Overclock msi gt 710 2gb GDDR5

@digiquandry
Your post contains valid concerns and good arguments.

Anyway it applies generally for OCing GPUs, not particularly to old now low end models.

What has applied 3 years ago still applies today. I mean people trying to OC a three years old GPU still have to consider today what they had to 3 years ago. And the side effects do not change only because it has got low performance compared to current models now.
I also build own PCs since 1997, but a decision to let run a device above of normal specifications AKA OCing is an own decision and answers the question: What's possible with MY GPU and is it enough, can I live with the fact running it above normal parameters, can I live with side effects, or do I need a new GPU?

I am sorry this bit is out of context to my post. By that you say that I instruct him to melt his computer!
My post is made with responsibility quoting same interests and results from another source. (Title: I have a Nvidia GT 710 and I want to squeeze more out of it.)
The benefit / success of OC'ing depends on many various factors. Also on own unique GPU.

And to keep it running in the range of 60-70 degrees Celsius (alarming limit 80) which I have posted is a moderate, responsible and safe advice I'd give to any good friend who is curious for OC results. To speak of melting here is far off. The game graphics either freeze or get distorted far before one has reached 'melting'. And there is also throttling far before reaching max temp of specification which is 95 degrees (another plus 15 to the posted alarming max).

In summer time my GPU is 78 without even OCed.

Sure, if the GPU dies while having it OC'ed there are thoughts that OCing has caused its death.
Giving advice to OC actually never turns out good for the advisor. I have no own benefit, but I can share experiences for how to.

My post is based on own practical experiences and experiences of friends. I do the same.
I am more like..if you want to know something, then you have to make own (reasonable) experiences to make REAL decisions.

Anyway here again a summary to avoid any misunderstandings:

OCing stresses the device and might shorten life time.
Keeping moderate posted temps does not negatively affect the GPU at least I don't know any practical example. Monitoring temperatures especially at hot days is essential, though.
Check proper air flow / cooling especially the dust filters if PC case is some years old already.
Do not enforce OCing to ultra limits. Keep voltages at normal.
Be careful. Some RAM speeds values are doubled since we talk about double data rates (DDR RAM). In fact the real clock is raised half of the value!
Reset to defaults whenever there is no gaming.
If it freezes try to exit frozen task instead of hard power off / reset to avoid I/O data loss (HDD/SSD r/w action while hitting reset button)
Use latest GPU driver.

Consider a GPU upgrade for relaxed gaming experience.

The actual question is: Does OCing make any real sense at all?
And that can be answered controversially.

I mostly run on normal clock speeds of course. I am no OC freak. My CPU is slightly OCed since purchase though (works better with my RAM speeds)

 

Yes.
Next time (new driver release) you might want to choose custom installation. It lets you check which components to install (and also if you want a clean installation removing old driver / components and settings first).
Saying that you surely know that you can play with official Nvidia settings to optimize particular games. Some features such as triple buffering might be useful to get a few more frames.
But here also you have to test, but within official features / parameters.

I respect MJs opinion. Running devices as they are is reasonable and literally without extra stress.

Anyway I want to add that to OC or not is no black / white dualism. It is a probability. The more you OC the more probable negative effects.

Each Nvidia model comes with a reference design. And the OEMs then make their own. (Asus, MSI, Gainward, Zotac, etc, etc).
Already there are different clock speeds for the exactly same model released! The difference is that it's covered by warranty, while you lose it doing own additional OC.

At chip production there is QC and QA (Quality Control and Quality Assurance). And mostly one batch of product is used to be released with different clock speeds. (Especially at CPUs where many are 'downgraded' to fit current market demands).
OC related sites try to inform customers about and there are OC freaks having their hobby.

The more you OC the more probable negative effects is anyway true anytime. The direct relation is temperature / thermal damage.

 

I don't think it'd be worth the efforts. OP said he's got game freezing already running below 70 degrees.
It would not add extra room for OC.
Finally he's gained knowledge by making an own test.

 

I agree with that now, too. But not because of the given arguments. I agree because of his own OC results.

I was first confused, because I posted he should enable custom fan speed / temp curve and he did not say he can't.
I checked then the particular MSI model it has passive cooling only.
70 degrees are still normal, though. Getting here already freezing made me thinking an additional fan does not help either.

Out of curiosity I just checked my GPU running Obduction (which is very GPU intense). Without OC it's 73 today. (Warm spring day).
Enabling custom curve lets it drop significantly.

 

I don't think when it came to the topic it was about to be personally right and not in the least to win a race with a low spec GPU.
It's about to offer an opportunity where the one can practically experience what's possible at a special hardware / game combination. (It's the purpose of trying to OC = thread title of OP)

It could have worked that he gains the missing few frames and he could be satisfied. But obviously it didn't.