• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

[HELP] Overclocking MSI GTX 750 Gaming

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
If you are gaming and only reaching 50C at worst, you are fine...


That is a 750ti... two different cards... Kepler and Maxwell, right?

youre right on that, they are of different cards but the description on the TI version is what exactly is doing on my part like max temps between 50-52 celsius and fan strikes at 32-33% and mine stays at 32%.
 
Judging by that review, I would say you'll be in the neighborhood of 1100-1125Mhz on the overclock (and that will "Boost" higher under load).

Start by setting your core to 1075, then go in 10-20Mhz increments from there.
 
I did a quick google search and didn't see many (any?) reviews for the 750 non-Ti. I figured they were similar.



For overclocking stability, it's all about what you're comfortable with. What I mean by that: when I find my card's "max", I usually back off 20-40Mhz. That extra little bit isn't going to yield you much in the way of performance gains, but it could prevent an in-game crash (which isn't destructive, but is annoying).

You can run a benchmark at each clock speed. That's one method. What I've been doing is just keeping a high-intensity game running (I use BF4). Then I'll gradually increase the core clock out and continue playing (looking for any signs of artifacting or a game/system crash - it's usually just a game crash). This is a little less time-consuming, I think, but do note that I've seen instances, especially when you get to the memory overclocking portion, where I can increase the speed while the game is running and not have issues, but if I increase the speed and then load the game after, I'll get a crash. Presumably, the vram is being stressed more at load time?

Well, that is very informative. Thanks! Then i guess, ill have to try the two testings then.
 
Which 3DMark? You'd want one of the newest ones (which also aren't small downloads) to make sure you get full load.
 
Judging by that review, I would say you'll be in the neighborhood of 1100-1125Mhz on the overclock (and that will "Boost" higher under load).

Start by setting your core to 1075, then go in 10-20Mhz increments from there.

Thanks much with all the info, ill have to check on it by now. See you all and thanks for the response. Till then...

- - - Updated - - -

Which 3DMark? You'd want one of the newest ones (which also aren't small downloads) to make sure you get full load.

Oh well, i just check out their site and youre right. There also big so ill just go with heaven as its a mainstream. Thanks buddy! c yah...
 
Thanks much with all the info, ill have to check on it by now. See you all and thanks for the response. Till then...

- - - Updated - - -



Oh well, i just check out their site and youre right. There also big so ill just go with heaven as its a mainstream. Thanks buddy! c yah...

Doesn't really matter, but I think Firestrike (part of the 3DMark suite), might actually be more mainstream? Again, unless you're benchmarking for fun (and not just to test stability), either is fine :thup:

P.S. Good job finding the edit button on your own ;)
 
Doesn't really matter, but I think Firestrike (part of the 3DMark suite), might actually be more mainstream? Again, unless you're benchmarking for fun (and not just to test stability), either is fine :thup:

P.S. Good job finding the edit button on your own ;)

Heaven (to me) has always been easier to run certain settings on. You don't have to pick different benchmarks to run for different settings, you just pick your res/quality and go.
 
+1 There's almost zero ability to adjust settings in the free version of 3DMark (unless I missed it, which is very possible as I didn't really care to run anything but the default for comparison's sake :))
 
One more question guys before i leave. Im curious, when im running some test like CPU or GPU. Do i need to close everything out like just run the test software and any monitoring software like HW monitor?
 
One more question guys before i leave. Im curious, when im running some test like CPU or GPU. Do i need to close everything out like just run the test software and any monitoring software like HW monitor?

It won't matter unless you're trying to get a top score in benchmarking.
If you're just testing an overclock, I wouldn't bother with closing stuff out. (unless it's something you could lose progress in if you crash the system)
 
Doesn't really matter, but I think Firestrike (part of the 3DMark suite), might actually be more mainstream? Again, unless you're benchmarking for fun (and not just to test stability), either is fine :thup:

P.S. Good job finding the edit button on your own ;)

Im overclocking for a mixed of stability test and know how. But i think this is also fun :D
 
One more question guys before i leave. Im curious, when im running some test like CPU or GPU. Do i need to close everything out like just run the test software and any monitoring software like HW monitor?

It won't matter unless you're trying to get a top score in benchmarking.
If you're just testing an overclock, I wouldn't bother with closing stuff out. (unless it's something you could lose progress in if you crash the system)

+1 It might even be more realistic to leave some stuff running (browser, music...) since that's what you'll have running in the background normally anyway. :thup:

Have fun :)


EDIT: As an asterisk to this, for GPU overclocking, you can leave stuff running in the background (it really shouldn't matter as those apps shouldn't be applying GPU load). With CPU overclocking, it might be better to kill that stuff, after all. The stress testing apps (Prime95, ...) will apply 100% CPU, so it might not be realistic to have other apps running, which unlike in the GPU overclocking scenario, will be applying additional CPU load. Games typically only put 50-75% stress on the CPU, so there's some headroom there for background processes which isn't available while stress-testing that component.
 
Last edited:
It won't matter unless you're trying to get a top score in benchmarking.
If you're just testing an overclock, I wouldn't bother with closing stuff out. (unless it's something you could lose progress in if you crash the system)

i see. im thinking some stuff could cause the test to fail or something but THANKS again! :thup:
 
+1 It might even be more realistic to leave some stuff running (browser, music...) since that's what you'll have running in the background normally anyway. :thup:

Have fun :)

Yeah, thanks. GLHF! you guys are great. Not just in mind but also in heart (helping people out) :D
 
Guys if youre still there, i got 1 more question. Is there an interval of testings like after one test, let your gpu cooldown for how many minutes till temp goes back to minimum and try again? Or it doesnt matter?

Also, do i need to max out the settings on Heaven like turn on max AA and other settings to test better? Also does resolution does matters and making it fullscreen or not?
 
Nope. Remember, you're shooting for realistic testing. You won't let your GPU cool down between maps when you're playing, right? ;)

I don't think you'll get anywhere near the thermal limits of your card, but once you do find the max core clock, you will want to play for a good 30 minutes+. Temps will gradually increase over that time and eventually level out when the ambient temperature in your case reaches an equilibrium (unless you have zero airflow, in which case, temps will just keep going up until your system crashes :eek:).
 
Nope. Remember, you're shooting for realistic testing. You won't let your GPU cool down between maps when you're playing, right? ;)

I don't think you'll get anywhere near the thermal limits of your card, but once you do find the max core clock, you will want to play for a good 30 minutes+. Temps will gradually increase over that time and eventually level out when the ambient temperature in your case reaches an equilibrium (unless you have zero airflow, in which case, temps will just keep going up until your system crashes :eek:).

ok thanks. how about my 2nd question. Maybe you havent noticed that since i just edited it a while ago.

"Also, do i need to max out the settings on Heaven like turn on max AA and other settings to test better? Also does resolution does matters and making it fullscreen or not?"
 
I would go ahead and max out the settings (unless you want to compare with other peoples' scores and they have some particular config).

Definitely make it fullscreen and I would use the native resolution of your monitor.
 
I would go ahead and max out the settings (unless you want to compare with other peoples' scores and they have some particular config).

Definitely make it fullscreen and I would use the native resolution of your monitor.

Thank you again :thup:
 
Just did overclocking/testing on Core Clock and looks like the safest i can add is +200. After i made it +240, it almost crashed after i click RUN on Heaven but after 4-5 seconds, it displayed and i could have run the benchmark again but i just considered it a limit already if im right on this.

On the upper right portion of Heaven screen on my last testing with +200 Core Clock, these are the info on the ff:

Graphics: 1533 MHz
Memory: 2505 MHz

Hope im doing it right :eek:
 
Back