GPU FPS Drops / Stutter - Lowering Memory Clock Solves the problem

[KuLee]

Greetings everyone,

Update:
Full system

Windows 7 - 64 bit clean installed just today (because i though it could be software related which it wasnt)
PSU Corsair vs650 - 650 Watt - Single +12V Rail 60 amps
GPU 9500gt - GTX660 both causing me problems
CPU - Intel 2 Quad Q9550
SSD CRUCIAL CT250MX - 250GB 2.5''
CPU Cooler : CoolerMaster Hyper 412S RR-H412-13FK-R1
MOBO: P5KPL/1600
RAM: Kingston DDR2 2x2GB = 4GB

I finally found out what is making my pc drop in frames and make all my game performances very very bad.

It was memory clock of my gpu for some reason when its on Default 400 Mhz i get really big fps drops which makes my game freeze for couple seconds

Reducing the memory clock to 200Mhz solved the problem but this way my games have total low fps and gpu hungry games are unplayable

Anyone know what is going on? any advices?

 

[KuLee]

Thats good,

Yes near the start he takes off the north bridge.

I cant think of anything else that would cause it as your temps are good on the CPU and GPU and same deal with 2 GPUs so its not thier vrams.

Touching North bridge heatsink with my finger feels to hot almost make me take my finger because it "almost" hurt is this normal?

 

[ingtar33]

Thats good,

Yes near the start he takes off the north bridge.

I cant think of anything else that would cause it as your temps are good on the CPU and GPU and same deal with 2 GPUs so its not thier vrams.

Touching North bridge heatsink with my finger feels to hot almost make me take my finger because it "almost" hurt is this normal?

sorry, went to bed in the middle of this thread, catching up now.

anything hot enough to actually burn you is probably close to too hot, if you lick your finger and touch it to the metal heatsink and you hear a sizzle, then it is 100% too hot; though to get that effect the heatsink has to be over 100C.

 

[KuLee]

Thats good,

Yes near the start he takes off the north bridge.

I cant think of anything else that would cause it as your temps are good on the CPU and GPU and same deal with 2 GPUs so its not thier vrams.

Touching North bridge heatsink with my finger feels to hot almost make me take my finger because it "almost" hurt is this normal?

sorry, went to bed in the middle of this thread, catching up now.

anything hot enough to actually burn you is probably close to too hot, if you lick your finger and touch it to the metal heatsink and you hear a sizzle, then it is 100% too hot; though to get that effect the heatsink has to be over 100C.

Thanks for sticking up with me i really appreciate, rigth now i really need somone to talk i hate being solo

Im trying to cool the NB and see if that solve my problem you can have a look 😛

 

[KuLee]

So no the fan cooling the NB did not solve my problem touching the NB heatsink wasnt super hot but not like touching southbridge southbridge was cool NB was not

 

[SkyNetRising]

Is FSB overclocked (more than 1333mhz)? That could cause those extreme NB temperatures.

 

[KuLee]

Is FSB overclocked (more than 1333mhz)? That could cause those extreme NB temperatures.

No no stock settings 333 x 8.5 = 2830

 

[KuLee]

Its probably not the NB Overheating as i cooled it really hard right now with some extra fans unless the thermal paste cooling is more important to it if not then nb is fine mb temps are 36c then it might be my card dying or i dont know i think ima test some another gpu

 

[KuLee]

Is FSB overclocked (more than 1333mhz)? That could cause those extreme NB temperatures.

I realised something weird my gpu is set at 550 core clock but in monitoring it shows that card is running at 594

 

[ingtar33]

ok, I want you to try something.

1) download DDU http://www.guru3d.com/files-details/display-driver-uninstaller-download.html
2) uninstall that 9500GT
3) install the GTX 660
4) keep the 660 in the top most pci-e slot
5) update your motherboard's bios
6) update all drivers for your motherboard (typically windows won't do this, you'll need to hunt them down on the manufacturer's website)
7) uninstall MSI afterburner or any other GPU overclocking software
8) power down your system and perform a CLR_CMOS (jumper + battery, unplug the system, let it sit for 5 minutes or so)
9) power on your system, set your bios date/time, and any drive settings you need to boot into windows (like changing from legacy IDE to ACHI, whatever settings you had origionally on your hard drives)
10) boot into safemode
11) run DDU, uninstall the nvidia drivers
12) reboot your system into windows (normal)
12) install nvidia driver WHQL 314.22 http://www.nvidia.com/object/win8-win7-winvista-64bit-314.22-whql-driver.html

test only the GTX 660

NVIDIA tends to drop support for their older cards in their latest driver releases. I know 314.22 was the last driver released prior to the release of the 700 series. So in theory it's the last driver released specifically for the 600 series cards.

 

[KuLee]

ok, I want you to try something.

1) download DDU http://www.guru3d.com/files-details/display-driver-uninstaller-download.html
2) uninstall that 9500GT
3) install the GTX 660
4) keep the 660 in the top most pci-e slot
5) update your motherboard's bios
6) update all drivers for your motherboard (typically windows won't do this, you'll need to hunt them down on the manufacturer's website)
7) uninstall MSI afterburner or any other GPU overclocking software
8) power down your system and perform a CLR_CMOS (jumper + battery, unplug the system, let it sit for 5 minutes or so)
9) power on your system, set your bios date/time, and any drive settings you need to boot into windows (like changing from legacy IDE to ACHI, whatever settings you had origionally on your hard drives)
10) boot into safemode
11) run DDU, uninstall the nvidia drivers
12) reboot your system into windows (normal)
12) install nvidia driver WHQL 314.22 http://www.nvidia.com/object/win8-win7-winvista-64bit-314.22-whql-driver.html

test only the GTX 660

NVIDIA tends to drop support for their older cards in their latest driver releases. I know 314.22 was the last driver released prior to the release of the 700 series. So in theory it's the last driver released specifically for the 600 series cards.

Okay i did everything to ensure that my PC running at best temps everything runs at 36C and cpu max load is 40-45C gpu 50-55C mobos drives all cooled to 36C im going to do what u said now but im not confident enough i am expecting DPC latency :/

 

[KuLee]

ok, I want you to try something.

1) download DDU http://www.guru3d.com/files-details/display-driver-uninstaller-download.html
2) uninstall that 9500GT
3) install the GTX 660
4) keep the 660 in the top most pci-e slot
5) update your motherboard's bios
6) update all drivers for your motherboard (typically windows won't do this, you'll need to hunt them down on the manufacturer's website)
7) uninstall MSI afterburner or any other GPU overclocking software
8) power down your system and perform a CLR_CMOS (jumper + battery, unplug the system, let it sit for 5 minutes or so)
9) power on your system, set your bios date/time, and any drive settings you need to boot into windows (like changing from legacy IDE to ACHI, whatever settings you had origionally on your hard drives)
10) boot into safemode
11) run DDU, uninstall the nvidia drivers
12) reboot your system into windows (normal)
12) install nvidia driver WHQL 314.22 http://www.nvidia.com/object/win8-win7-winvista-64bit-314.22-whql-driver.html

test only the GTX 660

NVIDIA tends to drop support for their older cards in their latest driver releases. I know 314.22 was the last driver released prior to the release of the 700 series. So in theory it's the last driver released specifically for the 600 series cards.

Did everything you said GTX 660 On right now and this is what happening

 

[ingtar33]

so basically nothing is behaving well.

ugh~ this will be a bear to track down. how is the performance when gaming? unchanged?

I'm half tempted to tell you to breadboard it. however that probably won't help.

well, since temps aren't the problem, and it looks like we can write off video drivers, bios updates, motherboard drivers. we're getting somewhere atleast. or we know what it ISNT.

I'm still drawn back to my first thoughts on this which was power issues. Can you replace that PSU? will it be a financial hard ship? do you have a spare somewhere you can borrow?

 

[KuLee]

so basically nothing is behaving well.

ugh~ this will be a bear to track down. how is the performance when gaming? unchanged?

I'm half tempted to tell you to breadboard it. however that probably won't help.

well, since temps aren't the problem, and it looks like we can write off video drivers, bios updates, motherboard drivers. we're getting somewhere atleast. or we know what it ISNT.

I'm still drawn back to my first thoughts on this which was power issues. Can you replace that PSU? will it be a financial hard ship? do you have a spare somewhere you can borrow?

i heard that it might be an IRQ Conflict and i have ACPI APIC enabled in bios ( Which is power management and IRQ Sharing) You thing setting it disabled will solve my problem? I already tried to disable it but then windows wont boot and im not even able to reinstall windows because i get an error there 0x00000025

 

[ingtar33]

yeah, if you don't install windows with that disabled you'll need to install windows again. I wouldn't bother. it's obviously never been turned off before.

IRQ conflicts happen with faulty devices, maybe something is failing? dying hard drive, dying dvd drive, dying motherboard, dying usb device. you can check the event viewer -> system log and see if any hardware is throwing up massive errors; might give us an idea of what might be failing. like "drive not detected" or "device stopped responding" type of errors.

 

[KuLee]

yeah, if you don't install windows with that disabled you'll need to install windows again. I wouldn't bother. it's obviously never been turned off before.

IRQ conflicts happen with faulty devices, maybe something is failing? dying hard drive, dying dvd drive, dying motherboard, dying usb device. you can check the event viewer -> system log and see if any hardware is throwing up massive errors; might give us an idea of what might be failing. like "drive not detected" or "device stopped responding" type of errors.

Or maybe just the gtx 660 dying and conflicts everything else lol how i am able to install windows tho? i disabled it and i cant install windows i get an error 0x00000025

 

[ingtar33]

yeah, if you don't install windows with that disabled you'll need to install windows again. I wouldn't bother. it's obviously never been turned off before.

IRQ conflicts happen with faulty devices, maybe something is failing? dying hard drive, dying dvd drive, dying motherboard, dying usb device. you can check the event viewer -> system log and see if any hardware is throwing up massive errors; might give us an idea of what might be failing. like "drive not detected" or "device stopped responding" type of errors.

Or maybe just the gtx 660 dying and conflicts everything else lol how i am able to install windows tho? i disabled it and i cant install windows i get an error 0x00000025

re-enable it. if you can't even boot into the bios, then clr_cmos