Sign in with
Sign up | Sign in
Your question
Closed

No performance increase on my new GTX 650 Ti

Last response: in Graphics & Displays
Share
December 21, 2012 4:24:20 PM

Hi, I just bought a Galaxy GTX 650 Ti to replace my old PNY GT 430, expecting a major boost in performance but that wasn't the case

I unisntalled my old drivers, slotted the new card correctly, plugged the power cables, downloaded and installled the latest NVIDIA drivers (310.70) and after countless tests in different games my FPS is exactly the same or worse...My Windows experience in Graphics did went up from 6.7 to 7.4 but that was it...

My specs:

CPU: E5700 dual core 3.0Ghz
MOBO: Intel G41 Express Chipset P49G by PChips
RAM: 4gb DDR3
PSU: ATX 500w PIV
OS: Windows 7 32bit
Display: 32inch Sony Bravia 720p HD TV (using the VGA to DVI adapter the card came with)


I also get this weird 3D/high contrast/yellowish look in all games if I leave the 3D vision programs from nvidia installed

Btw when I used GPU-Z on default clock it says 540Mhz and on GPU clock 966, which I found weird since this card is factory overclocked and the default clock should be 928mhz

After Heaven Benchmark v3.0 Basic

FPS: 21.7
Scores: 548
Min FPS: 5.4
Max FPS: 54.7

Settings

Render: direct3d11
Mode: 1920x1080 4xAA fullscreen
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 16x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled
Tessellation: extrem

I know my cpu isn't the best but its a decent 2 core processor and most games I own only utilize 2 cores at max (its only 10months old) My psu is a year and a half old and its not a branded supply but it's a 500w shoulnd't it perform ok on a card that only requires 400w?


Thanks in advance.
a c 271 U Graphics card
December 21, 2012 4:26:42 PM

Bump this thread and I may get my banhammer out.
Score
0
a b U Graphics card
December 21, 2012 4:27:43 PM

your cpu is the bottleneck. how do you know your games only utilize 2 cores if you have not tried a quad core? trust my you will see a difference
Score
0
Related resources
December 21, 2012 4:57:12 PM

Will a i3 3220 resolve my issue? also can I install that in my current mobo?
Others have told me the issue it's the psu and a branded 430w+ will solve it. Im on a tight budget atm after getting the new GPU(and a little frustated too after seen no good results) So I want to be sure which is the villain and what is affordable for me here before spending more money.
Score
0
a c 105 U Graphics card
December 21, 2012 6:18:54 PM

different socket. won't work.
Score
0
December 21, 2012 10:48:40 PM

Back on topic, will a new branded PSU like a corsair 500w fix my performance issues for the most part? (can a PSU bottleneck a gpu that much?) or I truly need a new mobo/cpu to get a decent performance in most games? and with decent I mean 1280x720 settings on high(some)/medium(most) with 30ish fps
Score
0
a c 105 U Graphics card
December 21, 2012 11:01:10 PM

we never left the topic. no, a power supply only supplies power to the system. it can hamper performance if it can't provide enough. right now your old processor/mobo/ram are holding you back and you need to replace all 3 because you can't use any of those old parts with new hardware.

only thing that will increase fps now would be those 3 items......... and then maybe a new PS......... depending on exactly what one you have now..??

if you're not having slow ups in anything I wouldn't even worry about it.

when you install drivers do a custom or advanced install. only install the drivers you need. in this case the physx and graphics driver only.. not the sound driver nor the auto update trash and anything else it thinks you're dumb enough to install.
Score
0
a b U Graphics card
December 21, 2012 11:12:58 PM

Well, a 650Ti just isn't that powerful a vidcard...
And that test (Heaven) was run at 4x AA, everything high, 16x AF, trilinear, and Tesselation: extreme.
That's what you get! ~22 FPS.
There's nothing wrong with your CPU.
Though you might in fact benefit from a more modern quad-core CPU, if you ever get a more powerful vidcard.
Score
0
a b U Graphics card
December 21, 2012 11:16:46 PM

"Btw when I used GPU-Z on default clock it says 540Mhz and on GPU clock 966, which I found weird since this card is factory overclocked and the default clock should be 928mhz"

GPUZ is up to date right? I know my MSI Afterbruner was an older version and when I upgraded to a 660 (From 550ti) it was displaying the Clock Speeds wrong.
Score
0
December 21, 2012 11:57:22 PM

simmons33 said:
"Btw when I used GPU-Z on default clock it says 540Mhz and on GPU clock 966, which I found weird since this card is factory overclocked and the default clock should be 928mhz"

GPUZ is up to date right? I know my MSI Afterbruner was an older version and when I upgraded to a 660 (From 550ti) it was displaying the Clock Speeds wrong.


I just updated to the gpu-z v0.6.6 and now it shows the rght clockspeed thanks :) 

Back on topic: Can the psu really bottleneck my gpu even though the system in booting and running benchmarks such as FurMark for 15mins+ without crashing/shutting down?
Im just really surprised there's virtually no performance increase over my year old GT 430, infact every single game appears to run at the exact same framerate...
Score
0
a b U Graphics card
December 22, 2012 12:15:37 AM

Well Id imagine that if doesnt have enough power to give your GPU then your GPU wouldnt be able to max out like it needs to.

I would honestly think the computer would just shut itself down.

Ive never used it so is it possible to run that Heaven test on Settings you would actually use in a real game? Did you run the same test with the same settings on your 430?

Is GPU useage being displayed as 99%?

Tessalation on a 650ti just doesnt seem all that possible and also retain a good framerate. My 660 still gets a bit of a knock down when I use tessalation. And I usually use the lowest setting possible in game that support it to retain 55+ FPS (Besides OFF of course)

And of course there is always the issue of a CPU bottleneck. If the CPU cant throw out the INFO faster than the GPU is giving then your in bad shape
Score
0
a c 107 U Graphics card
December 22, 2012 12:23:49 AM

Want to know whats sad. My 650 ti(with an i5 750) beat my 5870(with an i7 2600k @ 4.4) in that benchmark with ambient occlusion on.

Its honestly sad compared to what most users get, but then again a 650 ti is not a beast by any means

FPS: 79.8
Scores: 3384
Min FPS: 59.7
Max FPS: 97.7

Binary: Windows 32bit Visual C++ 1500 Release May 21 2010
Operating system: Windows 7 (build 7601, Service Pack 1) 64bit
CPU model: Intel(R) Core(TM) i5 CPU 750 @ 2.67GHz
CPU flags: 2666MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 SSE42 HTT
GPU model: NVIDIA GeForce GTX 650 Ti 9.18.13.1070 1024Mb
Settings
Render: direct3d11
Mode: 1920x1080 4xAA fullscreen
Shaders: medium
Textures: high
Filter: trilinear
Anisotropy: 8x
Translucence: enabled
Parallax: enabled
Reflection: enabled
DOF: disabled
HDR: disabled

Something strange about it. All games show the reverse for clear reasons.

If you ran out of power, I would guess it would just crash out on you, but the 650 ti is not that power hungry.

How many amps on the 12 volt rail?
Score
0
December 22, 2012 12:29:24 AM

Thats the strange thing, I just ran the 15mins FurMark test on 1920x1080 with 0x AA and got an avg of 19fps max temp 67degrees celcius, and yes the gpu was at 99%
my 430 used to score less than 10fps in that test BUT in actual games like Borderlands 2, Arkkam city and every single game I have installed (with the exact same settings) runs the same or even worse...

If the pc is stable atm(no trouble booting,/resets/shutdowns) and the GPU usage displays 99% on high demand should I forget about the psu as the issue?

EDIT for previous response:

AC input:
115v 60mhz 6a
230v 50mhz 5a
DC output:
+3.3v +5v +12v -12v +5v sb
15a 25a 15 0.5a 2a

+3.3v/ & +5v MAX:105w TOTAL POWER: 500w

Thats all the info in the PSU sticker
Score
0
a c 107 U Graphics card
December 22, 2012 12:31:33 AM

12 volt amps please?

In general, it crashes before it fails.

No way around it, if the cpu holds you back, games may not see improvement. Have you ever wanted to overclock a cpu :) 

EDIT you edited :p 

so 3.3@15a 5@25a 12@15a -12@0.5a +5SB@2a ?

15 amps for 12 volts is very low. That is 180 watts. That is like 250 watt power supply land.

I now worry. Lucky you have a 65 watt cpu, but DAMN that has got to be close.
Score
0
a c 107 U Graphics card
December 22, 2012 12:41:51 AM

bump to ensure you see my edit /\

Mods, i promise never to do it again. :) 
Score
0
December 22, 2012 12:53:59 AM

nukemaster said:
12 volt amps please?

In general, it crashes before it fails.

No way around it, if the cpu holds you back, games may not see improvement. Have you ever wanted to overclock a cpu :) 

EDIT you edited :p 

so 3.3@15a 5@25a 12@15a -12@0.5a +5SB@2a ?

15 amps for 12 volts is very low. That is 180 watts. That is like 250 watt power supply land.

I now worry. Lucky you have a 65 watt cpu, but DAMN that has got to be close.


I can't because my Bios doesn't have the option and my cpu isn't listed in all the OC software I have tried
to use.

Can you repeat the voltage/amp problem you just found more clearly to me?
you mean my psu is barely running my current system or it's doing it incorreclty?

thanks in advance.
Score
0
December 22, 2012 1:15:48 AM

hmmmm mousemonkey threatens to ban people if they bump this thread. TWO MINUTES AFTER THE THREAD IS OPENED. guess any bitch can mod at tom's......i have lost a lot of respect here. yes it was a stupid question, of course his cpu is the bottleneck. but to show that kind of intolerance.....geez. like no one has asked a stupid question before. show a little tolerance dude. the world doesnt conform to your expectations.


ps guess i am banned now :) 
Score
0
a c 107 U Graphics card
December 22, 2012 1:16:53 AM

I will try to clear it up.

Power supplies are rated in TOTAL power.

Now older or cheaper designs tend to place MORE power on the lower voltage rails(3.3/5). While newer and more expensive power supplies place more of the power(current/amps/watts whatever you want to call it) on the 12 volt rail(s).

So take your system.
3.3 volts x 15 amps = 50 watts(49.5)
5 volts x 25 amps = 125 watts
12 volts x 15 amps = 180 watts
-12 volts x 0.5 amps = 6 watts
5 volts(standby) x 2 amps = 10 watts
50+125+180+6+10 = 371. It seems something is missing. Can I get an image?

Either way, Newer power supplies deliver something in the way of 80-90% or more of the total power to the 12 volt rail(s). This is NOT to be confused with 80+ as that is a power efficiency standard(less heat/less noise/lower cost to operate/better for the environment).

At one point the ATX specs(what power supplies follow in one way or another) recommended no more then 18 amps per 12 volt rail so power supplies suddenly had tons of rails(4-6).

This added confusion because all the rails on most multirail systems get power from one big 12 volt rail and just have protection to prevent more then 18 amps(20 as the tripping point for most) from being taken from any rail(think of safety), but the total is not always the same as adding all those rails.

I will give you an example of a mid end easy to read power supply.

This is simply one of the BEST 300 watt power supplies you can get

Now if you look at the boxes you will see that the power supply can deliver 300 watts TOTAL and 288 of that can be from the 12 volt rails, but you will also notice that you see another 120 from the 2 lower rails(3.3 and 5, we will ignore the -12 and 5v standby, but the 5v standby is its own system).

This does not add up. it is WAY over 300 watts.

This is how power supplies are made now. You make one huge 12 volt rail then you can take all the lower voltages from it(with a DC-DC converter). the 3.3 and 5 actually come from the same 12 volt rail.

If you load them up to the max, you instantly loose 120 watts off that 288 watt rating.

Lucky for us, not much of the 5 volt is used(2.5 watts for a high power usb device as per the specs, much less for most keyboards and whatnot) and even less 3.3 volts is used.

On older systems the system used the other rails for much more so a large 12 volt rail was not needed and power supplies worked much differently(and MUCH less efficient as DC-DC conversion is just great :)  ).

You will also notice that the power supply has 2 17 amp 12 volt rails, but that does not mean it has 17 + 17(34) amps for its 12 volt rail. it is listed below as 288 watts(288watts / 12volts = 24amps).

Maybe this is too much(confusing), but a 12 volt at 15(180 watts) amps would either be misread OR just real old. It seems something is missing since I do not see 500 watts at all.

Also the 3.3 + 5 volt MAX of 105 makes no sense as well since 5 volts x 25 amps = 125 watts.
Score
0
December 22, 2012 2:32:35 AM

I can't take a picture atm but this is all the info on the sticker:

ATX Switching power suplly
MODEL:ATX-500w-PIV

AC input:

115v 60mhz 6a
230v 50mhz 5a

DC output:

+3.3v +5v +12v -12v +5v sb
15a 25a 15a 0.5a 2a

+3.3v/ & +5v MAX:105w TOTAL POWER: 500

(same as before I double checked)

Name of product NO: (TD-500)
made in 2005 @china

and then the safety cautions and thats it.

Even though it's obviously a crappy psu so far i haven't had a single freeze/shutdown/reset and my gpu keeps displaying 99% usage on furmark/heaven/gpu-z, so back to my original concern, can I eliminate the psu as the performance bottleneck and blame it on the mobo/cpu?
Looking at benchmarks from some of the games I own people with this card (even with mid-end cpus and 4gbs of ram) can run most games at high on 1080p with 30-40fps

I've tried countless drivers, re-seated everything on my rig couple of times, tested back my year old GT 430 and get the SAME fps in every single title...
Score
0
a c 107 U Graphics card
December 22, 2012 2:41:25 AM

I would say that power supply is quite old, but again, lucky for you the system is not power hungry.

do NOT OVERCLOCK.

I am betting on the cpu. Closest i can do is try to run the same benchmark with my cpu clocked down and 2 cores disabled and post some results in the morning(or maybe tonight if i get a chance.).

The problem is not quite knowing what clock speed to pick. newer cpus do more work per MHz.
Score
0
December 22, 2012 2:46:32 AM

nukemaster said:
I would say that power supply is quite old, but again, lucky for you the system is not power hungry.

do NOT OVERCLOCK.

I am betting on the cpu. Closest i can do is try to run the same benchmark with my cpu clocked down and 2 cores disabled and post some results in the morning(or maybe tonight if i get a chance.).

The problem is not quite knowing what clock speed to pick. newer cpus do more work per MHz.


I would really appreciate that thanks :) 
Score
0
a c 107 U Graphics card
December 22, 2012 3:02:11 AM

CRAP.

I have the wrong benchmark. I had Sanctuary installed and was thinking you had that.

Getting the right one.

Time for this card to suffer.
Score
0

Best solution

a c 107 U Graphics card
December 22, 2012 3:38:27 AM

Ok here we have it.

Test system
Win7 HP 64-bit
Silverstone SG05
FSP 300 Watt SFX PSU <- yeah thats right!
H55N USB3
I5 750 undervolted <- every watt counts
Geminii S + 1200 rpm fan(Globe)
2 x 4096 @ 9,8,8,24 (1333)
GTX 650 Ti(MSI Cyclone II)

I5 750 with 2 cores disabled. I set power savings to MAX to try to keep the clocks down. Check it out, even worse then you got :)  This setting did not keep the card that busy.

FPS: 12.4
Scores: 312
Min FPS: 5.8
Max FPS: 20.8
Hardware
Binary: Windows 32bit Visual C++ 1600 Release Mar 7 2012
Operating system: Windows 7 (build 7601, Service Pack 1) 64bit
CPU model: Intel(R) Core(TM) i5 CPU 750 @ 2.67GHz
CPU flags: 2666MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 SSE42 HTT
GPU model: NVIDIA GeForce GTX 650 Ti 9.18.13.1070 1024Mb
Settings
Render: direct3d11
Mode: 1920x1080 4xAA windowed
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 16x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled
Tessellation: extreme

Stock clocks 4 cores, please note the min frame may be off as i started task manager while it was running. I am not 100% sure if windowed effects the score, but i wanted to see cpu-z, gpu-z and hwmon while it ran.

Heaven Benchmark v3.0 Basic
FPS: 25.9
Scores: 653
Min FPS: 11.5
Max FPS: 62.7
Hardware
Binary:
Windows 32bit Visual C++ 1600 Release Mar 7 2012
Operating system: Windows 7 (build 7601, Service Pack 1) 64bit
CPU model: Intel(R) Core(TM) i5 CPU 750 @ 2.67GHz
CPU flags: 2666MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 SSE42 HTT
GPU model: NVIDIA GeForce GTX 650 Ti 9.18.13.1070 1024Mb
Settings
Render: direct3d11
Mode: 1920x1080 4xAA windowed
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 16x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled
Tessellation: extreme

All in all, it actually is less demanding then guild wars 2 from a system power consumption point of view. This averaged about 132 watts at the wall to run. Honestly by game standards despite being so hard on this mid end video card, many games are much harder on the cpu.

Note : This is a very lean system(so low power).
Share
December 22, 2012 3:57:15 AM

nukemaster said:
Ok here we have it.

Test system
Win7 HP 64-bit
Silverstone SG05
FSP 300 Watt SFX PSU <- yeah thats right!
H55N USB3
I5 750 undervolted <- every watt counts
Geminii S + 1200 rpm fan(Globe)
2 x 4096 @ 9,8,8,24 (1333)
GTX 650 Ti(MSI Cyclone II)

I5 750 with 2 cores disabled. I set power savings to MAX to try to keep the clocks down. Check it out, even worse then you got :)  This setting did not keep the card that busy.

FPS: 12.4
Scores: 312
Min FPS: 5.8
Max FPS: 20.8
Hardware
Binary: Windows 32bit Visual C++ 1600 Release Mar 7 2012
Operating system: Windows 7 (build 7601, Service Pack 1) 64bit
CPU model: Intel(R) Core(TM) i5 CPU 750 @ 2.67GHz
CPU flags: 2666MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 SSE42 HTT
GPU model: NVIDIA GeForce GTX 650 Ti 9.18.13.1070 1024Mb
Settings
Render: direct3d11
Mode: 1920x1080 4xAA windowed
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 16x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled
Tessellation: extreme

Stock clocks 4 cores, please note the min frame may be off as i started task manager while it was running. I am not 100% sure if windowed effects the score, but i wanted to see cpu-z, gpu-z and hwmon while it ran.

Heaven Benchmark v3.0 Basic
FPS: 25.9
Scores: 653
Min FPS: 11.5
Max FPS: 62.7
Hardware
Binary:
Windows 32bit Visual C++ 1600 Release Mar 7 2012
Operating system: Windows 7 (build 7601, Service Pack 1) 64bit
CPU model: Intel(R) Core(TM) i5 CPU 750 @ 2.67GHz
CPU flags: 2666MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 SSE42 HTT
GPU model: NVIDIA GeForce GTX 650 Ti 9.18.13.1070 1024Mb
Settings
Render: direct3d11
Mode: 1920x1080 4xAA windowed
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 16x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled
Tessellation: extreme

All in all, it actually is less demanding then guild wars 2 from a system power consumption point of view. This averaged about 132 watts at the wall to run. Honestly by game standards despite being so hard on this mid end video card, many games are much harder on the cpu.

Note : This is a very lean system(so low power).


Wow I didn't knew the cpu would make such a hugeee difference in performance (atleast now I know my gpu iis not faulty) when I asked here previously about getting this new gpu most people told me that with high resolutions and high AAx etc the cpu wouldn't bottleneck the machine because of the huge demand on the gpu, and that I could upgrade/see a nice performance increase with the 650ti since it wasn't a high-end or hungry card but turns out for me that wasn't the case.

Just one last question, which mobo/cpu combination should I get to get to see the full potential of this card without breaking the piggy bank(im on a tight budget atm)

Thanks in advance for your time and effort, you earned that 'best answer' thanks :) 
Score
0
December 22, 2012 6:30:08 AM

Best answer selected by Sprix.
Score
0
!