At a total loss why gaming is so flaky

e46_m3

Established Member
Joined
Mar 7, 2006
Messages
203
Reaction score
4
Points
75
Location
Nottingham
My current set up is:

Core 2 Quad 6600
Asus P5B
4GB RAM
1GB BFG GTX280
500GB Seagate SATA300 32MB cache HDD with 450GB free space
650W PSU
Creative Xi-Fi Xtreme Audio
Vista 32 bit SP1
DX10

I am totally peeved with PC gaming - my old set up (core 2 duo and 8800GT) used to do weird things in games so I bought a new CPU, PSU, graphics card and HDD and still the same things happen. Im running the latest Nvidia drivers (7.15.11.7813 - not Beta drivers) from their website and Vista was reinstalled from scratch yesterday so what could cause this?

When Im playing a game (UT3 just now for example) the game either just freezes up or the screen goes all different colours like purples and reds etc so its unplayable - only way to make it go away is to ctrl-alt-del (but not end the process/task) and then when I go back everything is back to normal and works ok for another few mins till it happens again making the game almost unplayable.

Any ideas? I thought it was my old set up not being powerful enough but its obviously not that and after spending over 500 quid on an upgrade Im well ****** that its still not working and games are virtually unplayable - the only common pieces between old and new are the mobo, soundcard and RAM and everything else has been replaced.

I havent tried it in any other games as this is exactly the same as what it did before on my old setup but I know if I install anything else Im going to get the same problem(s). It doesnt do anything anywhere else except in games hence my original thinking it might have been maybe the graphics card but this is brand spanking new and a GTX280 should wee all over UT3 without any problems.........

Help me - please :(
 
Sounds like a VGA issue to me. You haven't OCed anything have you? If so, reset everything back to stock speeds.

What make & model is your PSU. If you got a ****** one, then it might be getting stressed out by the GTX280 demands.
 
Sounds like a VGA issue to me. You haven't OCed anything have you? If so, reset everything back to stock speeds.

What make & model is your PSU. If you got a ****** one, then it might be getting stressed out by the GTX280 demands.


Nothing is overclocked at all, everything standard - I did wonder about the PSU but the one I took out from my last setup was a decent one. This time I did only go for a budget one but it "looked" ok spec wise so does anyone have any experience of this one? Good? Bad? OK? Crap?

EZ Cool 600W SLI Ready PSU - 12cm Fan, 2x PCI-E, 20+4pin, 1x SATA
 
That leaves the motherboard then! You could remove the soundcard and maybe experiment with removing and swapping the RAM around but only logical thing left is the mobo.
 
Shagged memory is where my money is.

download, burn to CD and boot from:

Memtest86 - Download Page

You'll see the errors if any down the bottom.

Let it run overnight.
 
i would possibly be weary of that psu ezcool is not a very good make imo, what are your voltages like at idle and load? what are your temperatures like at idle and load?? (cpu,mobo and gpu)
i would suspect the motherboard more than the ram, but mem test is always a good idea.

so just to confirm you where getting this exact same problem with your old system?? so you replaced the processor, hard drive, graphics card and power supply??

your new power supply is a 600watt ezcool? what was your old psu??
 
Shagged memory is where my money is.

download, burn to CD and boot from:

Memtest86 - Download Page

You'll see the errors if any down the bottom.

Let it run overnight.
If it was memory then Windows would be playing up too imo. My money is the cheap PSU - as I've never heard "EZ Cool" PSU brands.
 
i would possibly be weary of that psu ezcool is not a very good make imo, what are your voltages like at idle and load? what are your temperatures like at idle and load?? (cpu,mobo and gpu)
i would suspect the motherboard more than the ram, but mem test is always a good idea.

so just to confirm you where getting this exact same problem with your old system?? so you replaced the processor, hard drive, graphics card and power supply??

your new power supply is a 600watt ezcool? what was your old psu??

How can I check temperatures etc? Via a mobo utility or is there something which will do everything inc PSU voltages etc?
 
I read a gtx280 can use up to 240 watts, which is 12 volts x 20 amps. I suggest as yours is a SLI model PSU you plug one power line into the six pin socket and the other into the eight pin and see if that helps.

My personal preference was to go with the Hiper Type M 880W, their earlier models weren't that great but the 880w is nice with very stable voltages.

For testing temperatures and voltages I suggest the free HWMonitor from CPUID
For checking system info there's CPUz which you can get free from CPUID too.
 
Last edited:
How can I check temperatures etc? Via a mobo utility or is there something which will do everything inc PSU voltages etc?

speed fan is quite a good unility to use. its always best to check temeratures and voltages from within the bios. or if your more compitent with what your doing you can even use a multimeter directly on the cables on the psu (dont try this if your not 100% sure you know what your doing)

my bet is with temperatures or weak power supply. i have seen ez-cool psu's but they are generaly quite cheap psu's and ive use one which was very poor imo, the voltages where quite low at idle and on load where dropping way below the required minimum. (the psu should have been more than powerfully enough to power the system it was in)

have a look in the bios for temperatures and voltages if you can as they tend to be more accurate. be carefull when you exit the bios...... exit the bios without saving settings as that way you wont change things accidently.

nvidia have been having problems with the gtx260's and 280's some of them are running far to hot because of problems with the manufacturing process. but lets not jump there right now.
 
It's not very useful knowing what your idle temperatures and voltages are (unless they fluctuate wildly by a lot as you look at them doing nothing in which case you know you have problems), and your computer doesn't get much more idle than in the bios when you're not even running an operating system.

I still suggest HWMonitor rather than Speedfan as people tend to refer to the inside-the-core temperatures and I'm not sure whether Speedfan uses those particular diodes or does it for all four cores showing the maximums and voltages etc. That and I don't see a lot of people that overclock regularly who suggest it. It's usually either HWMonitor or Coretemp which both measure the same temperatures but HWMonitor shows voltages too so I prefer it.

I suggest you leave HWMonitor running in the background while you run 3dmark06 or something more stressful and see what the maximums are.

If you decide to try Orthos as a stress test then be aware you have to click Stop to stop it. If you close the window instead then you may think it's stopped but its still running full steam and just minimised to the system tray.

To stress all four cpu cores with Orthos you need to run two copies and set the affinity of one of them to the third and fourth cores (Ctrl-alt-del -> Processes -> right click Orthos -> Set affinity -> click 1 and 2 off and 3 and 4 on)

Orthos is extremely stressful to cpu's, even more than Crysis so if you can get through several hours of it your cpu can stand anything you can throw at it. Watch those temperatures carefully though. If any of the four cores reach 70c I'd suggest clicking Stop on both Orthos copies quickly.

The ATITool is quite a decent videocard stress test even for Nvidia cards though I wouldn't use it to change the videocard clock speeds.

Btw thanks CP-PC you reminded me of something I came across quite recently. I noticed some GTX280's capacitors can buzz or cause a PSU's capacitors to buzz (if they're glued in solidly they may not buzz but I'm sure are still under a lot of physical stress) when the framerates get crazily high during quite boring moments of a screen fade to black due to their much higher power requirements and speed, it can reach 4000fps to 5000fps. 3DMark06 does this on fadeouts between scenes and you can see for yourself using Fraps to show the framerate.

I think it's a good idea to go into the Nvidia control panel -> Manage 3D Settings -> and switch Vertical sync to Force on, and Triple buffering to on. Vertical sync will limit the framerate to the maximum your monitor can handle starting each new redraw at the top of the screen and avoids tearing in games where it tries to draw the next screen when halfway down the monitor. Triple buffering will allow it to always have the next screen ready to draw the moment the previous one has finished. This will reduce the stress on the capacitors and if you're unlucky enough to have a buzzing GPU/PSU it should fix it as a side effect.
 
Last edited:
i was mearly pointing out that values from the bios tend to be more accurate and idle voltages and temperatures can be usefull if only to verify software readings are the same as bios. crysis is more graphicaly intence than cpu, there are many programs out there to stess components. and as for your comment about early hiper models being poor....... what are you talking about?? some of the lower end hiper range (especialy the 580w) received multiple awards and have a very good reputation for being a very good reliable stable power supplys. im running one of the hiper 580's right now and its been flawless AND silent every since i installed it.
 
as for your comment about early hiper models being poor....... what are you talking about?? some of the lower end hiper range (especialy the 580w) received multiple awards and have a very good reputation
I said their earlier models weren't that great, that doesn't mean poor in a strictly like to like comparison against other brands, it just means they weren't at the standards the Type M and Type R 880w set, which are a much newer improved design even over the 780w. Back when the 580w came out it was probably head and neck above most of the others.

This isn't helping e46_m3 unfortunately so if you want to know more about it you'll have to use google to read all the reviews and user comments which is pretty much the way I research this stuff. One note, always check the exact model number and the year of a review which is usually only mentioned on the top of the first page, and not just the name of the item as they sometimes base everything on a pre-production version.

Back to helping, here's a simple test, try playing with the side of the computer off and see if it helps.

Try blowing into the CPU heatsink when the computer is off (don't spit and an aircan would be better, hold any fans don't let them turn as you blow through them) and if you get a duststorm it's time for a cleanout.

If everything has good temperatures and voltages I suggest checking on the Asus site to see if there's a relevent motherboard bios update you need. Don't flash a new bios unless you really really need it though. It's disastrous when it goes wrong.

If you solve the problem be sure to let us know.
 
i actually happen to know a fair amount about the matter mate and exactly how far back in the hiper psu line are you going? even when the 580watt type r came out, there where all the rage, the only differenc between the old type r and the new type m is that the m's are not modular or "flashy" like the type r's where. (plsu the type m's only have a single 120mm fan) but agree'd this is not really helping the op, we just need to wait for him to post back some voltages and temperatures or results of some of the sujestions given to him
 
@ CP-PC and Krakin: can you keep this thread on topic, this is the 2nd thread where you seem to have clashed, can't you both agree to disagree as it doesn't help the op.

Curly
 
Am on this today checking out the utils etc and will come back with the results as soon as I have them
 
@ CP-PC and Krakin: can you keep this thread on topic, this is the 2nd thread where you seem to have clashed, can't you both agree to disagree as it doesn't help the op.

Curly

dont get me wrong but i dont really understand why you felt the need to "chip in" as both me and krakin had said its getting off topic and agree'd that its not really helping the op
if by "clash" you mean difference of opinion then yeah i agree but thats going to happen multiple times daily between many many different users. but once again we digress

@ the op... looking forward to your results, so we can try and get to the bottom of this as your system is very expencive and pretty useless if you cant use it due to corrupt screens
 
Last edited:
Have ran 3D Mark and got the temp results etc but two things:

On ATiTool and Orthos whats the maximum CPU/GPU temperature I can safely let it get to? I want to leave certainly ATITool running for a few hours unattended byut I dont want it to do any harm to my system by overheating it especially as I only have stock cooling in my case.

Also when I tried to run Orthos it started and ran for about 2 minutes (both copies, one per core pair) and then gave me a BSOD/STOP screen for memory management and rebooted - could it be my memory which is screwed up then after all?

Before running 3D Mark the temps etc were:

normal-pre-tests.gif


After running 3D Mark I get the following:

after-3d-mark.gif


Any thoughts or comments thus far?
 
the +5v line looks a little low. cpu temps looks to be ok. gtx 280 temps looks ok aswell. your cpu should shut down or throttle if it gets to hot. and i think your gpu will do the same but can someone else confirm this?
maybe it is the ram then, going by your temperatures you should be fine to run mem test so try that next.
if everything is fine then you should get NO errors at all. it does take a while to run so its not going to be over in 10 mins like 3dmark is.

one thing i did notice is that your vcore has quite a droop on it, and im not sure why your +12v line jumped up to over 14volts

actually the vcore seems a little low to me as the q6600 runs at 1.25volts at stock does it not?? so it may be your power supply after all, as under load it looks like your cpu voltage drops to 1.1volts

the next step imo would be to run memtest as said earlier (he may be right after all about the ram) but we will soon see.

at the moment it may still be a multitude of things like your power supply (over 14 volts on the 12volt line seems weird to me although it doesnt drop below 12.6v... puzzleing)
it could be your power supply only providing 1.1volts to your processor
or it could be memory so run memtest and see what that does.

although if your processor is running on 1.1volts it may not be stable so mem test may fail which would make it look like a memory failiure but it may actually be your processor not having enough power.

your currently in use psu is your 650watt ez-cool correct? what was the make of your older 600watt psu?
 
Also looked at BFG website and under system requirements for my card it states:

"575W PCI Express®-compliant system power supply with a combined 12V current rating of 42A or more (Minimum system power requirement based on a PC configured with an Intel Core®2 Extreme QX9650 processor)"

Then found the website for EZCool in the UK and under my PSU it says Maximum Load of:

+3.3V 32A
+5V 40A
+12V 24A

C & J Computers - EZcool PSU - 600W 12cm

Does that mean that the maximum 12v current rating of this PSU is only 24A and, is therefore, WAY under powered for the job?
 
yeah looking at the specs for your psu its well under powered as the gtx280 can take up to 20amps from the 12v rail just by its self. what other psu have you got???

just for referance my hiper 580watt psu has a total of 38 amps on the 12v line which just goes to show ez-cool falls into the "cheapo" psu catagory

the 600watt quote for your ez-cool psu is a "peak" reading. i would get something like my hiper 580watt psu minimum
while the hiper 580 will power your system it will be near 90-100% load when your pc is running maxed out.

you said you have a 600 and 650watt psu's the 600 is the ez-cool your using now, is the 650 any more reliable?
 
Last edited:
yeah looking at the specs for your psu its well under powered as the gtx280 can take up to 20amps from the 12v rail just by its self. what other psu have you got???

I think my other one is even less up to the job - its another 600W but the sticker on it only lists 17A on +12v which is even less (its a Magna ATX600W). Think its going to be memory test anyway out of curiosity and then a new PSU and a decent one at that.

Have just tried running Hells Highway for the first time on this setup and, to be honest, it was dog slow and jerky to the point of being unplayable even down at 1280x800 with medium texture settings etc and I was hoping to be able to play way above that with this setup so will check the memory and then invest in a decent PSU.

Guess if Im honest I have never worried about the PSU really so long as it had the sufficient wattage but now Im running this card Im going to have to look at something more substantial - anyone got any suggestions? Remember I wont ever be overclocking this etc so just something middle of the road but up to the job and which, preferably, wont break the bank?
 
if the game was running but just being really slow that would point even more to lack of power.
lack of power would cause the processor and or graphics card to throttle its self or "cut in and out" causing things to be realy slow or jerky. i would not bother running memtest untill you have a better power supply because at present your system is underpowered which means its not stable. and instability can cause memtest to fail which would imply your ram is broken when in fact its fine. you could try running both "cheapo" power supplies in tandum in order to get more power available but it can be tricky if you dont know what your doing. so if you are going to try that then research it or ask here and ill see what i can do (i have done it a few times in the past, and ill be running dual power supplies in my new build which i will be using a gtx280 maybe 2 lol)
 
if your looking for recommendations on psu's the corsair HX and TX series come highly reccomended as to hiper, pc power and cooling (now taken over by OCz)

i would probably go for the corsair TX 650 or 750, they can be had for £60-£90 depending on where you shop. both will be power full enough for your system
the TX650 has 52amps on the 12v rail and the TX 750 has 60amps on the 12v rail plus if you get them from scan you can get free delivery :p
 
Thanks for all your input CP-PC - have decided am going to go for one of the TX PSU's, there are a couple on Ebay at a good price which I might bid on.

Just as an aside, have just been playing UT3 and I turned the resolution down to 1280x1024 (from 1900x1200 previously or something like that anyway - might have been 1900x1080!) and it didnt do anything - no colour blocking and no issues at all so Im guessing its the PSU and when I lower the requirements it puts less of a strain on everything and therefore there is less of a power requirement put on the PSU?

I will put a TX650 or 750 in hopefully next week when Im back from work and see what happens :smashin:
 

The latest video from AVForums

Is 4K Blu-ray Worth It?
Subscribe to our YouTube channel
Back
Top Bottom