Rig#2: Core2duo@3Ghz/ 4GB DDR2/ 9800gtx+
GTX580 OCed
Rig#2: Core2duo@3Ghz/ 4GB DDR2/ 9800gtx+
personal experience with my GTX 580:
i get around 10FPS more when overclocking to 888Mhz
also OC’ed my memory clock to 1100/2200/4400Mhz (GPU-Z would show 1100MHz and MSI Afterburner would show 2200MHz)
thats a summed up performanca boost of about 15% in-game
of course, a customized fanspeed is neccessary
also, make sure you GPU gets enough power, or you might see random game or driver crashes or shutdowns
and don’t set rendersmapling in-game to “supersample”
a GTX580 just can’t handle it and the quality difference is absolutely minimal – not worth it, to make it work 4 times harder
i also got 10-15FPS more by messing around at the nvidia-cp:
make sure "Antialiasing-Transparency is set to “off”
“Power management mode” to “prefer max performance”
“multi-display/mixed GPU acceleration” to “single display performance mode”
“texture filtering – quality” to “performance”
and “ambient occlusion” to “off”
if you now disable/lower reflections in-game and set rendersampling to “native”, you won’t need any GPU overclocking anymore
a GTX580 isn’t meant to be overclocked in the first place…
“Only the finest of potatoes in my zerkburgers.”
With GW2 90% of cases end up being the CPU limiting performance.
Try OC ur cpu and see what fps gains you get.
I run a 560ti factory OC’d to 950 on the core and 4580 effective on the memory. I run 1920×1080 max settings (except supersampling and somtimes reflections are set to down to terrain and sky) That gives me usualy 45 to 60 fps depending on where i am and whats goign on.
however the main performance giver is my CPU which is a OC’d 3930k at 4.6ghz . if it was stock i wouldnt be getting such high framerates (by gw2 standards) in game.
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 |
i’m seriously about to start thinking, that my CPU’s (i7 950) displayed clock-speed is lying to me
i mean @3,1GHz = 46-48FPS in the middle of divinity’s reach
and @3,6GHz = 46-48FPS in the middle of divinity’s reach
and @4,1GHz = 46-48FPS in the middle of divinity’s reach
the same applies to any other area, where i’ve tested it (some PvE areas, some cities)
the only difference: temps are higher at prime torture tests…obviously
but they remain at 55°C while playing GW2, no matter at which clock speed my CPU is running
i must do something wrong if OC’ing my GPU boosting my framerate, but OC’ing my CPU does not while everybody out there telling you how much CPU dependant GW2 is
“Only the finest of potatoes in my zerkburgers.”
personal experience with my GTX 580:
i get around 10FPS more when overclocking to 888Mhz
also OC’ed my memory clock to 1100/2200/4400Mhz (GPU-Z would show 1100MHz and MSI Afterburner would show 2200MHz)
thats a summed up performanca boost of about 15% in-game
of course, a customized fanspeed is neccessary
also, make sure you GPU gets enough power, or you might see random game or driver crashes or shutdownsand don’t set rendersmapling in-game to “supersample”
a GTX580 just can’t handle it and the quality difference is absolutely minimal – not worth it, to make it work 4 times harderi also got 10-15FPS more by messing around at the nvidia-cp:
make sure "Antialiasing-Transparency is set to “off”
“Power management mode” to “prefer max performance”
“multi-display/mixed GPU acceleration” to “single display performance mode”
“texture filtering – quality” to “performance”
and “ambient occlusion” to “off”if you now disable/lower reflections in-game and set rendersampling to “native”, you won’t need any GPU overclocking anymore
a GTX580 isn’t meant to be overclocked in the first place…
im using Asus overclocking utility, which stands true as it shows up core speed and memory in GPU-Z.
core speed 930, 1.1mV, 4400 Memory.
Ofcourse custom set fanspeed.
Sampling to Native. FXAA enabled, it has no noticable FPS hit, and its also quite ugly to be honest – but fast.
As i said i see no difference at all. The thing is while in the zerg fights in WvW i drop to 15fps, i get alot more when i lower all details to minimum (perfomance) in GW2 options.
Can i presume that GPU bottlenecks then?
Rig#2: Core2duo@3Ghz/ 4GB DDR2/ 9800gtx+
i’ll stick to BIOS OC…that asus-oc tool isn’t accurate…it always freezes my PC when doing even the slightest voltage changes
i doubt a GPU bottleneck…
i barely see any difference between native and supersampling @1920x1080p (max resolution of my screen)
the usage will go from 60-70% (temps at 69-76°C) to 95-100% (temps at 80-81°C)
so the minimal increase of quality isn’t worth it for me
you will see the most performance boost on native while being on a heavy loading area with lots of water/reflections (that are the spots where my GPU goes nuts on supersampling, even when turning reflections to “off” @stock clock settings)
nice clocks…i can’t get it to work stable with anything higher than 900MHz and 4200Mhz memory clock
“Only the finest of potatoes in my zerkburgers.”
personal experience with my GTX 580:
i get around 10FPS more when overclocking to 888Mhz
also OC’ed my memory clock to 1100/2200/4400Mhz (GPU-Z would show 1100MHz and MSI Afterburner would show 2200MHz)
thats a summed up performanca boost of about 15% in-game
of course, a customized fanspeed is neccessary
also, make sure you GPU gets enough power, or you might see random game or driver crashes or shutdownsand don’t set rendersmapling in-game to “supersample”
a GTX580 just can’t handle it and the quality difference is absolutely minimal – not worth it, to make it work 4 times harderi also got 10-15FPS more by messing around at the nvidia-cp:
make sure "Antialiasing-Transparency is set to “off”
“Power management mode” to “prefer max performance”
“multi-display/mixed GPU acceleration” to “single display performance mode”
“texture filtering – quality” to “performance”
and “ambient occlusion” to “off”if you now disable/lower reflections in-game and set rendersampling to “native”, you won’t need any GPU overclocking anymore
a GTX580 isn’t meant to be overclocked in the first place…
im using Asus overclocking utility, which stands true as it shows up core speed and memory in GPU-Z.
core speed 930, 1.1mV, 4400 Memory.
Ofcourse custom set fanspeed.
Sampling to Native. FXAA enabled, it has no noticable FPS hit, and its also quite ugly to be honest – but fast.
As i said i see no difference at all. The thing is while in the zerg fights in WvW i drop to 15fps, i get alot more when i lower all details to minimum (perfomance) in GW2 options.
Can i presume that GPU bottlenecks then?
If you get more FPS in the demanding area with lower settings then yes your GPU is the problem or driver. I have a lower GPU and have no problems at all around 35FPS is the big zergs and still use a driver form 11/12/11 AMD
i5 2500k 4.4 6970 950/1425
16GB 2133 ram
SSD
(edited by zerk.9701)
personal experience with my GTX 580:
i get around 10FPS more when overclocking to 888Mhz
also OC’ed my memory clock to 1100/2200/4400Mhz (GPU-Z would show 1100MHz and MSI Afterburner would show 2200MHz)
thats a summed up performanca boost of about 15% in-game
of course, a customized fanspeed is neccessary
also, make sure you GPU gets enough power, or you might see random game or driver crashes or shutdownsand don’t set rendersmapling in-game to “supersample”
a GTX580 just can’t handle it and the quality difference is absolutely minimal – not worth it, to make it work 4 times harderi also got 10-15FPS more by messing around at the nvidia-cp:
make sure "Antialiasing-Transparency is set to “off”
“Power management mode” to “prefer max performance”
“multi-display/mixed GPU acceleration” to “single display performance mode”
“texture filtering – quality” to “performance”
and “ambient occlusion” to “off”if you now disable/lower reflections in-game and set rendersampling to “native”, you won’t need any GPU overclocking anymore
a GTX580 isn’t meant to be overclocked in the first place…
im using Asus overclocking utility, which stands true as it shows up core speed and memory in GPU-Z.
core speed 930, 1.1mV, 4400 Memory.
Ofcourse custom set fanspeed.
Sampling to Native. FXAA enabled, it has no noticable FPS hit, and its also quite ugly to be honest – but fast.
As i said i see no difference at all. The thing is while in the zerg fights in WvW i drop to 15fps, i get alot more when i lower all details to minimum (perfomance) in GW2 options.
Can i presume that GPU bottlenecks then?
If you get more FPS in the demanding area with lower settings then yes your GPU is the problem or driver. I have a lower GPU and have no problems at all around 35FPS is the big zergs and still use a driver form 11/12/11 AMD
i5 2500k 4.4 6970 950/1425
16GB 2133 ram
SSD
Seems so. I could boost my CPU higher to maybe 4.3 or 4.5 , but my Tower totally sux and i have poor vendilation.
Im also using a 212TX aftermarket cooler but it screams for an additional fan, on the other side of the heatsink.
For all these reasons of poor heat management, i prefer not to go any higher in the clocks of the CPU.
On the other hand the GPU remains largely inaffected, whether i run on stock speed, around 780 or 930. Havent managed to procure a stable OC at 950 with 1.1mv though and i dont wanna push any higher.
The video card has a very nice internal mechanism that reverse to stock speed if for some reason the OC fails (or is it the Asus Tweek program?). The screen goes black for a moment and the videocard resets to stock settings, and you carry on gaming :>
Rig#2: Core2duo@3Ghz/ 4GB DDR2/ 9800gtx+
i’ll stick to BIOS OC…that asus-oc tool isn’t accurate…it always freezes my PC when doing even the slightest voltage changes
But if it wasnt accurate, wouldnt the value that shows in the Asus Tweak be different with the one in GPU-Z?
Rig#2: Core2duo@3Ghz/ 4GB DDR2/ 9800gtx+
don’t set rendersmapling in-game to “supersample” a GTX580 just can’t handle it and the quality difference is absolutely minimal – not worth it, to make it work 4 times harder
This is not true, I have the GTX 480 and i run with vsync enabled 60fps in PVE with rendering set to supersample. it does make a difference, I have a very smooth and crisp looking game now.
Settings like Reflections and Shadows are best on their lowest setting or disabled for maximum performance, also Shaders will put a bigger hit on the GPU that supersampling will.
512GB Samsung 950pro | 2TB SATA-3 | AMD Radeon RX 580 8GB
don’t set rendersmapling in-game to “supersample” a GTX580 just can’t handle it and the quality difference is absolutely minimal – not worth it, to make it work 4 times harder
This is not true, I have the GTX 480 and i run with vsync enabled 60fps in PVE with rendering set to supersample. it does make a difference, I have a very smooth and crisp looking game now.
Settings like Reflections and Shadows are best on their lowest setting or disabled for maximum performance, also Shaders will put a bigger hit on the GPU that supersampling will.
doesn’t your GPU catch fire or something like that? mine does on supersampling T-T
i never should’ve compared some GPU-Z logs, one on native and one on supersampling
40% usage vs 100% usage
71°C vs 81°C
not to mention that my framerate drops by 10FPS on supersample and will go nuts if anything like water is in sight (reflections @ “none”)
BTW…you’re getting like ~10FPS more than i do in PvE…i never liked logic anyways
“Only the finest of potatoes in my zerkburgers.”
(edited by wauwi.9162)
Wauwi.
You should have your GPU fan speed set to keep your GPU cool even at 100% usuage.
GPU’s are designed to be used at 100%. So heat shouldnt be a problem. If your uncomfortable with 81c, then use EVGA precision and setup a custom fan speed curve.
I have not OC’d my 560ti more than what it was as factory (950core 4580effecitve memory) but i did have to change the fan speed curve to my own one to keep the temperature from going above 75c. When temps hit 70c my fan bumps up to 70% and when it hits 75c the fan goes to 80%, that stops it from going any higher. I can run it all day at 100% load and it wont go above 75/76c.
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 |
i should have said that my fan speed is customized…it runs at ~80% when at 75°C and ~90% if at 80°C
i clean out the dust regularly
so my fan speed is always at ~90% if OC’ed + supersample and it still won’t reduce the temps
let’s just hope that santa will bring me a fresh new Gigabyte GTX650-TI twinforce
it’s half as cheap as a GTX580 and about 20% more powerful…theoretically
so i can use a sledgehammer on my GTX580
had such nice framerates two months ago…no need for OC’ing anything or tweaking anything…60FPS everywhere T_T
it sounds so unrealistic now…
“Only the finest of potatoes in my zerkburgers.”
hmm 650 vs 580 ..i honestly dont ‘think’ thats an upgrade. Ill look into it more but im pritty sure thats a downgrade.
EDIT: yea a 650 will be a downgrade. The only way it beats the 580 is in the wattage is uses.
This isan overal comparison. use your browsers search engine to select the 650 and the 580
http://www.videocardbenchmark.net/high_end_gpus.html
You would have to get a 670 or 680 to ‘upgrade’ IMo you should wait for the 700 series or 8000 series ATI’s. I am.
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 |
(edited by SolarNova.1052)
All thoes ppl caliming that gw2 stays at 60 fps at max setings at full hd yeah sure . . . if you run area free of ppl not orr events and such. Alot of preformance comes from drivers. Somtimes older drivers gives 10+ fps somtimes its the newer ones. And no all that ppl caliming that gw2 eats your cpu alive . . . .i have core i7-3770 turbo to 4.1k never i had more usage then 80% even at thoes huge karka event or wvw fights. Its my hd7850 clocked to 1050 core, 1300 mem thats getting choked at wvw and some events (80-99 gpu usage). If you want to play at smooth rate around 4h on i5 or i7 series is just right you wont rly see improvment from getting it to 5k if your drivers are up to date. 670gtx or hd7950 sounds around right to max gw2 on full hd and still get decent frame rates in zergs. Note that reflections and shadows on ultra re buged right now and they will give you fps drops when you turn around(even some dev confirmed it on forum). And for anyone dont touch ati hd7xxx series. They give nice preformance but the drivers(or half of the cards) are kitten. Half of ppl using anything from hd7850 to 7970 gets flicering,blueish artifacts etc in guild wars and few other games. I myself rmd the card 3 times and all of em had the same flicering artifacts bugs and texture bugs that gets either lesss or worse with some drivers. And no its not my cpu problems cuz my friend gtx670 had no prob in my rig.
Megido.
CPU wise im afraid you are missinformed. If you 3770k cpu does indeed reach 80% usuage playing GW2 i would very much like to know how many programs you have running the background becouse that is waaaaaay to high for a 4 core 8 thread cpu. It should see between 31% and 38% usuage at the point it becomes the bottleneck. Add on what ever % your computer ’idle’s at in your OS with your usual programs running and you could maybe add say 3-5% ontop. So say 35% to 43% usuage is the range of usuage a 3770k will bottlneck at.
As much as i hate going into it time after time ….
The reason for this being is that GW2 uses between 2 and 3 main threads for its performance. If any 1 of those reaches 100% usuage on a THREAD (not core) then a bottleneck occurse becouse GW2 is not programed to offload the caped out thread onto a seperate one.
The way you work out what % you ‘should’ see when your cpu is bottlenecked is…
100/number of cpu threads x 2.5 = (lowest usuage for bottleneck to usualy occure)
100/number of cpu threads x 3 = (highest usuage for bottleneck to usualy occure)
In the case of a 3770k its 100/8 × 2.5
this is not 100% accurate but its close. My cpu caps out between 20-25% (6cores 12 threads) ..but i have seen in extreme cases it reach a tad over 30%. This is on the rare occasions where more than the usualy 3 perfomance threads are having high usage. Think special events like the recent karka one ..where not only do you have normal perfomance threads that are heavily used like in WvW and in LA ..but also dynamic event threads and masses of npc enemies and such.
I wasnt monitoring the thread usuage at this point but its likely that some of the other 50 GW2 threads that GW2 uses between them were being used more and as such accounted for the increased cpu usuage..however it is still highly likely the same 2 or 3 threads where the ones maxing out and reducing framerate.
Hope this helps correct some of your missconceptions
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 |
megido got a point tho…i can’t see much of the great so called “CPU-demanding” that everyone tries to tell you…
if you got a decent 4-core @3GHz+ you don’t have to worry about your CPU
i can OC my i7 how much i want to without a single frame per second more
i’ve played more “CPU-demanding” games
i think, it “should” be as CPU-demanding as anyone tells you
the dev’s trying to optimize the the client’s usage of the user’s available CPU…i hope for an optimization patch within 2012 T_T
@Solar:
oh well…benchmarks
i just saw about 10 different one’s regarding different GPU’s
…they are all completely different
however, the GTX580 in fact seem to be about 10-15% more powerful
can’t afford anything better than a GTX650ti atm so i’ll just wait until the GPU-wars enters the next level
wait…wasn’t this thread about OC’ing a GPU and about it’s performance results?
“Only the finest of potatoes in my zerkburgers.”
(edited by wauwi.9162)