fluffdragon.1523

@"KrHome.1920" I like how you essentially ignored everything I posted except for one thing, which you immediately spun into a tangent. But thank you for pointing out that supersampling is rarely a good idea and confirming pretty much everything else I'd said :+1:

These threads keep popping up like weeds, but I'll answer here, too. **You will only ever see (extremely) diminishing terms with newer hardware.** It's entirely possible to see comparable performance running an old, base model GTX 660 (2GB) card on an antiquated Phenom II quad-core chip with only 8 GB of total memory. Why? Because that's still more than GW2 will ever make use of. There's also the fact that DirectX 9 (c, I think), the graphics API which GW2 makes use of, is *singly-threaded,* and moreover, *single-context,* meaning that only ONE thing can ever occur at a time, and that's essentially the main renderloop. What does this mean? That when a lot of things are happening on the screen (like an intense fight or a busy corner of LA), more and more has to happen within that one thread of execution -- the *same* thread as drives your rendering, texture / model IO, GPU on/offloading, *and* your DirectInput mouse and keyboard activity. This is why in really bad situations where framerate drops below about 14fps, you can start to see your keypresses fail to register. With an overhead of about 72ms per-frame at that rate, it's obvious to see why a keystroke might not get picked up by the next poll. A lot of people will tell you that having higher boost clocks on your CPU (or buying an Intel chip, which formerly had the best single-threaded performance) was the only way to go, thanks to very, *very* antiquated logic surrounding these sorts of graphics APIs. Unfortunately, that's no longer meaningful, particularly because most every modern game **except** GW2 has moved onto things like DX11 and DX12, if not Vulkan, which make use of asynchronous rendering and IO, entirely alleviating the problems outlined above. So ... no, @"Valtari.9041" . Your system is already overkill for the game. My recommendation at this point is to use VSync and the FPS limiter. The less extra stuff your CPU and GPU is doing (read: rendering more frames than your monitor can even display), the easier both can ramp up to take on that extra workload when necessary. It isn't a perfect solution, and no, it won't prevent those kinds of horrible frame drops you see, but it'll help smooth things out a little. And, as @"Infusion.7149" mentioned, consider using the DX12 proxy layer or dropping in [DXVK](https://github.com/doitsujin/dxvk/releases) into the "bin" folder of your client (these are files d3d9.dll and dxgi.dll under the x64 folder). The latter will make use of Vulkan and provide similar functionality, and is what's used by a number of linux gamers (as well as Steam, for allowing compatibility on linux systems).

> @"Dawdler.8521" said: > That said, interestingly enough there is *something* weird going on in the graphics pipeline, seen that in the past (I dont do so much PvE now). Boss fights could chug along at 25 fps and then suddenly, out of the blue with little change in visible player activity, it went to smoother 50+... only to drop down again like 5s later as the fight continue. The flaw in the graphics pipeline is that with DX9, input processing, GPU on/offloading, and IO all happen from within the graphics pipeline. You can observe this when an asset is taking a while to load, and the map appears empty for a time, then suddenly huge swaths of it populate -- or how at very low FPS (say, <12), skills can fail to go off because the difference between when your OS registers the keystroke and when GW2 attempts to process it is enough to void the buffer. This is because DX9 is a synchronous runtime -- event processing and rendering *cannot* occur in separate threads; the context simply won't allow it. Playing around with DXVK (the DirectX to Vulkan layer), and most of these issues are resolved, outside of GW2's asset loading problems (because in DX9, loading an asset and sending it to the GPU all occurs during the renderpass, effectively "between" frames, and this cannot really be circumvented based on *when* the engine initiates IO). *(As an aside, I've also observed substantial performance improvements in other titles by simply switching from DX11 to DX12, the latter of which fully supports asynchronous rendering and IO operations, allowing much, much faster loading (and deferred loading) of asset data, as well as the rendering thereof. Keep in mind that this **only** occurs for titles which fully support the target API; GW2 will only see nominal performance increases **unless** it were rewritten in such a way that made use of the new API features)*

Having played Necro almost religiously in GW1, that's what I was most excited for, and VERY pleased when the early trailers showed off the Life Siphon ability. Though to be fair, what made certain I was getting GW2 was playable Charr. However, during the initial missions -- and largely because of axe -- I suddenly felt that it just wasn't going to work out for me. At least, not in the way they'd implemented combat (I was not part of the betas, so it was entirely new to me). Mesmer was the second option I'd poked at, but it didn't really "feel right" for me, either. On a whim I picked Thief as the next thing that seemed interesting enough for a first character (despite really NOT liking Assassin in GW1), and instantly enjoyed the oddly frenetic playstyle of the initiative system. I've kept with the profession ever since. (My other characters, in approximate order, are: a Warrior, a Mesmer, a Guardian, and a Necromancer (reaper), who surprisingly bears the name of the one I played in GW1)

> @"Sobx.1758" said: > Thief is still perfectly fine for roaming, not sure why you'd use some third party site to determine if the class is useful or not. > > > And sure, I disagree with that balancing attempt, because if they don't want that mobility skill spammed then they could add cooldown on top of ini cost rev-style, so using it once wouldn't kitten the rest of the weapon skillset, but I wouldn't say that nerf somehow kills the shortbow or -even moreso- the class. At the risk of spamming this thread with a reaction image: **that's the joke.**

For me, it was the constant melodrama without any inner resistance or meaningful ruminations, combined with the nature of his interactions with the player, constantly switching between quailing over the least impediment, and grandstanding for the pact. Or, put simply: "Commander? A word …"

Bored and cruising the forums, so I'll post some thoughts from another thread of mine, as well as thoughts from chattering with a fellow player: * **Thief:** focus off-hand with synergy around Shadow Arts and *Deception* skills, inspired by the Am Fah and Jade Brotherhood gangs of Factions -- trades stealth and damage for utility, and the ability to summon shadowy manifestations, as well as a replacement to Steal, inspired by the GW1 Assassin kill, "Shadow Form," which grants defensive bonuses for its duration. * **Necromancer:** ~~basically Ritualist~~ (we all know that Revenant is a heavy-armor Ritualist, so this is moot). A greater focus on conditions and *Corruption* skills, styled after the miasma and affliction which plagued Cantha during the time of the Great Hero. Gains additional bonuses for the number and type of conditions it is afflicted by, and gains a new Shroud type, which transforms them into a shambling monstrosity. Weapon unknown at time of writing. * **Warrior:** honestly, what warrior lacks most is better ranged combat; both rifle and longbow, while effective, are kind of lack-luster. A shortbow warrior focused on skirmishing (with synergy in the Arms trait line, for condition application) would be an interesting direction and, with a few disengages, could turn Warrior into a really good mid-ranged combatant. * **Revenant:** *actually* Ritualist. Yes, I want revenant to get a ranged staff (unlikely, due to probable engine limitations) or a scepter weapon and the ability to conjure stationary "bound spirits" like the Ritualist in Factions. Vastly different play-style and possibly some much-needed application of range outside of hammer and Renegade. * **Engineer:** more of a focus on Alchemy and *Elixir* skills, and perhaps a slightly more profane take on both its use and its methods. Not quite sure what, but something that dabbles into the more real-world hunt for immortality, longevity, and very questionable applications of heavy metals. Scepter could be an interesting weapon choice here, as well as a hybrid mid-range staff as an alternative. * **Guardian:** counterinituitive though it was, I kind of liked the *conceptual* change to Guardian's Virtues, and think a more active playstyle around them might work again. Perhaps instead of firebrand's offensive (and condition-oriented) design, we could focus more on a classic supportive role, building off of the Virtues and Valor traitlines, whilst simultaneously providing something similar to the "Protection Spells" of GW1 Monk. Weirdly enough, I'm liking the mental image of ranged hammer for this aesthetic. * **Ranger:** neither existing Ranger elite spec has exploited their iconic archery, so something with synergy in Marksmanship and Nature Magic traitlines and *Survival* skills could be interesting -- really have a classic pathfinder feeling to them. * **Mesmer:** I'd honestly like to see mesmer move away from shatters and churning of illusions, though obviously we can't get rid of the latter. Perhaps give them a treatment similar to Scourge's shades; an option to either summon all 3 lesser illusions, or 1 big "greater illusion" that provides a far more persistent and formidable ally in a fight, and could satisfy @Ryo.5824's request for beefier illusions. * **Elementalist:** ... sorry, I've got nothing here. The profession is already such a mixed bag due to elemental attunements, that I really haven't any clue what to offer it.

Just looked into that GWMists thing or whatever ANet wants us to sign up for. Check out all the great new competitive builds for Teef! 

On the topic of player-based balance in PvE, I'd say this: specializations were a mistake. Back with the old pre April-2014 trait system, we had more traits and more diversity to work with. Sure some were extraneous and a few *needed* to get folded into baseline to really power up classes and stop having so many builds revolve around a small subset of traits, but that could have been done and the issues addressed, rather than overhauling the entire system. Not to mention we used to have 13 or 14 major traits *per trait line* compared to 9 now, *and* lower-tier traits could be slotted into higher brackets! Moreover, when we *still had attributes from traits,* it was fantastic. On my Thief, I used to run 6 Acrobatics for +3000 health, coupled with the old "Practiced Tolerance" that converted 7% of my precision to Vitality -- with 6 Critical Strikes, I had an additional ~1500 health on top of that, outstripping what I can do now with Daredevil and an ascended Marauder's staff. In short: we lost the vast majority of our build diversity when attributes were **only** granted by equipment. **A Proposal** I write this next part quite seriously: I'd love to see a return to [the old trait system.](https://wiki.guildwars2.com/index.php?title=Trait&oldid=781884) Not a complete reversion, but an adaptation; give us all 5 core trait lines at once, rather than forcing us to pick 3 (even if based on the allocation of trait points, that was the most common build setup). This would also mean it would be easier to add additional traits over time to fill in gaps, rather than forcing the retirement or movement of various traits to satisfy the restriction. There's also the fact that it would remove the hard power plateaus from the game when each trait line is now unlocked; start us off on traits at level 11 again, like back at release, and let the accumulation of trait points at each level be the limiting factor (and further incentive to level up -- meaning more potential gem store purchases). To solve Elite Specializations, make use of the old dual-class system from Guild Wars 1. For example, a Firebrand would add in the *Firebrand* trait as a *6th, additional* trait line, rather than forcing players to sacrifice one. Instead, it's the allocation of points which makes the compromise: do you take 6 ranks in your new trait line? 4? Maybe you only want 2 ranks. Regardless, it should be up to the players to decide. While I wouldn't be opposed to seeing the return of attributes on trait lines, I'm not sure how well it would work -- and likely was removed *because* of the risk of duplication from Elite specializations. However, with the proposed solution, Elite traitlines could waive extra attributes, serving as a further compromise -- dramatic changes and potentially more powerful passives, in exchange for better stats. Kind of like using [Runes in Guild Wars 1](https://wiki.guildwars.com/wiki/Rune), where higher tiers sacrificed more HP, but resulted in [their own absurd little build](https://wiki.guildwars.com/wiki/55_Monk).

> @"Dadnir.5038" said: > Honestly, in my opinion, _defiance_ need to be worked on, stat sets need to be more specialized, reflect effects should at least block incoming projectiles, auras need to be good by themselve (not only good if 3-4 traits are dedicated to them), mobs need to rely more on boons than on their inate attribute and the environment need to learn to use conditions in a healthy way (well that last point might be difficult since it's hardly the case in competitive modes already). THIS. Everything was well and good before we had Defiance to arbitrarily make encounters more difficult or otherwise prolong them. Vanilla GW2 wasn't a grueling challenge, but it still required either the clever cooperation of a small team (like a standard party) OR a lot of people working together. Nowadays, I watch world bosses and random veterans alike fall from DPS and conditions faster than their break bar is taken out -- because it's *simply more efficient.* ANet even had to [disable certain items in Fractals and Dungeons](https://en-forum.guildwars2.com/discussion/117145/game-update-notes-october-27-2020) simply because they were being used to work around a flawed mechanic:

> @"Pifil.5193" said: > > @"Eloc Freidon.5692" said: > > I say we let this event horribly fail and save the support marks for after. You'll be able to buy the boxes and trident after. > > I'm sure they've calibrated the tiers on this one so that it basically can't fail. **Taimi:** Look at that—you broke Scruffy's sarcasm meters.

> @"NorthernRedStar.3054" said: > Patch TL;DR: Thief is not allowed to 1) be able to duel _any_ classes; 2) thief is not allowed to be able to avoid other classes' damage or retain mobility despite their harshly nerfed damage. Isn't that what everyone on the forums has been asking for? According to what I usually read, if Teef had access to Infiltrator's Arrow, we could insta-gib the entire fortress of NPCs, no-clip through the walls and doors and right into the center of the keep, AND manage to kill a 30k health mega-turtle with a single glance, T Hanos style. Or was that from our access to stealth?

Rifle Deadeye's biggest flaw is that it works best when you use Kneel -- which is entirely counterintuitive in content such as Fractals, where you might have to move quickly or evade lethal effects to progress, and hunkering yourself down into one spot can potentially jeopardize your team mates. In PvE, however, there's far less of a problem. Find yourself a corner, use Kneel, and blast away. Or if you want, just roll pistols. Less damage baseline, more might generation, no change in mobility.

> @"Sammyjoe.4271" said: > > I'm asking for a well thought through balance patch for conditions. Power users got hit hard in a patch pointed toward us. My MAIN argument for conditions is, why does it make sense to let them deal big damage and have the choice of being extremely tanky as well? There should be a trade off like power users have. Being able to put big points into vitality, toughness, and even healing power while still being able to dish out huge damage blows my mind. Current stat prefixes certainly are contributing to this problem. Seraph, Trailblazer, Carrion, Dire, Plaguedoctor, and Bringer are all *primary* condition damage with vitality, toughness, an/or healing. Using *secondary* stats, Magi, Apothecary, Settler, Shaman, and Marshal can be strong alternatives with similar durability. Now, if you're talking PvP, where mix-and-match prefix (and suffix) combinations don't exist and everything is regulated to set values that sum to 3000, well ... that's a different matter altogether.

> @"Zephire.8049" said: > > In the text of the hero panel it does, but not on the character selection screen. If it's been a while since you've done the story and/or have a lot of characters, it can be a royal PITA to log into each character individually to see if they've joined an order and if so, which one. A symbol or other indicator on the character select screen would be a huge QoL improvement and save players a lot of time, especially as there are HoT achievements tied to each order that are available only to characters in that order.  Particularly for the fact that there USED to be a mechanism that showed the "account medals," (currently unfilled, here; stock image by ANet for the Guesting announce, [pre-megaserver](https://wiki.guildwars2.com/wiki/April_2014_Feature_Pack#Megaserver_system)). You'd think they could adapt this to show it per-character as well, in addition to account-wide status.