Do I just have a crappy GTX460? OC results suck

[nyker96]

I think the "average" is about 850 or so. some will go above and some will go below that. Yours is unfortunately below, even know cyclone is one of the best coolers out there. My personal GTX460 is also a soso sample, it's doing 869@1.05v which is soso. It's max is about 895@1.087. the better samples are doing 950s@1.087 I believe. But believe me, 830-850 is still a substantial OC compare to base of 625. pretty much a 1/3 OC, I'd say formidable by any measure.

 

[bryanW1995]

base was 675, but they ended up raising it to 720 b/c the yields were so good.

 

[MrK6]

A few things, IDC -

1) That sucks, sorry man XD.

2) That's a heck of an odd crash (full reboot) for it to be the graphics card's fault. Granted I haven't been on NVIDIA drives in a couple of years, but usually the should recover or BSOD at most if you're pushing the clocks too high. Have you tried other programs to stress test? How's gaming (if you're using the card for gaming)? How far can you push the card in games?

3) What PSU are you using and how much do you have plugged into it? Try unplugging non-essentials (extra HDD's, peripherals, the works) and re-running OCCT - see if it holds up with a lighter load on the rails.

4) Cooling doesn't look to be the problem at all. I'd say try more voltage with the 1.212V BIOS hack, but if the PSU is your problem then that could do more harm than good.

If it was me, I'd scrap OCCT GPU test and use actual programs/games to test for awhile. These synthetic programs/stress testers have their place, but I would never use just one program as an end-all (despite how awesome and easy that would be XD).

 

[Castiel]

.

That looks freaking amazing IDC

 

[Idontcare]

An update on some of the outstanding action items:

Testing w/only one LCD: I unplugged the second screen and reran the stability checks, my temps were unchanged (within a degree or two) and the rig locked per usual.

By "locked" I mean it sends the LCD into power-saving mode and then I am blind, can't get the screen to rewake unless I do a reboot.

A few things, IDC -

1) That sucks, sorry man XD.

2) That's a heck of an odd crash (full reboot) for it to be the graphics card's fault. Granted I haven't been on NVIDIA drives in a couple of years, but usually the should recover or BSOD at most if you're pushing the clocks too high. Have you tried other programs to stress test? How's gaming (if you're using the card for gaming)? How far can you push the card in games?

3) What PSU are you using and how much do you have plugged into it? Try unplugging non-essentials (extra HDD's, peripherals, the works) and re-running OCCT - see if it holds up with a lighter load on the rails.

4) Cooling doesn't look to be the problem at all. I'd say try more voltage with the 1.212V BIOS hack, but if the PSU is your problem then that could do more harm than good.

If it was me, I'd scrap OCCT GPU test and use actual programs/games to test for awhile. These synthetic programs/stress testers have their place, but I would never use just one program as an end-all (despite how awesome and easy that would be XD).

I thought I had good contact between the GPU and the cooler, my temps were around 60C and when I'd touch the fins of the GPU HSF they were definitely HOT.

But I went ahead and redid the TIM, using AS5 this time just to be sure.

I also reconfigured my case layout, the Black Pearl is one of the those cases that flips the mobo position so the bottom slots are actually positioned at the top of the case. I put my GPU in the top slot and the case fans in the top of the case push air into the intake fans on the GPU.

All else being the same, my loaded temps are now ~11C cooler. (running 49C vs 60C at stock overclock with OCCT)

Haven't tried OC'ing yet to see where my extra thermal budget takes me.

I did start a thread in the PSU forum to find a new PSU to buy just on the chance that my PSU is dying on me, I bought it in Nov 2006.

 

[MrK6]

By "locked" I mean it sends the LCD into power-saving mode and then I am blind, can't get the screen to rewake unless I do a reboot.

Ah, that sounds more graphics card error-like. I misunderstood and thought the PC itself was just rebooting. I see you have a 620HX PSU and they're darn good. That actually makes me move PSU issues lower on the differential XD. Also, Corsair has a 5 year warranty IIRC, so if it does turn out to be the problem, it might be worth it to RMA it and save some cash.

All else being the same, my loaded temps are now ~11C cooler. (running 49C vs 60C at stock overclock with OCCT)

Haven't tried OC'ing yet to see where my extra thermal budget takes me.

I'd be interested to see where it does as well.

I saw the discussion about VRM cooling, but have you checked the temps on your VRM's? I use GPU-Z and just run a log while I stress test. Even though you have active cooling on them, it might not be enough, especially in something like a stress test (under gaming they'd be much cooler). If your VRM's are getting too hot, that'd be enough to shut down your card.

 

[OILFIELDTRASH]

I picked up an evga superclocked external exhaust gtx 460 off craigslist today to replace my 9800gtx I used for a day. Lol. Anyways its at a stock voltage of 1.000v and I could only get it to 831mhz on the core. Not interested in adding voltage. I'm perfectly happy at the stock clock of the card at 765mhz. Max temp is 72c non overclocked and 74c overclocked to 831mhz. Is the temps normal? It's idling at around 35c.

 

[Idontcare]

I'd be interested to see where it does as well.

I saw the discussion about VRM cooling, but have you checked the temps on your VRM's? I use GPU-Z and just run a log while I stress test. Even though you have active cooling on them, it might not be enough, especially in something like a stress test (under gaming they'd be much cooler). If your VRM's are getting too hot, that'd be enough to shut down your card.

It gained me nearly 20MHz.

Originally I could do 836MHz at 1.087V and I'd hit 70C doing it (OCCT stable).

Now I can do 854MHz at 1.087V and my temps only hit 60C.

I'm not stable going any higher than 854MHz despite my temps being only 60C, but at least my temps have come to be in alignment with expectation.

I have not measure the temps on the vrms but I touched the backside of the card where the vrms are located and they be friggen hot!

 

[Idontcare]

Max temp is 72c non overclocked and 74c overclocked to 831mhz. Is the temps normal? It's idling at around 35c.

I don't know what is normal but my card idles at 22C with this AC cooler and full load is 60C.

Your clockspeed of 831MHz at 1V seems about right. I can't clock that high with only 1V but my temps are lower, so that says your GPU is definitely better silicon than mine. Be happy

 

[OILFIELDTRASH]

I don't know what is normal but my card idles at 22C with this AC cooler and full load is 60C.

Your clockspeed of 831MHz at 1V seems about right. I can't clock that high with only 1V but my temps are lower, so that says your GPU is definitely better silicon than mine. Be happy

Yeah I'm very happy. I'll just leave it at about 820mhz and call it good. I don't want it working too hard to keep it cool so 820mhz should be plenty. It's nearing gtx 470 performance at 820mhz. For $145 cash you can't beat that. I bet your 854mhz is dead even with a gtx 470. Have you run 3d mark vantage? I'd be interested to see the difference in score from my 831 to your 854mhz.

 

[Jules]

Not all reviews on O/c 460's are so rosey.

http://translate.google.com/transla...809-5/comparatif-14-geforce-gtx-460-1-go.html

Read this (if you can interpret through google translate). Don't be in too much of a hurry to push cards so hard.

IDC, you have a far superior cooler to a reference one, but this is good info in general, IMO.

Interesting article, Thanks

 

[BD231]

One other thing I'll mention about the AC cooler for anyone thinking of getting it for their GTX460 is that the mounting bracket that goes over the GPU is too big and if you glue on your memory heatsinks per the directions then you won't be left with enough clearance to get the GPU bracket in place.

I recommend you install the GPU brack first, then proceed to glue your memory heatsinks.

The only difference I can see is in your VRM cooling, I have one solid piece covering them all but I suppose three piece's would do just as well? Might have a bit more mass to your sink if you can find one that fits across all of them. Certainly glad you took that thermal gunk off, anything solid at room temperature is no good in my books but that's just my preference. I've tried getting my fan to lock in place but not even afterburner will do it for me, I'd have to change the fan profiles in my bios to get the fans to stop doing what they please. At best I can lower my max fan speed.

Did you get any headroom out of the memory with the ram sinks?? Its been a while since I did any research on BGA chips but last I heard they don't typically run hot at all and the thought of putting glue on them has me wondering if there's any benefit .

So far I've gotten *nothing* out of my memory overclocking, card just locks up at even a simple 50 added mhz.

 

[Idontcare]

The only difference I can see is in your VRM cooling, I have one solid piece covering them all but I suppose three piece's would do just as well? Might have a bit more mass to your sink if you can find one that fits across all of them. Certainly glad you took that thermal gunk off, anything solid at room temperature is no good in my books but that's just my preference. I've tried getting my fan to lock in place but not even afterburner will do it for me, I'd have to change the fan profiles in my bios to get the fans to stop doing what they please. At best I can lower my max fan speed.

Did you get any headroom out of the memory with the ram sinks?? Its been a while since I did any research on BGA chips but last I heard they don't typically run hot at all and the thought of putting glue on them has me wondering if there's any benefit .

So far I've gotten *nothing* out of my memory overclocking, card just locks up at even a simple 50 added mhz.

Yeah I just followed the directions, figured the guys who designed it must have had their reasons for doing it the way they did it.

I have not touched OC my memory either, gaming wise the most challenging thing the card is seeing at this time is Dragon Origins at 1920x1200. I turned up all the eye-candy (all 3 options for it anyway, lol)...gotta admit I really thought the AA would make the jaggies look good and all that, not really impressed. Maybe I'm doing something wrong.

Why do you suppose you can't control your fans like I can?

 

[SickBeast]

Yeah I just followed the directions, figured the guys who designed it must have had their reasons for doing it the way they did it.

I have not touched OC my memory either, gaming wise the most challenging thing the card is seeing at this time is Dragon Origins at 1920x1200. I turned up all the eye-candy (all 3 options for it anyway, lol)...gotta admit I really thought the AA would make the jaggies look good and all that, not really impressed. Maybe I'm doing something wrong.

Why do you suppose you can't control your fans like I can?

I recommend using 16x AA in the drivers along with multisample transparency AA. It looks really nice. Make sure you have the textures set to "high quality" as well, and disable the anisotropic optimizations. I've also had good luck with ramping my fan speed up to 80% once the temperature hits 50C.

If you can't run at those settings, just bump the AA down to 2X and you should be good. That's usually only the case if you run out of texture memory though, which is rare at 1920x1200 or lower.

 

[andrei3333]

My EVGA GTX 460 1GB @ 865MHz, card originally shipped with 763MHz all stock voltage

I think its weird that your OC of 830 is the same maximum on two different cards, but at the same time, they are from the same manufacturer and the same model. I think cards are binned and maybe you got two that just ended up having very similar limits.

Also you did not need to replace the cooler, that original cooler is a great one. Just leave it as is and enjoy the card. you wont even see a difference in going from 830 to 850 or so, so whats the point ?

 

[BD231]

Yeah I just followed the directions, figured the guys who designed it must have had their reasons for doing it the way they did it.

I have not touched OC my memory either, gaming wise the most challenging thing the card is seeing at this time is Dragon Origins at 1920x1200. I turned up all the eye-candy (all 3 options for it anyway, lol)...gotta admit I really thought the AA would make the jaggies look good and all that, not really impressed. Maybe I'm doing something wrong.

Why do you suppose you can't control your fans like I can?

Yeah I remember dragon age, GREAT, you've reminded me I've gotta install that one to see what it looks like on the 460. I've only played it on an hd4850 and judging from the jump in graphics quality I'm seeing in every game I have I'm guessing that game is going to look quite a bit better. Already tried metro 2033 vs the 4800 series and *wow* what a difference this card makes.

One thing I do notice like you is the AA, I don't like nVidias options past 2x. Older games fair better but I think there's something funky about nvidia's AA. Its not quite as effective it seems. I notice a lot less texture poping and a lot more detail in 3D though. Don't see a difference in 2D at all. You might try disabling the in game AA and turning on nVidia's AA in the control panel. See what happens when you turn on 32x aa or 16xQ.

Not sure about my fans, when I leave them on auto they don't spin up at all with this new cooler. Odd fan behavior for me.

 

[3DVagabond]

Interesting article, Thanks

Glad to see someone read it.