - keep working - this seems to be a very good coop mission.

The Bastige Island for OFP http://forums.bistudio.com/showthread.php?t=36171&highlight=bastige&page=1 was to my knowing one of the first mission that featured some of your points Cadmium77. In my opinion a masterpiece written by bn880, where the OFP engine was pushed beyond what until then had been seen as the limit. Our clan played this mission on several ocations. Newer seen similar missions in Arma, but they may exist without my knowing. One ting I'm 80 - 90% sure of, this kind of mission was or are unplayable on public servers. You need strong teamwork, and persistent players to pull this kind of mission off. bn880 is still around so who knows maybe he or others will suprise us once more. :D

In game scripted testing where AI play some part are by nature not 100% predictable, and performance score will vary within some boundaries from run to run. I accept that. Normally will run nr 2 and 3 of the same test procedure yield higher score than the first run, because texture and other static data are cached various places. Switching from full screen display to cinema view in an unpredictable way should not happen in a performance test mission. There are two possible answers: a) The mission contain some faulty code, witch may be of importance or not for the score. B) The mission indicate a bug in the ArmaII engine og render engine. IF these are true and if this affects the score, a workaround should be applied, if that is doable.

Hie I'm not home atm, but yesterday when I tested with vsync on/off triple buffering on/off etc i noticed some strange things. In one of maybe 15 runs, there are a dilaog in test 1 and if I rember correct it was somethin like person 1) Have a nice day person 2) Yes, sir. Strange that this dialog only occured once. In 30% of my test run with a 19" lcd 1280 * 1024 4:3 monitor, and all ingame setting the same, the test did show as ingame cinema (with black border up and down). Never seen this inside the game, when I'm playing. Not sure wat I did to provoke that, but I wil test later.

I agree with SgtH3nry3, newer buy HW before you need it. The price/performace score moving up in your favour, one tick or tack for each day. In 2019 you could have 500GHz in each of your core. http://en.wikipedia.org/wiki/Graphene And if BIS still are making Arma, they will launch Arma VI 29. may 2019 :D

I presume that Suma and Kegetys did correct me, and that they both tell the true tale, then I find it hard to belive that 64bit Cryses run better than 32bit Cryses on the same HW. Google search do not inidicate that 64bit Cryses is better or faster than 32bit Cryses. http://www.techarp.com/showarticle.aspx?artno=468

I rest my case, I did not know that the x86 arcitecthure/enviorment was uable to benefit from the wider 64-registers. Maybe thats good for Sun, IBM and Cray, or have AMD/Intel made a solution that in the longterm is the most valid approach.

This is not true. You can compute with 64b real numbers ("doubles") in 32b without any issues. As I said ..Without .. workarounds. [code #ref)64-bit add: 64bit-bit system: fetch A Fetch B add64 A,B 32-bit system: Fetch A1 //32 LSB Fetch A2 //32 MSB Fetch B1/ /32 LSB Fetch B2//32 MSB Add32 A1,B1 //add the "first half" Add32_r A2/B2 //add the "top half" _r[#ref] 64bit math could also be done in a 16bit computer/os, and then the compiler need to use more clever workaround. What is not true about that? The fact that this is hidden for the develope dont change anything. I did continue to say that I do not know about many applications that need the true 64bit match computing precision. And if ARMA 1 or II or what ever dont use 64bit math calculations in any significant way, yeah sure 64bit ARMA(n) is a waste. I also remebmer the transision from 8 to 16 and from 16 to 32bit. And the bigger memory footprint was always a big issue, until you got enough memory, better hw, and then nobody did care any more. Same will happen now for sure. Do we need 128bit computing? Sure but not in my liftetime maybe