Pages: [1]
  Print  
Author Topic: OpenArena hardware, framerates, latency, etc  (Read 7205 times)
*Ruthless*
Half-Nub


Cakes 0
Posts: 81



« on: March 30, 2014, 10:07:01 am »

Hello everyone,

I've been playing OpenArena for quite sometime. Like most, I've also greatly improved since my very first days in OpenArena. With that said, of course I have gotten a better understanding for the game. I'm sure we all know that OpenArena/Quake on certain hardware gives others an advantage. What I mean by that is, someone with a 2003 computer running OpenArena playing against someone with 2013 manufactured computer has the advantage over someone with a 2002 computer.

Although I know this, I want to know why is this the case? First, I want to say I have no problem with the advantage at all. I want to know why they have the advantage over others. I believe I'm very intelligent, but my education was never based on knowing certain technology.

My example:
I have to say I've noticed playing a very nice friend of mine, that he usually shoots me before I see him. Now some may say, he's better. That indeed is true, he's very skilled, and a well known player for such a long time. But, as I admit he is better, I'm by no means a scrub. I know when something doesn't feel proper. I take much pride in my skill, and I'm usually aware of many things going on with my play. My computer by no means is up to date. I sometimes run OpenArena on Windows XP, and Windows 7. I've noticed that he has a greater range of shots than I, and sometimes even shooting me in the absolutely tightest window. Sometimes when it looks like a wall is covering me, he is still able to connect with the shot. What I'm speculating is that hardware plays a great role, in someone being able to perform.

Also, I find that I sometimes have a delay in movement. Sometimes I take large breaks from OpenArena. But rust is very rare to me, but sometimes I feel that theres a huge delay in actions. When I say delay in actions, I mean turning a corner and feeling frames feel a bit slow. Taking damage .5 seconds late, and having to react to that .5 of delay. For me it's a big difference, because the way I developed my skills was based on precision, accuracy, and knowledge.

How big can the server location, and maintenance play a part? Can anyone comment? I want more knowledge in these sort of things. Let the discussion begin.
« Last Edit: March 30, 2014, 10:28:55 am by *Ruthless* » Logged
Gig
In the year 3000
***

Cakes 48
Posts: 4199


WWW
« Reply #1 on: March 30, 2014, 11:11:37 am »

There are many things involved. Of course, your internet connection is very important (e.g. having lower ping is better), but also other factors are, such as your com_maxfps value and other values you can find infos in the "tweak" page of the wiki. Optimized settings (e.g. a correct "rate" value) can make a difference. Also the game physics kind selected by the server in use affects players' gameplay.

About the "they hit behind corners" thing, check you have "delag hitscan" option enabled, too (somethere around the net there is a sort of very interesting FAQ from the author of the Unlagged Q3A mod -similar to OA delag feature-, which explains what happens when firing with an hitscan weapon on an unlagged server. Maybe it was this one: http://www.ra.is/unlagged/ - Note: OA delag feature has similar purpose, but probably is not directly derived from Unlagged mod... I don't know exactly from where "OA delag hitscan" code comes from).

Some links with interesting readings:
http://openarena.wikia.com/wiki/Framerate
http://openarena.wikia.com/wiki/Lagometer
http://openarena.wikia.com/wiki/Tweak
http://openarena.wikia.com/wiki/Game_physics

Gig from cell phone.
« Last Edit: March 30, 2014, 11:26:50 am by Gig » Logged

I never want to be aggressive, offensive or ironic with my posts. If you find something offending in my posts, read them again searching for a different mood there. If you still see something bad with them, please ask me infos. I can be wrong at times, but I never want to upset anyone.
grey matter
Member


Cakes 8
Posts: 381

>9k


« Reply #2 on: March 30, 2014, 03:16:19 pm »

How big can the server location, and maintenance play a part? Can anyone comment? I want more knowledge in these sort of things.
You might want to take a look at the Quake 3 Unlagged documentation (most of its code is included in OA).
Logged

This space is for rent.
fromhell
Administrator
GET A LIFE!
**********

Cakes 31
Posts: 14471



WWW
« Reply #3 on: March 30, 2014, 04:37:32 pm »

I do know older Nvidia hardware (Geforce2 etc) have drivers that default the vsync to on. Vsync is a hugely common cause of input latency.


r_detailTextures 0 will definitely help some weaker cards since modulated blend is expensive.  You might also want to try r_primitives 2, which will use compiled vertex arrays - some cards are buggy don't let their extension be detected for this to turn on automatically.  Sometimes it may also require r_ext_compiled_vertex_array to be set to 1.

adjusting r_lodBias to be higher will also help, as most OA models are high poly and include reduced versions.

Around between 0.7.7 and 0.8.0, the implementation of SDL definitely played a part in making the input more sluggish.  Some people deny this however and want me dead over it (despite a conducted player poll saying otherwise), I think it's gotten better in the git master however since the input is a bit more responsive in lower framerates lately...   and SDL input latency doesn't just affect this game, but other games where input is critical (like VVVVVV which is fixed to 30hz and the input response is literally half of that)
« Last Edit: March 30, 2014, 04:46:24 pm by fromhell » Logged

asking when OA3 will be done won't get OA3 done.
Progress of OA3 currently occurs behind closed doors alone

I do not provide technical support either.

new code development on github
Gig
In the year 3000
***

Cakes 48
Posts: 4199


WWW
« Reply #4 on: March 31, 2014, 12:57:41 am »

How big can the server location, and maintenance play a part? Can anyone comment? I want more knowledge in these sort of things.
You might want to take a look at the Quake 3 Unlagged documentation (most of its code is included in OA).
Then, is OA "delag hitscan" code really based upon "unlagged" q3 mod mode (and not just "inspired" by it)?

If unlagged code is GPL'd (that FAQ seems to do not mention licensing), what's about importing some of its side features, such as the (cheat-protected) cg_drawbbox variable (you can see some screenshots here) or g_trueping?
« Last Edit: April 01, 2014, 11:06:17 am by Gig » Logged

I never want to be aggressive, offensive or ironic with my posts. If you find something offending in my posts, read them again searching for a different mood there. If you still see something bad with them, please ask me infos. I can be wrong at times, but I never want to upset anyone.
Gig
In the year 3000
***

Cakes 48
Posts: 4199


WWW
« Reply #5 on: April 01, 2014, 01:46:00 am »

About Unlagged mod, I found a download of its source code here. But it doesn't seem to mention GPL anywhere... :-/
Logged

I never want to be aggressive, offensive or ironic with my posts. If you find something offending in my posts, read them again searching for a different mood there. If you still see something bad with them, please ask me infos. I can be wrong at times, but I never want to upset anyone.
grey matter
Member


Cakes 8
Posts: 381

>9k


« Reply #6 on: April 01, 2014, 11:02:00 am »

I don't know whether this is the original file, but OA's v0.8.8 gamecode on Github has a g_unlagged.c with GPL (v2+) header. An unlagged-2.01-src-gpl.zip exists elsewhere on the internet though.

From a quick look it seems like OA uses Unlagged mostly unmodified. g_truePing can be added serverside, cg_drawBBox requires some shader (haven't checked whether that one/the clientside code is already present in OA).
Logged

This space is for rent.
*Ruthless*
Half-Nub


Cakes 0
Posts: 81



« Reply #7 on: April 01, 2014, 06:37:05 pm »

I'm pretty sure that I have tweaks set properly. I use to play without any tweaks years ago, but a friend showed me the game cfgs do need to be tweaked. Before I've done tweaks I hadn't realized the large delays. Although when I try playing without a cfg now, it's pretty much unplayable hehe.

It's kind of hard to also play these days, since most servers are European. The locations of course vary depending on who runs the servers. But I usually have a playable, predictable time even in Europe. It could be many things why I feel theres a bit of a delay, also others have said they feel this as well sometimes. It's common between players who are from the United States, but I do hear it sometimes from even a European player.

What I may think sometimes, is that most of the servers aren't properly being ran. Maybe they could be missing some crucial settings in the server cfg. Does anyone believe this to be true as well? Especially when some servers feel totally different than others. This is left up to the server operator, but I've definitely noticed some servers have different movement options. As I said in my earlier post, my entire game is setup on precision, and timing. Sometimes landing in different spots from another servers settings, can play a role I believe as well.

I'm waiting for some players from the current or past OpenArena, to hop on this topic, and speak about this.
Logged
grey matter
Member


Cakes 8
Posts: 381

>9k


« Reply #8 on: April 02, 2014, 08:02:28 am »

I'm pretty sure that I have tweaks set properly. I use to play without any tweaks years ago, but a friend showed me the game cfgs do need to be tweaked. Before I've done tweaks I hadn't realized the large delays. Although when I try playing without a cfg now, it's pretty much unplayable hehe.
Oh, you certainly want to tweak fps (try some of the magic values, e.g. 125), maxpackets, packetdup, vsync etc. Those delays might be small, but they do add up.

What I may think sometimes, is that most of the servers aren't properly being ran. Maybe they could be missing some crucial settings in the server cfg. Does anyone believe this to be true as well? Especially when some servers feel totally different than others. This is left up to the server operator, but I've definitely noticed some servers have different movement options.
I think many server operators don't really pay attention to how well their server is running. Some are hosted on a crappy VPS with bad connection, slow disk and with shared gameserver hosting the providers often overbook server resources in general.

Regarding core server settings I don't think there's much to adjust. One might want to tweak max rate, max ping, fps and so on but the defaults should be fine (see Unlagged FAQ as well).

Regarding the movement, read the "Game physics" article that Gig linked to, there are several options for server operators to chose from and it's mostly a matter of taste.


To come back to your original question: it matters whether your hardware is fast enough to run the game (at e.g. a constant 125 fps) and with a proper connection (< 100 ms, little packetloss, jitter etc.). Above that, you don't gain that much of an advantage. Extremely high fps cause some bugs in Quake 3 movement and prediction. Having a LAN like connection/latency is not that important with Unlagged.
Logged

This space is for rent.
*Ruthless*
Half-Nub


Cakes 0
Posts: 81



« Reply #9 on: April 02, 2014, 05:34:30 pm »

grey matter
I've been wondering for quite sometime. How much does it matter if someone is running multiple servers on a VPS? Would that cause the servers to give a bad experience? Same question goes for a dedicated server. I know some people that run multiple servers on their one VPS/dedicated host. I suspect this could also be an issue?

I just had another game on the same server with another friend of mines, he noticed the same delay that I experience. Although it was a different map this time, we both felt the problem with our guns fire. Also, on my end I keep noticing splash damage/ direct damage from rockets and plasma take at least a few extra milliseconds to register I've just been smoked.
Logged
Gig
In the year 3000
***

Cakes 48
Posts: 4199


WWW
« Reply #10 on: April 03, 2014, 12:57:52 am »

Also, on my end I keep noticing splash damage/ direct damage from rockets and plasma take at least a few extra milliseconds to register I've just been smoked.
"Projectile" weapons (unlike "hitscan" -instant hit- weapons), are not "delagged"...
Logged

I never want to be aggressive, offensive or ironic with my posts. If you find something offending in my posts, read them again searching for a different mood there. If you still see something bad with them, please ask me infos. I can be wrong at times, but I never want to upset anyone.
*Ruthless*
Half-Nub


Cakes 0
Posts: 81



« Reply #11 on: April 03, 2014, 07:41:14 am »

But this is a sudden change. It's not usually this bad. I figure it's the servers though. Because this particular server has at least 10 others from the same host account.
Logged
grey matter
Member


Cakes 8
Posts: 381

>9k


« Reply #12 on: April 03, 2014, 01:23:46 pm »

I've been wondering for quite sometime. How much does it matter if someone is running multiple servers on a VPS? Would that cause the servers to give a bad experience? Same question goes for a dedicated server. I know some people that run multiple servers on their one VPS/dedicated host. I suspect this could also be an issue?
While I can't give you absolute numbers on this; the more unrelated processes you run on your server the higher your latency and jitter will/can become.
The hardware requirements for (io)q3ded however are quite moderate, given it's age (think Pentium III) and as such it should run well on any recent/decent dedicated server.
I don't quite recall whether one ioq3ded process needed 100% CPU before the introduction of com_busyWait, so this might be an issue.
Many people also forget that a VPS usually has limited I/O capacities as well. So if one process causes disk i/o (e.g. another q3ded instance writing logs, loading a map etc.), the whole VPS might become throttled or stall due to iowait. Like I've mentioned before, the VPS host itself might be running at capacity as well. Most VPS providers don't have the network required for low latency applications such as games/VoIP but are rather geared towards webservers with their bulk traffic.

On a sidenote I got the feeling that some players with bad connection slow down the entire server. I have no evidence for this, but general latency etc. usually return to regular levels when they leave. This might be caused by a high cl_maxPackets setting on those clients.

tl;dr: Multiple gameserver instances on a dedicated server shouldn't be a problem, but VPSs are often unsuited in general.
Logged

This space is for rent.
*Ruthless*
Half-Nub


Cakes 0
Posts: 81



« Reply #13 on: April 04, 2014, 08:12:28 am »

Thanks. I kind of figured thats what your answer would be. I feel the same way about some clients joining the servers, and causing somewhat of a disruption. I as well have no proof of this, just a speculation over the years of playing.
Logged
Pages: [1]
  Print  
 
Jump to: