BEGINNER'S GUIDE to Overclocking your GPU (Galaxy GTX 670 @ MSI Afterburner)

preview_player
Показать описание
This is a tutorial on how to overclock your AMD or Nvidia Graphics Card (7850 / 7870 / 7950 / 7970 / GTX 660ti 670 680 690 etc.), I specifically am using a Galaxy GTX 670 in this example, but the settings and the methodology is what is important.

ON AMD CARDS settings will be different, I would aim as a safe guide to maybe shoot for around a 10% overclock and personally wouldn't go over 20% for a Graphics card.

ALSO keep in mind that you can probably push your card higher, it is just that Nvidia Kepler Series (6 series) already overclocks the cards pretty high straight out of the box. Another really important thing to remember is test your VRAM using the right test and to also remember that it is the most prone to failure thing on your card, so be extra careful when overclocking it.

AND BE VERY CAREFUL WITH OVERCLOCKING YOUR GPU!! IT IS A LOT MORE SENSITIVE THAN A CPU (BECAUSE A LOT MORE CAN GO WRONG!), SO PROCEED WITH CAUTION AND IF YOU ARE NOT SURE SEND ME A MESSAGE / COMMENT.

The changes are all made through MSI afterburner, your card may or may not have some additional features.

Links for the 2 programs needed are here below (windows 7 64bit, if you are on a different OS / 32-bit version you may need different versions of the following):

MSI Afterburner (configuration program)

MSI Kombustor (stress testing software) -

Keep in mind that your graphics card DOES NOT have to be a MSI graphics card to use these programs (mine is a galaxy). You can use your gigabyte card / asus card / EVGA or any other maker out there.)

Any questions please ask them below or send me a PM and also subscribe to keep the BET going!

And if this helped you then please thumbs up the vid as it in turns helps me to produce more videos and keep content coming!
Рекомендации по теме
Комментарии
Автор

Depends on your temps mainly, at the moment I have scaled back my GPU as it is starting to get really hot where I am. I usually have clocks a bit higher (100mhz on the GPU and only 100mhz on the memory too) and my fan speed on auto (as it does get loud if left at 70%). I usually only switch this on before games.
Ultimately it is up to you, do you want a slight increase with the fan speeds on auto, or do you want to go heavy on the OC with the cost of background noise. The choice is yours :).

techyescity
Автор

Yeah you will see a big improvement over a 550ti. I would just up the TPD to +10% and see what overclocks you get from there. Nvidia 6 series is really good in that the boost feature works amazingly well.

techyescity
Автор

You can reinstall the heatsink fan with some arctic silver 5, that would help, can also purchase an aftermarket cooling solution as well.

techyescity
Автор

Depends how frequent they are, if you are noticing an occassional one then you are near your max, slow it down a tad, if your card is glitching a lot then I would back it down quite a bit.

techyescity
Автор

Life span depends on a few things, but yes overclocking (higher temps mainly) will reduce a graphic cards lifespan. I would mainly watch the memory clocks, as the onboard gddr5 memory is usually the first and most common thing to poop itself on graphics cards.

techyescity
Автор

I want to thank you again Bryan for another amazing video. I was having some temp problems (70°+) at stock clock and I never even thought of overclocking because I didn't want my temps to go higher. Turns out my GPU fans were staying at 30, and once I bumped that up to 69 it immediately dropped to 61° at overclock speeds. Now I am at a boost clock of 1125 and am seeing around 1180+ in game and in test. Thanks again, happy I subbed.

lilturk
Автор

Generally any Z77 board will support overclocking, just the more expensive you get the more features you can. I solid board for the price is the extreme4, I have it and its an amazing board.

techyescity
Автор

Yeah if the actual heatsink doesn't get hot but the GPU itself in Kombutsor is hot than that can indicate a poor connection between the heatsink and the GPU (I have had it 2 times in the past its not too uncommon). I also read reviews about that Asus version having weird problems too, may wish to check it out and if you have those problems just RMA the card.

techyescity
Автор

should be able to go into the options and unlock both voltage and power limit. on a laptop though you should tread carefully.

techyescity
Автор

Up to you, if you can handle the extra noise from the fans then why not.

techyescity
Автор

I am pretty sure the 7 series has boost clock as well? Maybe try go for a clock speed of 1000 and make sure fan speeds are at 60%+ to be safe.

techyescity
Автор

I like to keep it safe especially when doing this guide for people who are learning, but love the enthusiasm lol.

techyescity
Автор

Oh, well on your DIMM slots how many sticks of ram do you have inserted? If one then you are out of luck and will have to get a friend to test his ram on your board. If you have 2 sticks in there just take out the one to the right and then if it works fine you know that stick is faulty, if it is still playing up switch the stick you pulled out with the other stick and check if it is all fine.

techyescity
Автор

Yeah thats pretty much it, but TDP is neccessary in this case cause for the GPU and CPU they all have a "sweet spot" and its just a mechanism that helps keep everything at the "sweet spot", when putting the better aftermarket cooling solutions on you are able to find your new "sweet spot". TDP is just a number that states this is the maximum your part will use when its stressed and if using more power we will throttle the badboy down for you, maybe think of it more so as a "heat safety limiter".

techyescity
Автор

A difference none the less!! Great to hear!

techyescity
Автор

Weird... I couldn't help you there, I am guessing it may have to do with the GPU's bios revision? as that is usually what blocks programs from performing certain things. Just weird how you can't even access the 114%...

techyescity
Автор

Have to give us more specs, like mobo / cpu / os. But a full reinstall is the first thing I recommend, then after that see if the problem is still there.

techyescity
Автор

Nah if the PSU is the problem your card will do really weird things like cut out. I would say try 15%? Is it an OC model with an aftermarket cooling solution or just the reference cooler? Also sometimes the GPU can have a bad contact with the heatsink so you may wish to reseat the heatsink with some proper thermal paste.

techyescity
Автор

What psu do you have? Could be a few things, the heatsink could not be seated on the GPU properly, also might just be a bad model, or could be the psu not giving enough power to the GPU, but if you underclock it and it works it would indicate that it is most likely the GPU overheating.

techyescity
Автор

Thermal Design Power is just a number given by the manufacturers (in this case Nvidia) that stipulates what the card will most likely max out at (in terms of power draw) when being stressed. I believe it is 170W for the GTX670, so when you change the limit you are essentially saying "my card can draw up to 204W (if 120% limit is set)", I think the GTX680 has a TDP of 195W so increasing this on the GTX670 could (most the time it won't) make a difference.

techyescity