Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
I'll nab that and give it a shot when I can here. Thanks for the information.
I re-tested this situation with the public beta.
I played Endless 2 since its a much more "hands off" level to test on. I used the same unit layout as I did the last time. I tried testing at Speed 16 from start to finish. Finish in this case, sadly, being the same app crash and "DefendersQuest.exe" has stopped working message.
The game ran noticeably better for longer. Previously after wave 60 things were largely becoming a slideshow and by ~63 it would crash. This time up to wave 50 worked great. At 60 there was a noticeable drop in performance. By 70 it seemed to have a ton of frameskipping and the input was becoming unresponsive. 71 - 73 it was basically a slideshow and inputs would be ignored or massively delayed. On wave 74 it crashed.
A very noticeable improvement, but it still seems to getting worse performance the longer a level goes on. Is it possible that its a garbage collection issue of some kind thats building up over long missions like these? As I've not noticed any significant slowdowns in non-endless missions even if the amount of enemies and partile effects flying about the screen are more than in the endless levels.
I know in a program I was working on recently I had two dumb mistakes that gradually brought it to a halt. One was that a minor particle effect had a typo in the code for its death timer so despite fading away visually it was never removed from memory so the game would track it forever. The other was with a projectile that could pierce through multiple targets where I was storing which targets it already dealt damage to in a list data structure I would create so that I could ensure it didn't damage the same target multiple times. But, I forgot to destroy the list after the projectile was destroyed so even though the projectile was gone I still had a list and new projectiles meant more and more lists.
I know some programming, from hobby projects, but never really knew much about external debugging and performance tools so its certainly useful to hear about them.
Unsure if it might be related to this issue, but through testing I've found that the status effects from the "Dragon Fire" (Main character spell # 6) really bogs things down. It certainly slows things down while everything has the burning effect going at once. Even after it wore off things never seemed to go back fully to the same speed again. Though its possible that its just because by the time the effects end the game overall has bogged down more than prior to casting it.
EDIT - Actually, just crashed again in a way that might have useful information. I had tried doing the first challenge of New Game+. Only units available at the time were my Berserkers and my Archers. So aside from my "Dragon Fire" spell there wasn't really a heap of fancy status effects or projectiles flying about. I made it to the last wave, 50, and it was nearing the slideshow phase. I didn't crash though, I died and failed.
However, after my defeat in that challenge I went to start the next storyline mission "The Battle of Karsk." As soon as the actual level started the game was already a slideshow and crashed out before it even started spawning the baddies.
So I'd guess whatever is causing the issue can occur even with just the minimal things available at the time, enemy variety for that mission, Berserkers,and Archers.
Anyways, I'm going to go ahead and finish deploying 2.0.3a, fix a few conventional bugs, and then hop onto the memory profiler.
The beta version is an improvement for sure. Can't comment on the other updates since there hadn't been any updates, aside from the beta which I wasn't aware of at the time, after I reached the long missions, like Endless 1 & 2, and started having crashes. Either way the beta is a noticeable performance boost and aside from these crashes I've not encountered any other bugs, before or after opting into the beta.
As for my system specs...
Operating System: Windows 7 Professional
System Type: 64 bit operating system.
CPU: AMD FX 6200 - (3.80 GHz) Six Core Processor
RAM: 16 GB
GPU: AMD Radeon GA-990FXA-UD3
If you need to know anything beyond that just lemme know what.
Using version Steam DQ 1 2.0.4 c (also says Hardware + ANGLE in bottom left), which I believe is the most current version, I wasn't able to crash it. I tried damned hard too as well.
I re-tested this version by once again playing Endless 2 and using the same unit setup as before.
I went up to wave ~107 or higher before I quit. No crashing. I put it through its paces by doing all I could to crash it such as pausing and firing off my frenzy, push, dragon fire, and heal spell while also placing a crystal which I shot with lightening. My lightening is modfied by the chain lightening book, but even with all that, and constant firing off of the lightening, it still held strong.
Granted, the framerate by this point was boardering on slideshow due to the number of enemies and particle effects and such but thats to be expected. Only thing that I think could really "solve" that issue would be to limit how many enemies could be on screen at one time and hold off further waves until room opened up.
The actual problem though seem to be fixed. No evidence of a memory leak as there was no crashing and the slowdowns were only there from the sheer volume of enemies and objects and not due to progressively bloated memory usage. Checking at times in task manager to see how much memory was being taken up showed it holding quite stable, with increases of course as the waves progressed but none of the previous ones permanent ballooning.
Kudos on the fixes! Looking forward to seeing you / your groups next project.
For what it's worth, I've added a few more tweaks that might speed up performance a bit more (2.0.5b, test_public). I agree that it was a bit silly to let a level like this get so full of enemies -- something we'll be more careful with in the sequel.