Question for everyone

[Batman]

Ok I goto Fragfest, me and Gond are sitting next to each other. We both turn our graphics cards to quality in windows, then we go into CS and set everything on High. We played all weekend with those settings and it was very smooth and very pretty.. Then I come home, and I dont make any changes to my settings and I go into a GC CS server and it ran VERY bad, so then I try a smaller server and its still VERY bad like 30-40 fps.. All weekend at FF I was gettin 80-99 fps on the lan..

 

So here is my question: If my game ran great at FF over a lan, and the packet loss and choke make my game run crappy, HOW is a faster CPU/MB gonna make my game better when its obviously the internet connection making my game bad??

 

I asked Gunman this question lastnight and he said "get a better NIC, like the gigabyte one", and I said "thats the NIC im using!!!"

[NOFX]

um... a better NIC isn't really going to make your cable modem any faster

 

about your question. I don't know man, maybe if you had a beast video card you would be getting 200+ fps at the LAN ?

[Preacher]

No ammount of hardware boostage will make your internet speed faster, but the processing power of a good video card and the capabilities of a higher end CPU will make the very most of what connection you do have.

 

The LAN has 100mb/s or 1000mb/s (actual 50mb/s or 500mb/s) where your ave DSL/cable is 1.5mb/s peak and an actual 500kb/s. So the better the connection you have via ISP the faster the info can get to you, but no matter how fast it gets to you your hardware will make a large difference in how fast it's APPLIED

[JackieChan]

Sorry batty, I stole your video card and gave you my bad one..... magically

[Norguard]

Another one learns the sad truth about the horrors of teh intarwebnets.

 

I complain about choke giving me 20 fps, and people tell me that my x800 is defunct. The sad truth is that it does have an impact with Source. Take solace that most games aren't that bad. It isn't the fault of Valve's Source engine... ...exactly. It's the internet's problem, here.

 

There's too much stuff to be calculated between PCs in the Source engine. If we were playing Half-Life, you'd still be getting 200+ fps if you were on lite-cable, with the same number of players. Same is true of Q3A or, for the most part, UT2k4.

 

There's a lot that goes on in Source, and it all has to be sent and verified and recieved and verified again, and the servers regulate the playing field, to make sure that nobody's at too much of a disadvantage - so that somebody on a 1.2 with a GeForce 4MX is on as level a playing field as somebody with an X2 4800 with two GeForce 7800s as possible. If everybody playing has a great system, and a decent connection, there will be little or no slowdown... Or on the flipside, a decent system and a great connection. Servers can calculate, say 300fps. Your PC requests keyframes, which are exact, regardless of which PC you're on, and then interpolates the rest. The more it has to interpolate, the harder it has to work, and the more jumps there are between keyframes and everything else... Is that Valve's fault? I'd say it's Valve's fault no more than it's Blizzards fault if your character jitters in World of WarCraft, for trying to make it as compatible as possible. In all honesty, we are at the point where some people who play have a PC more than 5 years old, and others have a PC that's less than 5 months old, trying to compete on the same field...

 

...your PC may actually be too good right now... ...at least, when certain people, or certain groups of people are playing.

 

...if it really gets to you, switch to a less intensive game for a bit. Anything but the latest MMO (Or DooM 3). If you are having legitimate choke problems with CS1.6 or Q3A, consider a new ISP. If you're not, play some dreaded single-player for a bit. The sad truth is, as much as people bag Valve's netcode, it's pretty sweet for what it does, and it's not going to get any better until the gaming community possesses systems which are even in the same ballpark with one another.

 

When we can all play CS:S at 100+ fps at all times, I'm betting your choke, or worse - loss, will be all but a thing of the past.

Edited July 26, 2005 by Norguard

[NOFX]

i think we should pentalize the ones who have a 1.2Ghz and a Geforce4MX. They shouldnt write the code to make up for it.

[bear^]

great post norguard!

 

i feel learned now

 

[Norguard]

NOFX, would you think it fair to say that the minimum requirements for playing Half-Life 2 or any Half-Life 2 mods are an AMD64 3000 or greater, 1GB PC3200 RAM, dual channel, a SATA drive and an x800 or better videocard? All of that hardware, save for the x800 is already 2+ years old, so everybody should have it installed.

 

All lesser systems I've played on have dropped significantly when everybody in a 24 man server rushes DD in Aztec. I've gone from 60fps to less than 10 in those firefights on machines that I would have no hesitation playing Unreal II, or even BattleField 2 on. Those machines are liabilities that you're not willing to keep around?

 

Those machines are machines like an AMD XP2500+ with 1GB PC2700, dual channel and a 9600XT videocard. On the Intel/nVidia side, that may as well be a P4 2.8 with 1GB PC4000 RAM and an GeForce 5700.

 

If you told those people they weren't allowed to play on your server until they spent $1000+ to upgrade, they wouldn't be too happy with you. If those people bought the game, and installed it and got a pop-up from VALVe saying "Whoops! Go back to StarCraft, kid. You'll need to buy a new PC for this game.", when they probably bought that stuff just a year or two ago, and gaming is just a hobby to them... ...there'd be heads rolling.

 

PC fragging has never, ever been even. I remember spanking people in WinQuake. I was one of the few and proud with an 8MB 2D card, hooked up to not one, but two 12MB VooDoo II cards in SLI. I also had a Diamond 112k modem. Two phone lines dialed into the same ISP, and their bandwidth combined. Was it fair that I was playing at half the ping of everybody else, at twice the framerate, at 1024x768 when they were playing at 320x240, or 640x480? No, probably not... ...but then they caught up with RivaTNT2s and VooDoo3s and GeForces and DSL.

 

If all were to be even in the realm of online gaming, we may as well all be given trendy-coloured iMacs (because everybody knows it's what's on the outside that counts) and have every game coded for the 32MB on-board Bondi iMac videocard. No graphics options - everything from CS:S to DooM 3 to Elder Scrolls 4 must look like a mixture of CS1.5 and grey cardboard smeared with oatmeal.

Edited July 26, 2005 by Norguard

[NOFX]

well..... a 2500+ and a 1.2Ghz are alot different, so is a GF4 MX and a 9600XT.

 

ahh those where the days, when you had a beast machine, you had a clear advantage and I loved it. I used to play UO and when Ihad a T1 and everyone else was on dial up, they didnt stand a chance..

 

But you also have to look at it this way, you are being penalized for having a beast machine. What is the use in owning one if you can't get the performance its capable of?

 

The code should be written for the better graphics card, not the crappy one and I think people should be able to play on any machine they like, regardless what it is, but it should be their machine that runs crappy and the code should not be written to try and level the playing field by making my machine lose performance.

 

Those machines are machines like an AMD XP2500+ with 1GB PC2700, dual channel and a 9600XT videocard.  On the Intel/nVidia side, that may as well be a P4 2.8 with 1GB PC4000 RAM and an GeForce 5700.

 

not tryin to start an intel/AMD fanboy flame thread, but about three years ago I decided to upgrade from my 1800+

 

I first bought a 2500+ to use with my 9500 pro, which is actually faster than the 9600 pro, but not as good as the XT. Well, I OC'd the 2500+ to a 3200+ and it sucked, not much performance difference, so I returned it and got a P4 2.6 OC'd to 3.0Ghz, which runs significantly better and scored better on every benchmark I had including 3dMark03.

 

Is my 2.6Ghz really that slow of a machine to where it makes ppl lag? I have always noticed I am still usually the fastest person in any given server with a low ping.

Edited July 26, 2005 by NOFX

[NOFX]

after looking at my score I will take that back, 3dmark03 had jsut a tad higher score, by like a hundred maybe. But PC mark blew the AMD away.

 

However, looking at my scores, it all depends on the video card. I had my 1800+ and then 9500 OC'd to around 312/312 and it scored wayyy better than it did stock on my P4

[Preacher]

That's because teh AMD rulez Joo

[appalachian_fox]

You must have been in the same server I was running my 486 DX-50 in.

[Cujo]

after looking at my score I will take that back, 3dmark03 had jsut a tad higher score, by like a hundred maybe.  But PC mark blew the AMD away. 

 

However, looking at my scores, it all depends on the video card.  I had my 1800+ and then 9500 OC'd to around 312/312 and it scored wayyy better than it did stock on my P4

<{POST_SNAPBACK}>

forget synthetic benchmarks. i'd take a xp3200 over a p4 3.0 any day of the week, especially for source.

 

 

 

as for the topic. i have a relatively high end machine and i hate being penalized for it. since the last update i now get spikes upwards of 10 loss and at least 10-20 choke. on our match server doing a 5v5 i get max of 8 choke on round start and no loss. so it seems to a server and a client issue.

 

i think the main problem is that ppl run their gfx settings too high. who cares what the game looks like. you should set it up so you get the best possible fps. if everyone did that then the servers would be a lot smoother.

 

i run at 1280x960 but all gfx are turned low except shadows. on our match server my fps rarely drops below 60. if everyone set it up so that on a good server their fps never dropped below 60 then we'd have A LOT less lag and everyone would have much smoother game play.

 

i wish we could set up a test server with 20 ppl on and everyone set their settings to the highest and then everyone change to the lowest. i believe choke and loss would change dramatically and all would have a better gaming experience.

Edited July 26, 2005 by *|CsLs|*Cujo

[NOFX]

i think the shadows are the biggest enemy to video cards, u would probably get better performance with everything up and your shadows down.

 

Well see the thing is, most people here base their processor only on counter-strike when you say which is better. An intel with hyper-threading vs a 3200+ would dominate when encoding a DVD. Or when doing some multi-tasking.

Edited July 27, 2005 by NOFX

[Norguard]

True dat, yo.

 

...okay, I'll be the first to admit it... ...I'm white. Totally, and undeniably caucasian.

 

Some people don't have the luxury, though, Cuj. You know as well as I that if you've got a 9800XT and an XP2400+, the better upgrade is to get a midrange Athlon64 CPU... ...at least for Source. There is a huge fps boost once your processing power reaches a certain level (I'd say it's around the 3GHz barrier). Physics love you and so does interpolation. The reason I added an x800 on top was so that I could run the graphics up and keep a 60fps+ overhead. I totally agree that those findings would be just like you say, but I think a lot of it is old CPUs and not necessarily GPUs in all circumstances.

 

At least as far as Source goes. Put it this way: when DooM 3 came out, there were no cards that could run it full tilt at more than 30fps. Now most of the cards can do it and push 100 if you take the frame limit off. I'm assuming the exact same thing for Unreal Tournament 2k7. Problem with Half-Life 2 is that it's way more CPU intensive than either will be, and people are really only interested in GPU upgrades. I'm looking forward to seeing if this is one of the engines that's going to be updated to support physics acceleration.

 

As for the shadows, they're just simple shaders. Nothing huge and volumetric... ...that's actually why they're so useful. You can see them all over the place. It's a low performance cost if you've got PS2.0 support, and it gives you a huge advantage. Even if you set your shader detail to minimum, which will net you the biggest boost, outside of using no AA and using BiLinear instead of Anisotropic, keeping the shadows lets you play smart and get the proverbial leg up.

 

And I'll admit that a P4 with HT is enough to blow an AthlonXP out of the water when using Adobe filters in Photoshop or Premiere...

 

...but Socket939 (and later Socket754) AMD64 chips, Venice core and later, all have the exact same SSE instruction sets as every P4 chip, putting a huge dent in the P4 lead in those areas.

Edited July 27, 2005 by Norguard

[Laz.e.rus]

OK, heres a question: How much differnece does it make changing your virtual mem? I have always had to det to initial of 2700MB, max of 4000MB. I recently changed it to 600/600 to get an old game to play ( dont know why it worked, but the forums said do it and it worked).

I jsut changed them back. Does this/ would this affect gameplay in any way with diff settings? How so? How Much?

 

Thanks!

[NOFX]

OK, heres a question: How much differnece does it make changing your virtual mem? I have always had to det to initial of 2700MB, max of 4000MB. I recently changed it to 600/600 to get an old game to play ( dont know why it worked, but the forums said do it and it worked).

I jsut changed them back. Does this/ would this affect gameplay in any way with diff settings? How so? How Much?

 

Thanks!

<{POST_SNAPBACK}>

 

well it will depend on how much RAM you have. If you have 2 GB of ram installed, your virtual memory will hardly be used. Put it this way, when you run out of RAM, virtual memory means your hard drive will take over the responsibilities of your memory(which means its way slower)

[*BiGBonES**]

I used to play cs 1.0 with a ping of 400 and got on just fine

[appalachian_fox]

As far as the virtual memory goes, I've actually had a good bit of luck setting both the upper and lower to the same value and then defragging my hard disk. This means the virtual memory "disk" will not grow or shrink, and by defragging it you put the whole virtual memory block in one place, avoiding unecessary seeks. Fragmented virtual memory files have appeared to be the culprit for many a memory-intensive processes we've run at work, and ever since I've fixed the virtual memory size and defragemented on all the new machines here I've noticed the really hard-core memory munchers don't crash nearly as often.

 

How well that will affect your gaming...I don't know. Again, like NoFX said, it depends on how much RAM you already have. If you don't touch your virtual memory, it won't help at all. If you are into your virtual memory, it may add a few FPS, but it'll probably still be darn near unplayable anyway. That's all speculation.

 

Couldn't hurt, though.

[NOFX]

Here is my previous experience with games and virtual memory.

 

When I first started playing desert combat, I had a 1800+ with 768 megs of RAM and a 9500 pro. Well, I upgraded to a 3.0Ghz with 512 Megs. It sucked. I was running decent, but it would skip allll over the place, especially in firefights and heavy action. I'm assuming this was because it was access my virtual memory. I install another 512 and it couldn't have ran smoother.

[Laz.e.rus]

rgr. So, since I have a gig, Im prolly not really even using it. correct? That old game wouldnt even run unless I set min and high limits to the same size, then it loaded right up. Thts what got me thinking of how much my games may depend on it and what diff it would make leaving it at 600/600 vs my reg limits of 2700/4000.

[Norguard]

A lot of old games will actually check your VMem before loading...

 

...StarCraft, for instance, won't load if you disable VMem. Not that it could possibly take up enough RAM to exhaust a GB or more... ...it just refuses to run.

 

There will likely be times in Source or DooM 3, WoW or BF2 where 1GB RAM just isn't enough. Then it pours into VMem. The performance of VMem is crap compared to RAM. Spinning disks will never perform on par with solid-state. RAM can move 1000s of MB a second, whereas HDDs can only move 100s of MBs... It's when both your RAM and your VMem are full that you get the chunk. The proverbial ice shard at the bottom of the straw, that's holding everything up, while you try your hardest to suck it all through the pipeline and keep it going.

 

Don't worry a whole lot about it, though. Half a gig should be plenty and a gig should probably be tops with what you've got. If any more is needed, I think it's safe to assume there's a leak.

 

I actually quite like Fox's idea as far as handling it, and keeping it clean, as well...

 

...the next time I have a hard drive, I'm going to play by those rules, too.

Edited July 28, 2005 by Norguard

[Batman]

After doing alot of settings tests, no matter what settings I have on my system GC servers if there are more than 24 people in the server are always 20-30 fps lower than other servers, like DOH's server..

 

So when I came home the first server I went into was stock and there were 27 people in the game, so it ran very choppy...

When I have everything turned up in CS and go into DOH's server its much smoother than GC servers, its very strange.. But doh's server is 24 max people I think.. So my problem is: I must have a limit to smooth gameplay / # of peeps in a server..

 

**Note this is in no way a smear on GC servers, its strictly in the matter of science..** Its science darn it!!!

[appalachian_fox]

NoFX, I'll take your anecdotal evidence as fact supporting my theory

 

Anybody else have experience in this area?

[Wolfsblood]

I may have had an experience with something similar. My wife is using my old 2800+ with a 9600xt to play SWG. The computer had 512 RAM. She said she was getting some frame rate drops in heavy firefights, and when she was in a large hunting party. I dropped an additional 512 meg in, and since then she has had no more issues with framerates dropping.

 

Just my personal experience.