Some insight on FPS issue?
I’m not really sure what the problem with this engine is .
Desktop
Amd Phenom II X4 955 Black Edition 3.2GHz
4GB DDR3
NVidia Geforce GTX 260 216OC
Resolutions
1280×720 Windowed
1600×900 Fullscreen
During the first BWEs I was getting 20-40 FPS in most places and 10-19 FPS in big zergs.
During the last couple BWEs after optimization took place this system was totally unplayable. 10-19 FPS all the time no matter what resolution or settings.
Laptop
I7 2670QM 2.2GHz (4 Core 2.7 Turbo Boost, 3.1 2 Core Turbo Boost, 3.2 1 Core Turbo Boost)
12 GB DDR3
Nvidia Geforce GTX 570M 3GB DDR5
Resolutions
1280×720 Windowed
1600×900 Windowed
1920×1080 Fullscreen
All resolutions net the same FPS pretty much. 40-100 FPS PVE, 20-100 FPS WvWvW. Max FPS when no one is around and 20 FPS when people are near in WvWvW.
I thought the reason would be obvious. Your comparing an i7 Quad to a 6yr old C2D.
It works like this:
bad cpu / good graph card = bad FPS
STRONG CPU / bad graph card = GOOD FPS
The idea is to get as close to 99% cpu/gpu usage all of the time, but obviously that will never happen because one or the other is going to choke, it always does. There is never a perfect match. Using a Core2Duo with a GTX680 will get identical or sometimes worse FPS than an old 8800GT. Because the Gpu is ALWAYS waiting for instructions from the processor.
If you eliminate the processor bottleneck eg: i7 3960x, then you’ll ALWAYS see close to 99-100% gpu usage because the gpu is not waiting for instructions from the processor. Which in turn = more FPS.
GTX 6GB Titan@1160Mhz
3007WFP@2560x16000
(edited by Cbuzz.5083)
So i’d be good idea to leave that GTX 260 and get Bran new CPU.
While looking on price let’s say something like
AMD PHENOM II X4 965
AMD X6 FX-6200
AMD X4 FX-4170?
They seem to be fine on par with i5 ( a bit slower ) tho a difference in a price by 100% is noticable
So i’d be good idea to leave that GTX 260 and get Bran new CPU.
While looking on price let’s say something like
AMD PHENOM II X4 965
AMD X6 FX-6200
AMD X4 FX-4170?They seem to be fine on par with i5 ( a bit slower ) tho a difference in a price by 100% is noticable
While you will see strong FPS increase with a new processor, that Gpu is still old. Also do not get an Amd processor (see GW2 & Amd thread), they are not on par with i5/7, not even close.
GTX 6GB Titan@1160Mhz
3007WFP@2560x16000
The idea is to get as close to 99% cpu/gpu usage all of the time, but obviously that will never happen because one or the other is going to choke, it always does. There is never a perfect match. Using a Core2Duo with a GTX680 will get identical or sometimes worse FPS than an old 8800GT. Because the Gpu is ALWAYS waiting for instructions from the processor.
If you eliminate the processor bottleneck eg: i7 3960x, then you’ll ALWAYS see close to 99-100% gpu usage because the gpu is not waiting for instructions from the processor. Which in turn = more FPS.
There are some flaws here.
1. Many people are reporting no where NEAR 99% load on their CPU, but still its “bottlenecking”. Myself, my CPU load never goes higher than 70%, and often, my GPU is at 0%. This isnt a true hardware bottleneck, its a “soft cap” set by whatever the developers coded the CPU to do. Unused CPU resources go to waste, resulting in bad performance.
2. Gpu always waiting on CPU is realy dependent on how the programmers coded the game. For instance, Alien vs Predator 2, has a VERY low cpu overhead, maybe netting around 11% CPU load, but your GPU will be constantly working because thats how they designed the game. A game like GW2 (still based on Dx9, which is YEARS old), hardly pushes off ANYTHING to the GPU.
Bad design decision from a developer standpoint imo. If i had to guess, id say its related to the Umbra tech they inherited. Umbra allows for a special type of 3d object occlusion that is supposed to help the game perform better. This is for things like 3d buildings, trees, hills, and other static world objects. I have a feeling this tech is also being used for player models as well. Using this tech with player models adds a huge new element of dynamic calculations that have to be made before the Umbra tech can even start working. And this isnt even considering the fact that in large battles, there are more player models to keep track of than say, buildings in a city that you can see on your screen at any one time.
The devs have their hands full regardless, trying to figure out how to optimize this, especially 2 of the major parts of the games hyped, Dragon/Big Boss fights, and WvWvW, are the 2 things most affected by these performance issues.
The idea is to get as close to 99% cpu/gpu usage all of the time, but obviously that will never happen because one or the other is going to choke, it always does. There is never a perfect match. Using a Core2Duo with a GTX680 will get identical or sometimes worse FPS than an old 8800GT. Because the Gpu is ALWAYS waiting for instructions from the processor.
If you eliminate the processor bottleneck eg: i7 3960x, then you’ll ALWAYS see close to 99-100% gpu usage because the gpu is not waiting for instructions from the processor. Which in turn = more FPS.
There are some flaws here.
1. Many people are reporting no where NEAR 99% load on their CPU, but still its “bottlenecking”. Myself, my CPU load never goes higher than 70%, and often, my GPU is at 0%. This isnt a true hardware bottleneck, its a “soft cap” set by whatever the developers coded the CPU to do. Unused CPU resources go to waste, resulting in bad performance.
2. Gpu always waiting on CPU is realy dependent on how the programmers coded the game. For instance, Alien vs Predator 2, has a VERY low cpu overhead, maybe netting around 11% CPU load, but your GPU will be constantly working because thats how they designed the game. A game like GW2 (still based on Dx9, which is YEARS old), hardly pushes off ANYTHING to the GPU.
Bad design decision from a developer standpoint imo. If i had to guess, id say its related to the Umbra tech they inherited. Umbra allows for a special type of 3d object occlusion that is supposed to help the game perform better. This is for things like 3d buildings, trees, hills, and other static world objects. I have a feeling this tech is also being used for player models as well. Using this tech with player models adds a huge new element of dynamic calculations that have to be made before the Umbra tech can even start working. And this isnt even considering the fact that in large battles, there are more player models to keep track of than say, buildings in a city that you can see on your screen at any one time.
The devs have their hands full regardless, trying to figure out how to optimize this, especially 2 of the major parts of the games hyped, Dragon/Big Boss fights, and WvWvW, are the 2 things most affected by these performance issues.
You don’t need to try & convince me otherwise. I’ve been doing this forever & I know how it works. I’ve got 6 PC systems here which I play with/test daily. ALL of which conform to what I’ve stated above to the TEA.
GTX 6GB Titan@1160Mhz
3007WFP@2560x16000
When your GPU is able to do FP32 math more efficiently than the CPU and there’s a unified way to do so for AMD/nvidia, let me know.
Otherwise, http://i.imgur.com/uiBK8.jpg. No game can use 4 renderers to split up the load, and there’s no escaping concurrency.
MMO’s are dependant on your CPU mainly.
I used ot run an old core 2 extreme (QX6850) ..which is a quad core 3ghz processor. I OC’d to 3.7ghz. but even with that mmo’s would have lower FPS than you would expect. My graphcis card would be sitting at 50% usage ..my cpu at 70%.
Wheres the problem you say?
Well. The CPU was holding my GPU back ..not becouse it was maxed out in load, but it was maxed out in the speed in which it could handle the load it was given. Bassicaly the archecture of the CPU could not handle the information it was being given efficently enough to pass it onto the GPU. So from experiance i can say ..you wont see your old C2D or C2E reach 99% usgae ..but it will still be slowing you down.
What you DO want to see is your CPU below say 50% usage and your GPU maxed at 99% asuming you have vertical sync OFF. if you have VSyn ON you may not see your GPU reach 99% if it comes across an area where it could make 100 fps but your screen is set to show a max of 60 fps via vsync.
I for example have a intel i7 3930k OC to 4.6ghz ..totaly overkill for any game and mmo out atm it barely reaches 15% usage.. however my GPU is a 560ti ..deffinatly NOT overkill. With Vsync on im maxed at 60 fps. I get 60 fps most of the time but it will dip down to the 30’s and 40’s when it comes across somthing really intensive. And when that happens my GPU usage is at 99%
You dont need to go overbored like i have. But you should try get a i5 or i7 ..over 3ghz as standard (not including boost). That way your CPU wont be the bottleneck.
…Well i say wont be ..it wont be until you get into seriose Graphics settings and GPU power.. at which point in a mmo ..you will still see advantages all the way up to 4ghz (which is OC’ing territory)
EVGA GTX 780 Classified w/ EK block | XSPC D5 Photon 270 Res/Pump | NexXxos Monsta 240 Rad
CM Storm Stryker case | Seasonic 1000W PSU | Asux Xonar D2X & Logitech Z5500 Sound system |