Miuramir
-
Posts
282 -
Joined
-
My first serious MMORPG was FF-XI. It initially came out for the PS2 just over eight years ago (May 16, 2002); I started shortly after the US PS2 release in March of 2004.
FF-XI features a detailed and visible weather system which ran smoothly on the PS2, although originally it required an above-average PC to run them (in ~2005 many PC users were still turning off visible weather effects for raids, etc.); by the time I had a better PC and switched from PS2 to PC (spring 2008) it was a complete non-issue.
Windy days featured small leaves being blown along, rain and snow were visible falling, stormy days had lightning flashes, some areas had heat shimmer on hot days, and so on. The weather interacted with a few quests (could only find a particular tiny object during a sunny day (as you supposedly found it by the sun glint), an elemental stone needing to be exposed to several different sorts of weather to activate it, etc.), affected fishing results, and had a significant effect on spells (IIRC +10% damage for aligned to weather (wind spell during windy weather), -10% damage for counter-aligned to weather (earth spell during windy weather)).
All this was done with a system that imposed very minimal load on the server or the network. The weather for an outdoor zone was randomly determined from a table appropriate to that zone's climate several periods in advance (to allow weather forecasting to work); this was pretty much a single random roll on a table for each zone every period. Sending the weather for the in-game day to the clients may have been as little as a byte every period, as part of the update of various world statuses. I no longer recall exactly what the weather change period was, IIRC it was a game day (slightly less than a real hour); and storing several periods locally so that, e.g. certain NPCs could offer opinions on tomorrow's weather presumably took only a handful of bytes.
On the local side, an outside area was either marked as being "under cover" or was considered "open". You could stand a bit back from the mouth of a cave in the mountains (under cover), and look out into the snowstorm (open). A good fraction of this is already done in CoH; as part of the Rikti invasion code, the game knows whether you're "under cover" or not; if you're under a bridge you don't get Rikti spawns, so this aspect would be easily portable.
The trick with all of this was that there weren't any directly player affectable persistent effects... the snow didn't build up into snowdrifts that you could dig into, or the like. There are already a considerable number of similar effects in CoH; pretty much everything handled by the particle system (leaves, shell casings, etc.) qualifies for instance. There are in fact more annoying current issues, like Propel objects and (IIRC) alakazamreact objects being different for each user (would it really be *that* much of a load to pass a single byte to designate the object, rather than each client generating them randomly locally?)
Looking at it another way, we already have three states of "weather": Normal, Rikti Invastion, and Zombie Invasion. There's clearly a mechanism to signal from the server to the clients to change into and out of the invasion weather states; hypothetically that could be repurposed to add "rainy" as another state, and send that to the zone every so often instead of "normal". -
Quote:I was looking at doing exactly the above, and after research decided it wasn't going to be a significant improvement. Consider:Originally Posted by VoxPanda
and speaking of sad panda's...
I went from a gtx275 to a gtx480: with no real change in FPS.
For ~$500, immature drivers, and more heat and noise, the above did not seem to be a good deal, and unfortunately you've helped confirm that. What resolution are you running at? My first-level reaction is that the increased pixel pushing power and slightly increased overall processing power should help folks with high-res monitors in general, but that for Ultra Mode the reduced texture performance would eat into that.Code:| 275 | 480 | % change -----------|------|------|--------- GPixel/sec | 17.7 | 33.6 | +90% GTexel/sec | 50.6 | 42.0 | -17% GFLOP/sec | 1011 | 1345 | +33%
Fermi, as we have it now, is a pale reflection of what it was supposed to be; the best available configuration in reality is the crippled-clock, disabled-shaders version that should have been the discount bottom if things had worked as designed. It remains to be seen whether that can eventually be fixed or not.
Note that going ATI isn't an easy fix, either; to get one that is at least equivalent on all the above stats requires going with at least a 5850 (~$300 - $390 just to about stay the same), and it looks like you'd need to go up to a 5870 (~$400 - $520) to see noticeable improvement across the board.
Considering that I paid $220 for my GTX 275, the price/performance of the current crop of cards is terrible. -
Quote:Given that this is a forum for other players to try and provide assistance, it would be appreciated if you didn't shout at the people who are volunteering their time to try and help you and troubleshoot the game.Originally Posted by venombeyond82
It is not my Video Drivers, It is not MY Computer, It is not Ventrillo, is it only happening AFTER Issue 17 went live, and onto the main servers. I am not sure HOW MUCH clearer I can be!
I have tried even turning down my Video settings, turning off ALL Ultra Mode Settings, Heck , even turning it down lower than what I originally had it prier to issue 17. It is STILL happening.
My video Card, is an ATI HD 4870 1 Gig GDDR5 Memory. My Processor is a 3.2 Ghz Athlon. I have 4 Gigs of Memory, and I use 3.25 of it. By the time I am in Atlas park for more than 5 minutes, my memory usage goes up from 504 Megs, to 1.58 Megs. CoH should not be using nearly that much.
You're using old drivers, which may or may not be related. The 10.4 drivers are currently recommended for City of Heroes, and I'm not aware of any reports of problems with them that aren't still problems with older drivers as well.
It's still not clear why you think there's a problem; in a 32-bit OS, it's allowed for a single program to use up to just under 2 GB of memory to improve performance. Using up to a gig wasn't terribly unusual in previous issues, so a gig and a half under I17 doesn't seem particularly problematic. Plenty of programs use more memory than that under routine circumstances. Heck, I play an ASCII game that slams right up against the 2 GB limit (admittedly, Dwarf Fortress is a bit more complex than, say, Nethack
On the system I can check most conveniently, CoH uses about 638 meg real, 743 peak working set just sitting at the login screen, but that seems to be a stable figure. By experimenting with the character select screen, real memory usage goes up by about 3 MB for every character it loads. Letting it sit for tens of minutes (edit: now nearly an hour), it remains stable as long as I don't bring up another new character to render; but it doesn't seem to be dropping back either, as the characters are still cached (clicking on one I've already clicked on is noticeably faster than clicking on one I've not yet clicked on).
This seems to indicate it's not a memory leak in any of the normal senses of the word, but that the program is aggressively caching ultra-mode costumes to improve performance. Even if some of this gets forced out into virtual memory, it may well be faster to pull the costume info out of the virtual store rather than re-assemble it out of the component pieces from the piggs.
The catch here is that this is happening in system memory, not graphics memory, so commands like /unloadgfx don't apply here. It would be handy to have a new command that tells CoH to manually flush the main memory costume cache; but as long as it starts recycling / cycling them out in a sane fashion before it hits the 2 GB limit, there shouldn't be any serious problems.
Testing procedure you can try if you still think your system in particular is having problems (not out of the question, some other program could be interfering with CoH):
* Reboot the system as you would normally
* Start the Task Manager; tab Processes; set Refresh to Low; and make sure at least the following columns are visible: Under XP: Image Name, Mem Usage, Peak Mem Usage, VM Size; Under W7: Image Name, Memory (Private Working Set), Working Set (Memory), Peak Working Set (Memory). (You may need to use View / Select Columns to see some of these.)
* Minimize the Task Manager, and start CoH. Once it's gotten to the login screen, give it a moment to stabilize, ALT-TAB to the Task Manager, and record the above listed columns. Give it about 5 minutes, record again, and another 5 minutes, record again. On a stable system, this should not be increasing any significant amount.
* Log in, and pick a server that has several characters (hopefully at least 5). Give it a moment to finish loading your top / default character, ALT-TAB to the Task Manager, and record the numbers. It's normal for these to be a bit higher now. Give it 5 minutes, record again, and another 5, record again. These numbers shouldn't be increasing significantly.
* Click on another character in the selection screen, let it finish rendering, and record memory. In my case, it looked like the real memory went up by between 2 and 4 MB; this may well vary with character costume complexity. Wait and record to make sure it doesn't continue to increase.
* Repeat the above steps with another character, and yet another. Based on my results, it looks like there will probably be either a flat, or varying around a central value, increase each time you tell it to render a new character on the loading screen; but once you've done so, there are no continuing increases or leaks.
* Go back to one of the earlier characters, and record again; based on my results, it looks like memory should NOT go up again, as it's already loaded and cached it. (It should probably also load faster, although on a fast system it may not be easy to tell.)
* Pick a character that has access to a SG / VG base that is not heavily used (preferably a fairly simple one without thousands of decor objects, etc.) and ideally has at least one costume that doesn't have animated bits, such as a moving tail, cape, aura, etc.; and hopefully at least one more complex costume. Log onto that character, switch to the plain costume if relevant, zoom in close, look down, and get them to a base portal and into the base. Wait for the base to finish loading, and go stand someplace boring staring at a chunk of wall or floor that's not animating, glowing, etc. Give things a moment to stabilize, and record memory. It's likely that it's significantly higher, depending on how much stuff you've encountered on your way here. Do the wait 5, record, wait 5, record again; it shouldn't be increasing.
* Assuming memory is stable above, try some controlled tests with more complex stuff. Switch to a costume with a cape, tail, or aura; and see if there's some sort of problem with that. Stare at the entrance pillar or an animated base object, do the memory check dance. Again, while there may be small increases when you first kick into something different, the memory usage should stabilize and not keep increasing.
* Testing at this point gets a bit less controlled, unless you've got multiple accounts and a second computer to run them, or cooperative friends. Ideally, you'd have someone else zone into your test base, and see how the memory increase compares to what you figured out earlier at the character select screen, and repeat with a few different characters and costumes. Lacking that, you could zone out of the base, record about how many people are in your field of view, and zone back in, and compare then.
* Finally, head out to someplace popular; AFK near Atlas, at the Cap black market, or some such. Every 5 minutes, ALT-TAB and record the memory. My hypothesis based on what I've seen is that as new people come into your view and/or vicinity, memory usage will grow up to some point, at which point the character cache will have filled and CoH will start recycling slots, and memory usage will stabilize. If this isn't the case, you'll have solid documentation of exactly where and how this fails to happen, which should be exactly what they need to find and fix the problem.
Obviously, if you find one of the earlier steps where the memory increases without bound and crashes your system, the problem's been localized at an earlier point. Don't stop the process as long as the memory usage remains under 2 GB, however.
Report the numbers here in detail, along with other info on your system; a CoH Helper report would be particularly useful. If you do have memory growth until crash, a Hijack This log in a following post would be useful (don't worry about it unless you have serious problems). -
Quote:In all fairness, it comes up with the 6th anniversary message again, and you do have to look more carefully to notice there's another tab. I wasn't even aware that the MoTD could *have* tabs until today. It would have been much better to put the warning message on the front tab, and the repeat of the anniversary text on the second tab.Originally Posted by DumpleBerry
Ranks up there with all the people QQing in WW and the BM after closing the MOTD without reading the Known Issues popup. I had several people tell me that they "never got it." Everyone gets that message, that's why it drives me nuts sometimes.
Additionally, the fact that it's listed only in an obscure time zone (BST? Bangladesh Standard Time is the primary meaning, but I'm guessing it's more likely to be British Summer Time; there's several others it could be including Burma Standard Time, Bissau Standard Time, and Burkina Faso Standard Time) isn't as helpful as it could have been with only a few minutes more work on their part.
If you're going to list only one time reference, it should be UTC, the world standard since around 1972 (replacing the old GMT, and almost the same unless you're worried about seconds per year). Obviously listing UTC and the server's timezone would be better yet.
Side note.. Is there a way to bring the main MoTD back up? I'd have thought /motd would do it, but it's not a command. -
Quote:Can you be somewhat more specific? There are known problems with several voice chat programs that have nothing to do with memory leaks, but with keyboard-trapping programs (ie, the "push to talk" features).Originally Posted by venombeyond82
Yesterday, Issue 17 came out, and I was quite happy. While my video card isn't top of the line, I can run Ultramode Low to Medium range, and everything runs smoothly with out any FPS issues. I'm on ATI 10.1 Drivers as a few of my other friends who have ATI are as well. The problem isn't my video, its the memory.
Before Yesterday, and before I17, While playing CoH and talking on Ventrillo, I never experienced any major Memory Leaks. But now it seems to be happening quite a lot. It has caused sound to skip on me, both in game and in Ventrillo talking to my friends. It has caused major lag issues left and right, and none of this happened before yesterday. My roommate who also plays City of Heroes, and who has a PC 98% close to what mines is, is having the exact same problem.
I've spoken early to 20-30 people I know on my global friends list, or in Pocket D or in Atlas, and I've had a lot of people telling me they have the same problem, that now, CoH is sucking up far more memory than it ever has, causing horrible memory leaks, and causing these issues.
I hope something can be done about this soon, as if it is not fixed in the future, it will keep me from playing properly, and I may have to leave.
A memory leak implies that the program takes more memory as it is run in an uncontrollable fashion, and does not clean it out. If you suspect this is happening, record the memory usage shortly after logging into a character, and then have them do nothing for a while (stare at a wall someplace to avoid taking up additional space for passing characters, etc.), and record the memory increase with time.
You can see if things go back down after using /unloadgfx, in which case it may be related to graphics driver problems. Make sure you're using the current one for your card, and let us know what it is. -
Quote:This sounds like either a power throttling or heat throttling problem. In general, most modern, and some older, CPUs and GPUs will automatically drop down to a much lower speed if they think they're overheating; and some will do so if they think they're not getting enough power. Frequently this is a real problem, but it can also be caused by various sorts of false positives.Originally Posted by xofour
I am experiencing dramatic cycling fps rate dropping. I can play at a normal 30-60 fps rate for 5-8 minutes. Then it will drop to 3 or less for roughly 5 minutes. It will then snap back into full fledged playable mode for another 5-8 minutes...rinse repeat. Even in safe mode It follows the same cycle, but with higher numbers. It doesn't seem to matter whether I am in a mission, or on the map, nor does it matter whether I am in paragon city, or rogue islands. I used showfps 1 & my watch to determine the timing.
Try the following: Shortly after starting the game and getting onto a character, check the time and run the following series of commands from the chat window:
/maxfps 20
/maxinactivefps 2
/showfps 1
Then play normally (although somewhat lower frame rate), and see if it lets you play longer before it it hits the slowdown. If you don't see a significant difference, try a more severe step, like /maxfps 10 or 12.
If the above makes a difference, there's a high chance that the problem is cooling or power. Make sure you've carefully cleaned all the various fans and intakes; graphics cards with plastic covers are notorious for accumulating lint and gunk under the cover and blocking effective airflow, for instance. -
Quote:This may be related to problems with key / mouse mapping programs that have been causing performance and/or lockup problems. Do you have an old, no-drivers-required corded mouse that you can try with the Revolution's dongle not plugged in at all?Originally Posted by AquaJAWS
I've seen lots of stuff about keyboards not responding due to ATI graphics cards, but nothing about this yet.
Intel Core 2 Quad CPU, 4 GB RAM, nVidia GeForce GTS 250 graphics card, wireless Logitech MX Revolution mouse.
I didn't notice this on test, however I didn't play too heavily before the live release. Once the game starts, my mouse's responsiveness becomes very sparatic. It'll respond for a few seconds, then stop for a few at random. It seems to be the worst on the login and character creation screen, but still happens when playing. Keyboard response is fine for me (Logitech Corded Wave Keyboard) but the mouse issue is very frustrating. I stop the game, the mouse works perfectly with everything else in the computer. If the game is on, it I gets weird even on other open programs if I alt-tab to a web browser or MIDs. I seem to have noticed a little better behavior without the new shadow tech on, but not by much.
What version of the Logitech drivers are you using? You could try uninstalling the Logitech drivers, seeing if it changes anything, and then try installing the most recent ones. -
Quote:There are several known problems with programs that lurk in the background and monitor keystrokes causing either performance problems or lockup. Unfortunately there are all sorts of programs that do this, frequently without telling the user.Originally Posted by Organica
Oh, I did that.
Current theory is that people are reporting problems with sound or with other programs running at the same time. I have music turned off from the older machine but not the one I'm having problems with... and I may have had other programs running each time I tried running it, so gonna try once more.
Try disabling everything you can think of that adds a keystroke to do anything that's not in standard Windows. Known suspects include an ATI hotkey utility, several sorts of voice chat programs, various mouse / keyboard macro programs, etc. -
Quote:There is some sort of bug with the throttling system that seems to be locking people out of the market interface. Usually exiting CoH completely and restarting clears it up for most people, but apparently for some it doesn't. If restarting fixes the problem, /bug it. If restarting doesn't fix the problem, /bug it and then /petition it. It can't hurt to post here with as much details as you can... what sort of things did you have bid on, stored, won but not collected, etc. prior to I17, and what sequence of market activities you took after I17.Originally Posted by Echonis
I also have the same issue with the grayed out "Get". Can't pull any of my purchases off the market. The "Post" also doesn't work to sell, but I too have been able to sell by entering the price and pressing the enter key.
It is possible that this is related to having some combination of items already on the market when I17 hit; there were a few reports of problems in beta, but there are very few people who had lots of standing bids on the beta servers, and no way to go back and test upgrade problems since the beta servers were already in I17 mode. -
Quote:I don't want to sound harsh, but that's a really old card. The R200 technology was introduced in late 2001, and while the 9200 was technically the second redesign and came out in mid 2003, both of those redesigns were primarily about reducing cost, and actually reduced capability in several areas.Originally Posted by LordSunburst
um....
Radeon 9200 series
ati2dvag.dll
is this what you are asking about?
Your card hasn't been supported by ATI for more than three years it looks like; the most recent Catalyst drivers that apply appear to be 6.11, from late 2006. If you're running an even earlier version, try upgrading to those and starting the game in Safe Mode; if you're already on those and Safe Mode doesn't even start, you'll need to file a support request but I'd not be optimistic.
We'll try and help, but in round numbers, the minimum card Positron recommended five months ago to run Ultra Mode is more than thirty times faster than your card. (Radeon 9200, about 1,000 MT / sec; GeForce 9800 GT, about 33,600 MT / sec; and in several ways more powerful than the numbers reflect.)
If you do need a new card, you are likely limited by both an old bus and an old power supply.
The 9200 came in AGP x8 and PCI versions, try and figure out which you have. AGP comes in several versions, and you need to figure out what sort / version your motherboard actually supports. PCI is in general worse from a technical standpoint, but somewhat more standard, and there are some more recent cards made for it.
You will probably also need to figure out what the output of your power supply is; this is supposed to be on a sticker somewhere on it. Newer power supplies should have the (more detailed and important) +12v rating in Amps, but older ones may have just an overall rating in Watts. -
Quote:IIRC, Reichsman had a special exception artificially low regen rate for his HP; since regen is normally a percentage of HP and he's got way more than normal. Would the numbers work out if somehow the special case for that got lost somehow in all the I17 changes?Originally Posted by UberGuy
His HP were trending down, but I can't even think about how long it would have taken to defeat him, assuming it was even actually possible. I may just not have caught him getting a HP increase from his regen rate, as his base tick rate is low compared to most stuff. Edit: I was basing that of his actual HP numbers. We never did enough damage to visually shrink his HP bar that I noticed.
-
This got linked over on the LJ community; it's not terribly clear if anyone can vote, or only artists. Assuming it's the former...
1: Frost (materials, style, and subject all a great fit to the theme; plus funny)
2: Shia (materials and some style elements highly thematic, yet the figure comes off quite well; plus some cute, some sexy)
3: U-Naught (nice mix of thematic materials and style with more sophisticated techniques; plus highly thematic setting / story. While the official text says "box of milk", my take on the expression would be closer to "How much high-fructose corn syrup do they *put* in this stuff?")
Honorable mention: ChristopherRobin (thematic materials, nice shot somewhat marred by framing / setting (should have either centered it better, or worked the sill into the piece more); I'm not really giving it credit for bonus sexy because with the medium it doesn't leap out, although looking at it closely it might qualify if in a higher-res format) -
Victory, redside, 7p - 10p Eastern.
The news post here on the forums told people to gather in Atlas, Talos, or Cap; the red admin message at 7p told people the event was in Atlas or Mercy, which caused some confusion / annoyance.
Other than the one initial message, there was no further communication. If someone had been a few moments late logging in, or switching characters, and logged into Cap as instructed to find nothing happening, they may well have concluded that the event was canceled or broken. (Eventually a somewhat forlorn Eochai was found in Cap, wandering in the water vaguely near the black market; we took it out while assembling a 3rd team nearly two hours later, but nothing further happened.)
The main (only?) action was in Mercy. Initially, the zone was tagged for a standard zombie invasion (red sky, automated spawns on people / groups outside, etc.) There was a decent-sized group near the upper arbiter, and we slogged through zombies for a bit despite there being not much point to it.
Eventually someone shouted in broadcast that they'd seen Eochai over near the fateweaver, and we got folks moving in that direction. Nothing much seemed to be there, so folks milled about shooting zombies some more until someone discovered that the spawns were over on the main street leading from the starting Arachnos base, about 500 yards SW of Kalinda. People gathered there, and as far as we could tell that's where the action was for the rest of the night. At least it was reasonably convenient to the "hospital" in the starting area.
At some point soon after showing up, there was someone handing out yellow anniversary temp titles. It wasn't clear if this was repeated, but if so only intermittently.
We still had zombies running for the first part of this, but eventually they cleared and didn't reappear. The "special" spawns appeared to be in an approximate pattern; I'm not sure I've got the sequence right but the following should give you the general idea of the combinations.
* Eochai (Halloween event version; Hallow Spirit badge, merits) and Jack in Irons (Halloween event version; Iron Warrior badge, merits)
* Winter Lord (Winter event version; Frozen Fury badge, merits)
* Caleb (HellBane badge; merits)
* Deathsurge (sometimes along with another one; Surging badge, merits)
* Quarry and Stropharia (no badges or merits for either)
* Lattice and Thorn (no badges or merits for either)
* Some combination of the Freedom Phalanx, usually two at a time but we did get the full crew at least once. Only appearing were the 6 that comprise the Blackguard badge, from Recluse's Victory: Manticore (Deadeye badge), Back Alley Brawler (Uppercut badge), Positron (Destroyer badge), Sister Psyche (Interrogator badge), Synapse (Breakneck badge), Statesman (Oppressor badge). Plus the Blackguard badge for eventually getting the full set of 6.
The usual "team must do 10% to get credit" rules were in effect, so unless you got unusually lucky with teaming, you'd generally need to go through several iterations to get the badges; my odds seem to be somewhere between 1 in 2 and 1 in 3 for getting the badge after a defeat.
I had the available badges earned on my first toon by about 8p, call it 45 or so minutes of fighting after the initial confusion and zombies. Switched, and my second went better (possibly because she was a DP/Dark and control-clicked Twilight Grasp), getting things in about 30 minutes or so. Shortly thereafter, as I was about to switch toons, the above list stopped being spawned, and we got a peculiar Rikti non-invasion.
IIRC the sky didn't change color, and we were getting the ship raid version of things (level 53 and 54, did NOT count for the invasion badges). It went OK for a bit, and then spawning in at least a half-dozen each level 54 Heavy Assault Suits and Rikti Priests, along with a bunch of Magi and lesser stuff, caused what was basically our first zone wipe. People were hospitaling, coming back, and getting blasted again. Eventually folks regrouped enough and kept slogging; the problem was that there were no real rewards for this; no badges, no badge credit, no merits, and little exp. Once it became clear that we couldn't get credit for anything, I took a bit of a break and switched characters again.
We ended up gradually forming a team in Cap (of people who had parked there because that's what the instructions said) to take down the solitary Eochai, and then headed to Mercy. By now a bit after 9p, and the Rikti were gone and it was back to the above rotation. The third team was mostly lower level, and although we had a 50 leading, we were having problems getting enough damage through to get credit; I only got about half the badges in the remaining not quite an hour.
Overall feedback:
* The lack of communication was annoying. When I switched characters, I'd run into people that had no idea anything was happening, or where they should go. Having a red admin message every 15 minutes or so, at least every 30, with instructions for which zone(s) and approximately where people should gather, would have been a really good idea. Telling people to meet in a zone that wasn't being used was a really *bad* idea. While the spot picked wasn't bad in terms of being generally near the starting base "hospital" yet out of sight of the base to avoid disrupting new characters, etc. picking a spot that had no distinguishing landmarks to gather around or tell people to head toward wasn't the best plan.
* Despite being billed as an interactive event, there didn't seem to be much interactivity going on. The vast majority of the time, it might as well have been scripted on a timer; if there really was someone live there, they may as well have been clicking on their "spawn next monster clump" macro every few minutes while they caught up on their reading or whatever. To the best of my recollection, none of the foes did anything to indicate they weren't under routine AI control; no speech, no dancing, no responding to the few players trying to interact with them (calling out specific members of the Freedom Phalanx, etc.). Additionally, there didn't seem to usually be any sense of theme, malice, or whimsy in the spawns and placement, just rote cycling through the predetermined monster list. The oversized high-power Rikti spawn *may* have been in response to the people asking for a Flier or Dropship, or just random chance.
Last year, it was far more interesting (although somewhat harder to find); moments like taking on three Ghost Widows and a Kronos Titan spawned in on top of the Atlas Globe I'll be remembering for as long as I play the game, but this year will probably fade to "I got some alts a few obscure badges in exchange for some debt". -
Quote:So, you're basically working up some sort of suggestion proposal for "common dual IOs" that could be purchased from a fixed-price source (NPC, table, etc.) and/or dropped like more common IO recipes? Sounds interesting, keep us informed. (Chances are I'd be in favor of it; the set bonuses would still leave the existing set recipes more powerful and valuable, but it would greatly improve the mid-level situation where there's a shortage of available recipes to frankenslot with, and help provide a smoother path / power curve into the world of advanced IO slotting.)Originally Posted by Samuel_Tow
The original idea is sort of a legacy of something I and a few others have been throwing around for some time - the divorce of the dual-aspect enhancement concept from that of "rare loot." In other words, I'm looking at dual-aspect enhancements irrespective of the sets they are in and what else those sets supply.
As you seemed to have realized after looking at Krogoth's example, the key is that multi-aspect enhancements are best at enhancing multiple aspects. Go figure Less tautologically, they completely fail with powers with only one enhanceable aspect; on powers with two or three aspects, they are primarily not for improving power levels, but reducing slot numbers required (which allows you to use those slots to make other powers more effective); on powers with four aspects (very common: acc, dam, end, rech), they provide a significant boost; on powers with five or more aspects (e.g. attack with a debuff component) they are the only practical way to fully enhance.
-
Quote:You haven't given much info here, but some common things to check:Originally Posted by Gypsy_of_Paradox
I'm trying to run a dual monitor on my laptop. I've got an HDMI cord plugged into both my laptop and LCD monitor. I followed the setup and the LCD is still blank, no idea what I missed.
My laptop is an ASUS G50VT-X1 and the monitor is an Emerson 19" LCD HDTV. What am I missing?
* Does your laptop actually support dual screens? Some don't have the capability to do so; you can use the external or the internal, but not both. (Others may support it, but have drastically lower performance; I've used one that could easily run CoH on the internal 1400x900 LCD, but if you connected an external 1920x1080 HDTV the performance would tank to unplayable levels).
* Do you have reasonably current drivers for your video card? The default drivers that come with many OSs may do OK for simple use, but not support dual monitors, etc.
* Have you enabled your OS's dual-screen output? Frequently on a laptop this would be toggled by Fn + a function key, sometimes labled "LCD/CRT" or having a picture of a monitor. Usually this will toggle through three modes by default; built-in screen only, presentation mode with the external screen cloned from the internal, external screen only. Setting to have internal and external as separate displays may require advanced settings.
* Have you configured your driver and OS to have the sort of dual-screen arrangement you want? In some versions of Windows for instance, there are multiple ways to set them up.
* Have you verified that the external monitor is on and set to HDMI input? Some don't auto-detect well, you may need to change it from "Auto" to "HDMI" or some such.
* Have you checked in the laptop's BIOS to make sure HDMI output is enabled?
* Is the laptop plugged into working wall power? Some laptops don't talk to some of their external ports if on battery power.
* Once everything is set and connected, have you rebooted with it all connected and turned on? Some laptops check for an external connection on boot, and if not present disable the port to save power. -
Quote:While I've not dealt with this exact problem, you have reasonable odds of doing a "repair install" and not loosing your data. Try to get hold of an XP disk with as close to the service pack that you had running as you can. (For instance, if you originally installed with SP1, and have over the years upgraded to SP3, trying to Repair Install with your original XP-SP1 disk will be a hassle; borrow a XP-SP3 disk if you can.)Originally Posted by Speculator
I'm hoping someone here can help me. I was trying to update my video drivers and I used Driver Sweeper to clean out my old drivers. But I accidentally deleted my chipset drivers too. Now my computer won't boot, it goes straight to a blue screen. Is there a way of restoring my drivers without re-installing XP and losing all my data?
You could also install your hard disk as a secondary hard disk in another (working) computer, or install it into an external USB enclosure that you then attached to another (working) computer, to back up your data. This is probably a good idea for paranoia's sake even if you think the Repair Install will work. -
Quote:If I were in the market for an Ultra Mode card (which I'm not, I'm happy with my GTX 275 now and all signs point to it doing at least OK for Ultra in the near term), I would wait until whichever release that has Ultra Mode (probably I17) is either announced to be going into Open Beta, or you get a Closed Beta invite; and purchase from a company with a reasonable return policy. That way, if you're not satisfied with it for whatever reason in actual Ultra Mode testing, you'll hopefully still be within the return period.Originally Posted by Samuel_Tow
Quick question - if my reason is Ultra Mode, wouldn't it actually make the most sense to buy now or soon, given that I17 and Ultra Mode seem to be looking at an April release date?
There's simply too much variance in systems and users, and too little public information about Ultra Mode, to give more than suggestions. If it were my money, I'd want to be sure that what I ended up with will give the sort of experience I want, through actual testing. -
Quote:That's weird at best. The only thing I can come up with would be a guess that it isn't the knockback itself, but something connecting to calling the ragdoll physics routines heavily. Do you have one of the rare old CoH-supported PhysX hardware cards in your system, one of the newer Nvidia PhysX cards (not CoH supported, but might be confusing Windows), or running the physics in software like most users?Originally Posted by PumBumbler
...
Normally most people won't even see the problem, but if you have lots of knockback it may happen if you have a slowish storage subsystem and you are running Microsoft Security Essentials. Note that I had no lag otherwise until I used knockback powers on 3 mobs and up. After about 10 mobs and up the game would be essentially rendered unplayable (.1 fps or less), and the network connection would stall as well since the system was freezing up entirely. -
Quote:A short-term fix is to use the "/maxfps #" command, with # set to something that will not over-tax your card. For instance, there is absolutely no reason to have maxfps set higher than the refresh rate of your monitor (usually 60 if you've got a LCD). There was an issue recently with CO and STO where if you didn't set a max fps, in some situations with higher-end cards it would go into a death cycle of computing far faster than the screen could even update, overheating the card badly; this problem seems to have similar roots combined with some fan control issues.Originally Posted by Hyperstrike
Apparently something in the 196.75 version of nVidia's drivers is causing temperature controls on nVidia cards to fail, exposing the cards to a danger of overheating.
Slashdot Article
Parent Article
nVidia Forum Thread #1
nVidia Forum Thread #2
If you're running this revision of drivers, it's recommended that you uninstall and roll back to the previous version of the drivers.
maxfps is most effective if it's an integer divisor of your refresh rate; for a typical 60 Hz refresh LCD, the useful settings are 60, 30, 20, 15, 12, 10, 6, 5, 4, 3, 2, or 1. Unless you have a high-end rig, the visual differences setting maxfps to 30 are quite small, yet can have considerable cooling benefits.
(As a side note, there is also /maxinactivefps #, which controls the rendering rate if CoH is in the background; if you think you might need to let your card cool off, you can set that to a very low number (like 3 or even 1) and minimize CoH. This is also useful to reduce the load from the other client(s) while running more than one account.) -
Two basic problems leap out: first off, the Nvidia GeForce GT 220 is a weak card, not really intended for serious gaming these days (for instance, it's between half and a third as powerful as various versions of the 8800). It's a bit odd in what seems to be an otherwise pretty nice PC. Secondly, you're running antique drivers on it, which isn't helping.
Download and install the current correct drivers for your OS and card as a first step. The following link should be correct currently:
http://www.nvidia.com/object/win7_wi...6.21_whql.html
There's a chance this will clean up your actual bugs, and you'll just be left with the underperformace, which we may or may not be able to do much about. There's also a chance you'll run into problems, as the 64-bit drivers aren't as reliable as the 32-bit ones; supposedly the 196 series is better about that, however. In either case, follow up here. -
Quote:Technically OT for this thread, but FYI these turn out to not have been either random or Microsoft's fault. PCs that were bluescreening after the patch (better known as MS10-015) turn out to have had hacked or at least damaged system files from one rootkit or another, frequently one of the Tidserv variants:Originally Posted by Coolio
Reason I asked about the OP's OS is M$ have nuked a few PC's with their kb977165 update.
By nuked I mean random Blue Screens (BSODs) and similar.
http://www.symantec.com/connect/blog...v-and-ms10-015
http://www.prevx.com/blog/143/BSOD-a...apologize.html
Basically, the various rootkits had rewritten certain calls to point to RVAs (Relative Virtual Addresses), pre-calculated addresses specific to the malware. After the patch changed the code, the RVA calls then pointed to different and likely nonsensical locations, causing a BSOD. (Similar in concept to the problem where really old scanner missions can end up with weirdly wrong enemies after a patch that changes the files they point to.) -
Quote:This continues to be both false and misleading. Nvidia 8 series and more recent, including all of the above cards, support OpenGL 3.2 (and OpenCL 1.0). OpenGL 3.1 support was available in the 190 series drivers back in August 2009, and the 191 series drivers from October 2009 upgraded support to full OpenGL 3.2. In comparison, as best I could tell ATI didn't offer OpenGL 3.2 support until the Catalyst 9.12 drivers from December 2009.Originally Posted by je_saist
The Nvidia GTX 2xx cards, the GTX 260, GTX 275, GTX 280, GTX 285, and the dual chip GTX 295 all have been end-of-lifed. No more chips are being made. The starting point for these cards, if you can find them, is around $200.
More importantly, these cards do not support DirectX 11 or OpenGL 3.2, and they cost more for the same performance than the old RadeonHD 4x00 cards that do not support DX11 or OpenGL 3.2.
It is true that the 200 series cards you list only support DX10.1, not DX11. Since City of Heroes doesn't use DirectX, this is irrelevant to the direct discussion, but many people play other games. What this actually means even indirectly is pretty debatable; the majority of games don't require anything better than DX9, because that basically means writing off all the potential buyers running Windows XP. There's less than a half-dozen games out that even support DX11, and none of them seem to require better than DX9; even the list of announced games that will support (not require) DX11 is pretty short. I'd guess the absolute earliest we'd see a major AAA DX11-required game would be Christmas 2010, likely to be a FPS of some sort; but it wouldn't surprise me if it's up to a full year later than that, and more like two years before it's more than a niche. Remember, DX10 came out back in 2006, and here in 2010 very few games require it.
On another related subject; as for the "go ahead and buy now" commenters, the rumors seem to be flying around as to an ATI refresh in either April or mid-year (June / July). This may make more attractive cards available, reduce the price of existing cards, or both. This really isn't a good time to buy a graphics card unless you need one, especially if your main reason is Ultra Mode; I'd suggest saving up for at least another month or two. -
RAID 0 will eventually lead to tears. In exchange for hypothetical minor performance gains, which may or may not actually be realized in whatever application you're using (remember that simply running through a RAID system has a certain amount of overhead losses), you're increasing the complexity of your system in a way that dramatically increases failure chance, and makes recovery in case of certain failures more difficult. (You cannot, for instance, just pop the drives out of the system into an external case to get at your data in case of failure elsewhere in the system. )
Hardware RAID 1 theoretically provides enough reliability increase to outweigh the added complexity. Whether it's worthwhile for home / desktop computers is a matter of debate; a good backup plan (e.g. external drive) protects against far more sorts of failure.
I'd just go with the larger single drive, and put some of that "upgrade later" money toward a good external backup drive when you can. -
Quote:Every time they've talked about the system in detail, they've said that there would be special rewards available to those who "stayed true" to their side. At one point early on, it was somewhat implied that whatever they were working on for post-50 improvement would be limited to exemplary heroes and villains (at the time, this was supposedly a way to earn up to 10 extra enhancement slots). A lot of water has passed under the bridge since then, and I'm doubtful this is still part of the design the endgame task force is working on; but they have at least considered substantial rewards for not switching.Originally Posted by Silas
c) I wonder what/how much we'll lose if we don't sideswitch all the way. Unless there's some kind of absurd penalty for doing so they devs have yet to announce, I'll be turning all my villains into Rogues and all my heroes into Vigilantes so I can play on both sides. Moar content+teams ftw.
-
Quote:Time-based crashes are usually related to either running out of memory, or buildup of heat. Given how long it takes you, it's more likely to be memory related. You seem to have a reasonable amount of system memory, but your video card is old and has old drivers, so it's more likely to be a graphics memory problem. The crash on zoning also supports this, as your card is trying to load the graphics for the new zone.Originally Posted by Ash_Trey
I seem to crash after about 1-2 hours of gameplay. There are two ways this seems to happen. One I will be in the middle of a fight, say the end of TF and I things just freeze and I get booted to desktop. Which I start the game right back up sometimes it verifies files other times it won't.
The second way I crash is same time frame, 1-2 hours of gameplay and it wll be when I'm loading from one zone to another. Loading bar won't move and if I click I get the busy mouse icon and again I get kicked back out to desktop.
Try the following to see if it helps: After you have finished loading into a new zone, wait about 30 seconds or so and type the command:
/unloadgfx
Your screen will probably look odd for a few seconds as your card is told to dump all its cached graphical info and reload only what it needs for the current zone. Hopefully, this will allow you to play longer between crashes, possibly indefinitely as long as you keep remembering to use the command to clear things out after you zone.
ETA: You should be aware that even that if the above workaround does help, this may be a warning sign that your video card is gradually failing. If you eventually update the drivers to current ones and the problem goes away, you're probably OK; if it does not, you may want to start saving up for a new card. I suggest posting information about your computer and budget here if you decide to upgrade.