Guild Wars 2

“On a 64-bit OS, 32-bit programs with LAA are still limited to a maximum of 4GB of mem usage.”

While true, most games only use 1.5-3GB available memory. So fixing the 2GB process limit is what helps those memory hungry games. So while a 64bit executable would be optimal, a 32bit + LAA is a working compromise.

@Refraction.1892, thx i tried this once, but did forget about this! Now i already have switched from a 32 to 64bit OS, but others may try this first, since u can just “upgrade” from a 32 to 64bit OS, which means lots of reinstalls.

The image shows what typically happens if the GPU memory gets corrupted, try lowering your RAM + GPU cloak via some overclocking tools and see at what frequency this error is gone.

dacooli.6712:

AndyPandy.3471:

dacooli.6712:

GTX 680 and i5 2500 4.4 and I lag. Anet is really bad..

Hehe welcome to GW2 the worlds most CPU bound game to date! Maybe this is a conspiracy with Intel to force u to buy new CPU’s?

i5 2500 4.4 is like i7 3770 3,9 in gaming, FACT. so my cpu is rly high end.

Nah this guy i was arguing with a while back, would disagree. He was running a 700$ i7, over-clocked to 5.4GHz and guess what he claimed to have no FPS issues at all, i wonder why…..

I bet if he had a 2000$ xenon cpu running at 5 Ghz, he would also “recommend” this setup for GW2.

GW2+forum is basically heaven for any hardware “pimp”, there 4000$+ gaming rigs finally pay off. So they run around noting to everyone running a i7 under 5 GHz: “u cant run GW2 with a atari/C64”

“LAA” is the reason why changing from a 32bit OS to a 64Bit will solve those pesky out of memory errors. GW2 uses around 2GB memory on longer WvW sessions, which unfortunately is exact the maximum amount a 32bit OS will allow per process. With LAA on 64bit this is lifted, therefor no crash.

The only problem i have is: WHY the heck is the application crashing? There is a “out of memory” error return code for every memory allocation function u call. If your memory malloc fails, do a D3D texture memory flush, this normally frees up the managed system memory, which will allow to prevent this kind of error.

This is lazy beginners coding and memory handling only PC company’s can get away with. If u run out of memory on a PC and have no fallback or error handling in place, your memory management system simply sucks!

dacooli.6712:

GTX 680 and i5 2500 4.4 and I lag. Anet is really bad..

Hehe welcome to GW2 the worlds most CPU bound game to date! Maybe this is a conspiracy with Intel to force u to buy new CPU’s?

Conditioned.2467:

AndyPandy make a video!

A video of what?

Btw “The other night when I was in wvw I noticed”, so basically u can’t reproduce this situation and ofc its not possible that the restarting of GW2 could be the reason for the FPS gain? Did u get the same bad FPS after a relog/reboot or did u directly jump to the conclusion of your HPET script?

Is your “test setup” in any way reproduce-able, if so how? What happens atm if u enable disable your script on your PC running GW2, does it still affect your GW2 FPS?

I simply wonder that u put all the effort into explaining what HPET is and how u enable it, but u seem to skip the part trying to reproducible validate that it actually effects GW2 FPS?

“I’m getting 40+ FPS in LA’s Trader Forum on auto-detect, compared to 15-25 I was getting before.”

Yet a other “mythbuster, magic” 160% performance increase yay! When will people at least setup a valid test, before opening a thread about it. I understand, that everyone wants to-be a FPS hero, but this does not mean u can skip the verification and validation of a result.

“Traders Forum” u know what i get 10-90 FPS in this area, depending where i stand/view. This zone is possible the worst FPS test setup u can possible come up, same with random WvW zones.

Those “gaming optimizers” on modern gaming PC, do basically nothing. If they act as a over-clocking tool than thats the best u can hope for. If u don’t render a 3dsMax scene while starting GW2, there is nothing running that can be shut-down for more than 1% more CPU performance.

Claiming a 160% FPS increase by installing a software, is as crazy and misleading as claiming by lowering your CPU temperature from 60 to 30 degree, will make your games run better/smoother.

So plz do a decent test in a controlled environment. The best u can find is a personal story instance or a empty place, that will not change if u log on/off. Than enable/disable your “change” a couple of times and see how it affects the FPS. Than do a reboot and test again, if u still see a 160% FPS increase open a thread with the exact location, system specs and steps u did in your testing. Than we can try to reproduce and review your test method and are able to independently verify it or not.

shastawong.3486:

“Parking disabled”

3DMark ScoreP6882
CombinedTest13.8 FPS

“Parking enabled”

3DMark ScoreP6743
CombinedTest14.55 FPS

So basically within the normal random margin of error, so no change for 3DMark. We still need a real test, preferable with a video for GW2.

I still don’t understand why using a different timer will notable improve game FPS? So while changing the timer u get higher precision and the timer call itself may cost less cycles, but those timers are used like a couple of dozen times per frame only, so there should no measurable FPS gain?

Tip: Don’t skill for retaliation, it will still “override” fury. Check the mesmer bug thread.

Just following up on this, since i changed from Win7 32bit, to Win7 64Bit with 4GB Ram. Before i would get this “out of memory” crash every play session once, after about 1-3 hours of playing WvW. After moving to 64bit, i had no such crash after 2 days playing WvW.

So if u get this particular error, moving to 64Bit even with a 4GB system, will probably solve those crashes.

“And I turned hpet on and Voila I sit again at 58 fps stabily.”

Could u reproduce this effect, by disabling Hpet and gaining lower FPS again and than turning it back on again for the gain?

Can u maybe fraps a video to show the steps and the results?

I’m a bit confused, since i’m a programmer and i know what the HPET timer is used for and what it does and i see no reason at all, why it would effect gaming FPS. So either there must be a “crazy” software bug involved or your claim is simply wrong and the FPS gain is related to some other factors.

Can we get some none anecdotal or placebo tests on this? This basically works always the same.

1) XY with no technical background or clue what he was or is doing posts some “magical” FPS improvement
2) The sheep’s follow the steps, also no idea how or why this should work at all.
3) Some post “drastic” FPS improvements and post “thx” other do get nothing.
4) If u review the change technically, there should be no increase at all.
5) Some “tech guy” makes a real FPS comparison on a stable test scene before and after and finds no change at all, matching the technically expected behavior.

So did any of the “wow i got 50%+ more FPS” posters here have a reproducible test? Where did u test FPS before the change, where after, could u reproduce the gain/loose by toggle the “change” on/off.

Can u Fraps/record a video of the steps and the change, since we all know a video tells more than 1000 words.

michaeljhuman.3940:

Every game benchtest I have seen shows that if you lower your game graphics settings enough (lower resolution especially) that eventually you become CPU bound.

It’s not that the CPU load goes up, it’s that the GPU stops being the bottleneck and the CPU does.

This is true for very low resolutions, benchmarks actually “need” this to test CPU performance for none synthetic tests. They also lower the graphic settings to the absolute minimum, they basically force/hack the render engine into a CPU bound mode, with vsync off.
Thats oki for the CPU benchmark and i guess thats where this confusing idea comes from that by lowering your resolution from like 1920 to 1650, u suddenly become CPU bound or the CPU works “more”. Most games are either always partially CPU bound or only if u lower settings/resolution to a absolute minimum.

My guess in GW2 case, is that the engine is suffering from poor thread locking and synchronizing code. Basically some threads are waiting for others to finish and will lock each other. The CPU/GPU behavior, usage i see is very bad and fluctuates extremely. There seem to-be a multitude of problems, ranging from bad render code to bad threading code.

GOSU.9574:

Now about WvW being a copy of DAoCs RvR…laughable I must say. I can only imagine you didn’t play DAoC in the first 6 months when everyone was getting 1 shotted from range.

I agree on the trolls parts, but hey thats what we have those topics for!

Actually on DAoC i was a Beta2 tester and hell after this buggy mess going from beta2->3 i had enough of beta testing, so yes i was starting DAoC RvR kinda 6 months after the release.
Anyway i was more referring to the base concepts of a open mass PvP zone, with keeps, stat boni and siege warfare, which is missing in Rift/WoW/swtor.

Hey they even “copied” the low FPS/Lag problems in DAoC RvR to GW2 :p

Craven.5468:

Not everything is a contest of extremism.

Thats why i prefer “flavor” over “different” as description, while i also use “WoW skin” for games like Rift/SWTOR. I would say EvE and Vindictus are different, since at least some core mechanics are changed drastically.

mhh what game are u playing that is so “different” than WoW? Vindictus? Minecraft? EvE?

I play GW2 and except for the WvW part, which is a DAoC “copy” its just a new “flavor” in the same old MMO fantasy genre, but guess thats the safe money bet these days…. or not pointing at EA/Bioware.

I also don’t like that they had to drop the core concepts of GW1 in favor for a more WoW/Warhammer like skill/trait system. Its more flexible than WoW, but i miss the GW1 system for PvP/WvW, compared to how many “real” combinations i get out of the current system. In short u still only have like 3-5 builds per class, with small variations. This feels more like Rift, with its multi tree system.

Maybe u define “different” um differently

Ruggy.7819:

making use of light and smoke fields with different finishers made my earlier Ascalonian catacombs explorable run a complete piece of cake

Thats also what i noticed, ae “blast” heals and blinds really outperform most other combo fields, especially if u stack a bunch of guardians/wars in front of a WvW door :p

Enfurion.6478:

Regarding Chaos Armor, it is a lot better than you think. Aside from the random boon/condition thing, you are guaranteed Protection.

No its not and also stated at the wiki entry, the “guaranteed Protection” is only given on the 40sec skill version to yourself, not for the combo field version. This is often misstated and makes the combo field version much less usefully, sine it only nets 2x 33% “lucky” protection/blind chances on hit, which will noticeable reduce the incoming dmg, which is the main reason u want chaos armor in the first place. In addition the AE version of chaos armor only last 3sec instead of 5, so its even less usefully, since u need to get hit in this 3 sec and than have the 33% chance to get your protection/blind also just for just 3 sec.

I still like the combo/leap chaos armor, since mesmer have many usefully etheral fields, simply don’t assume a null-field + ae chaos armor makes any difference for dmg, just pop the null-field if needed and add the ae/leap chaos armor as a small lucky bonus effect.

The combo with chaos storm + chaos armor “looks” so strong, since u get 1-3 aegis out of the chaos storm, thats what most will notice, simply don’t overestimate the chaos armor part if used with other ethereal combo fields.

I guess most WvW or sPvP mesmers hope they don’t “shine Sauron’s Eye” on us. Mainly because there are some buggy traits and combinations that feel a little to good to-be true.

If they would “fix” Condition Duration, HP for Clones/Illusions and add the “missing” extra bounce to clones, i pretty much know what build i will play, until they “rebalanced” it again.

Atm i have no problem at all in WvW/PvP, i’m actually very surprised how well i’m doing and in WvW i rarely loose any 1v1 or 1v2 fights.

I simply rationalize this with “my build and play skills are so good”

@juno.1840

In PvE i always prefer a bleeding build over a confusion build, because its much more reliable and spamable, than a confusion build.
In my full crit/bleed build, i apply like 10 stacks of bleeding in like 3-4 seconds to 1-3 mobs, which if i’m lucky can ramp up to 15-20 stacks. This needs nearly no skills or setup at all, compared to how confusions work. Also in PvE most mobs wont attack once per second.

The only exception in PvE, where i wished i had a confusion build is on crazy fast attacking bosses, that already have 25 stacks of bleeding.

VirtualBS.3165:

Using v-sync without triple buffering will reduce your fps to 30 the moment you lower from 60+ to 59fps, and stay between 30 and 59fps.

The triple-buffering option in the nVidia control panel only works for OpenGL, so you’ll need for the game to support triple-buffering or use a 3rd party utility (D3DOverrider).

Since the “.300” nVidia drivers there is a alternative called “adaptive vsync”, it basically just enables/disables vsync dynamically, if i understand it correctly.

Rpgtabbycat.5869:

an example of the combo field: Ranger drops the Healing Spring, Warrior does a whirling attack inside of it and suddenly the Healing Spring becomes an AOE heal.

Nice in theory, kinda useless in practice. Its not a “AOE” heal, first those bolts heal for like 125HP, than there will be like 5 bolts flying randomly around in a few seconds. Now simply try to “catch” such a bolt, even if u perma “whirl” inside the spring, u might get lucky and get 2 bolts every 10 seconds for a whooping 250HP heal….
This is the same with all whirling finishers, the bolts will mostly miss. They need to adjust those either to deal more dmg/heal or make those auto target players.

The biggest problem atm, is that its really hard to effectively test combo fields, no one has a real idea how much dmg/heal/effect u get out of what combo. The most testable are blind/condition removal fields, but than again how do i test this if i have no reliable source to get conditions. This was much better solved in GW1 and the test area.

“The lower the resolution you run, the more work your Cpu has to do because the gpu has to render less pixels & there is physically less on screen.”

What the heck are u talking about?
This makes no sense at all, u have any idea how GPU and CPU interacts in game engines? U can obviously overclock your 500€,- CPU and are “over” 30 years, so spend some extra time to understand the relation between GPU/CPU/Resolution and more specific vsync, tripple buffering and the new “target framerate” or “adaptive vsync”.

My “guess” is u referring to the interaction between GPU/CPU load and vsync on changing resolutions, which in some extreme cases will result in higher CPU load, if u go from 20 FPS to 150+ FPS. This only happens on drawcall bound PC games, if vsync is disabled and u go from like 1920×1024 to 800×600.

Sorry to many questions at once

This should get u started: http://www.youtube.com/watch?v=7eZ0NNd7b60

sPvP aka GW2 BG’s, scales u up to lvl80 and unlocks all traits/skills, its called “heart of mist” check the gw2 wiki to get u started.

WvW scales u also up, but as mesmer i recommend at least lvl20-40, so u have access to some very good PvP traits.

I enjoy a feedback/greatsword build in WvW, but will try to modify it for confusions. In WvW u see zergs all the time, so ae dmg/effects are nice.

A mesmer is a fun to play class, u can do a lot of crazy stuff, but u need to read this forum and guides/videos to know all the nice tricks we can do. Some tooltips are also still wrong and traits don’t work correctly.

Have fun!

I guess the difference is that confusion can be applied as ae effect on a whole zerg, while u cant do this with bleeding. U can also pop a iWarden in one of the 3-4 combo fields for extra ae bolts.

Like noted confusion works best against trigger heavy classes in sPvP and it works surprisingly well in WvW, if also combined with reflection.

I still don’t like confusion builds personally, since its really hard to reliable do any significant dmg. U also have to abandon staff/gs in WvW, assuming focus is a must have for WvW and u want the iMage.

Keep in mind that if they target u via the ui feature, every cloaking ability will break this, so safe your cloaks and decoy, until they spot u and attack u. Also max. toughness combined with the cripple feat, will make your clones much more annoying. Thats because clones will last 2-4 hits and most ae abilities wont completely kill them. This means instead of 3 dead clones, u still have 3 clones out that are at 10-20% HP, making them viable targets, which than cripple for maximum annoyance.

Like noted, the key is to perma recreate clones + stealth, so they need those 1-3 seconds to search for u, while they might auto attack clones and trigger conditions/cripple. Also try to get in there “dead” view spot, so that they see clones in front of them, while u try to get behind or out of the view.

Cbuzz.5083:

At that low resolution you are passing most of the work to your Cpu which isn’t particularly strong. Why did you even get a 660Ti for that resolution?

A 150€,- CPU for GW2 “isn’t particularly strong”, there is something seriously broken with GW2 than. Also what exactly means “passing most of the work to your Cpu”, what has the resolution to-do with “passing” GPU work, if u cap to vsync anyway.

PS: Just noticed your sig “3930k@5Ghz” oki a kitten kid, this explains why. So yes u can “solve” the GW2 crappy engine performance, by throwing excessive money at a problem, which should not exist in the first place.

Cloesd.3025:

Then someone were to use a blast finisher, to give AOE chaos armor:
http://wiki.guildwars2.com/wiki/Chaos_Armor

Ontop of this you gain Chaos armor:
Which upon being hit gives you:

• Protection (33% damage reduction)
• Regeneration (390~ HP heal).
• Swiftness again.

Enemies that hit you get:

• Blindness
• Cripplied
• Chilled
• Weakness

Keep in mind that chaos armor has only a 33% chance to give u either boon for 5 seconds. U may also end up getting 3x swiftness.
If u don’t get hit, chaos armor via combo, does nothing at all except for a nice purple bubble! Which like some animal coloring may indicate “don’t shoot/eat me” !

The problem seem to-be bad or buggy memory reservation code and still some memory leaks. GW2 will “use” up to 1.5GB, while reserving around 1.9GB of RAM, during longer gaming periods. This does not look like a problem on a 3GB, Win32 system (Note: only 3GB are normally usable on a 4GB/32bit windows). The problem is that a application under Win32 can only use up to 2GB, so reserving 1.9+GB will at some point exceed the 2GB limit and GW2 will crash. (@ANet: lazy memory manager coding! Ask your console guys how to manage memory!)

On a 64bit OS, first u get to use the full 4GB memory and also the limit of 2GB per application is lifted to what max. ram your system has. So until they fixed all the memory leaks or implement some sort of memory flushing, on such out of memory errors, the only fix is to use a 64Bit OS.

PS: @Anet, put some out of memory macro on your memory allocation calls and flush D3D texture memory or recreate the device! This is still cheap and better than a crash and having to re-queue for WvW.

Agemnon.4608:

Bottlenecks are when there’s a part of your computer holding all other parts back.

This article may be old but the principle still holds:

http://www.tomshardware.com/reviews/game-performance-bottleneck,2737.html

Going with your article, doesn’t the article show the exact opposite?
80% of the presented games did run 10-20% faster going from 1 to 2 cores, while going from 2->3->4 made either no difference at all or only in a few games and than it was also just 10-15% more?

Going by this article, having more than 2 cores for those games seems like a waste of money. If i could drawn any conclusion, than that “mostly” 2 cores are enough and that u have to actually test every game to be able to make any valid statement.

“Actually you will hardly see an increase. Youll will notice some, but only due to the extra memory and architecture of the chip. If you were to benchmark your system indepth, you wouldnt find much of a increase at all. mainly due to the extreme bottlenecking both with the card and the fact that you would have an older motherboard that would struggle to cope with the new GPU”

Not relevant for the OP, but what the heck are u people talking about?
I just garbed a GTX 660 from MSI and my FPS went from 25 to 60. I also can now crank up most settings, except shadows?
So doubling my FPS, while increasing the overall quality is “hardly see an increase” and “wouldnt find much of a increase at all”?

Renegadeimp.8439:

If you have a low end system now, basically any money you put into it will just be a waste.

This doesn’t make any sense and contradicts all benchmarks i have seen in the last couple of years, what are u talking about?

If i have a old Core2 Celeron “famous” 4300 and a 9600GT and now replace the 9600GT with a shiny new gtx 660TI, the FPS will triple or quadruple? So how is this a “waste”?

Agemnon.4608:

“SO my advice would be grab a cheap PentiumG dual core and spend the extra 100€ on a better graphics card”

Any of the current generation cards would be bottlenecked by a dual core.

What do u exactly mean by “bottlenecked” ?

So lets assume i have 300€ budget to spend on a upgrade, for GW2 i would buy a 50€ CPU and 250€ GPU.
Do u mean u get better FPS, if u spend 150€ on a CPU and 150€ on a GPU? If so can u provide some data links that shows this reasoning?

I also don’t get how a SSD will “bottleneck” FPS? The SSD will improve loading times and maybe texture streaming, depending on the game. Since GW2 does not use texture streaming i don’t understand how a SSD will impact FPS, while playing GW2?

Not sure what u talking about, the data i found contradicts this?
http://www.tomshardware.com/reviews/guild-wars-2-performance-benchmark,3268-7.html

The table shows that even a PentiumG Dual Core outperforms any AMD system. It also shows that a 50€ PentiumG dual core will get u 80% the FPS of a 150€ i5 quad core.

So actually u get 80% performance for 1/3 of the cost using just a dual core SandyBridge intel CPU.

This indicates to me that GW2 scales very bad with cores and its more a matter of raw speed per cpu cycle and CPU architecture.

SO my advice would be grab a cheap PentiumG dual core and spend the extra 100€ on a better graphics card

No there is no “consensus”, but i simply use this rule. What “special” effects can a rune set(1-6) or multiple 1-2 rune sets give me. If i just want the stats, i just put crests in my armor+jewelery, since its 288 total stats vs 211 + a effect.

As tip all the duration % runes and “on crit” runes are nice, since its hard to get this via traits or other means.

As example, u can get +45% bleed duration via 3 different sets, which on a condition build will outperform any other thing u can put in.
The boon duration runes are also very strong, since it allows u to stack might boons faster.

So try find effects, which u normally have no access to or hard to get otherwise. For me runes are defined by there effects and not stats, thats where i would use crests over runes.

I also don’t like the often noted “+10 stat 2% crit” set, since i normally dont care about the 3 stats i don’t use and having +60 vitality in a PvE power/crit build does nothing.

Thedenofsin.7340:

Garruda.4201:

illusions also take on the the “proc” stats for weapon sigil…so a sigil of earth would grant the iDuelest and clones you create while using the pistlol the 60% bleed proc.

Has this been verified?

Obviously not, since its easy to test in the mist and no sigils and armor rune effects wont transfer to illusions. If “Garruda.4201” did some test at all, he did it with the sharper images trait active, assuming the bleeds come from the sigil, while they actually came from crits + trait. If u don’t take this trait and the earth sigil u can easily test and verify that no illusion will apply bleeds, except the staff clones that does this anyway.

Yeah this explains some problems i also noticed, shame on u arenaNet. Not only are tool-tips righteous wrong or badly formulated, but skills/traits works in such a strange way that without forum/wiki u have no clue whats going on.

Looks like a word is missing, “Reduces recharge of pitol skills by 20%.”

Btw its “pistol” not “pitol”, also should note “Reduces recharge of your pistol skills by 20%.”

I assume it only reduces your skills and like all other 20% skills not the abilities of phantasms. It does stack however with the 20% Illusionist Celerity feat, only the UI shows one of those, while the recycle UI shows the final recharge times correctly.

Hi,

the sPvP combination “Rabid” is only available via dungeons or as “green” karma vendors items. Is there a reason why all other sPvP combinations can be crafted? This seems like a bug/oversight in the crafting system.

So can we none “Power” users pls also get exotic accessories and armor via crafting like everyone else?