ASUS GeForce RTX 4080 Super TUF OC

[W1zzard]

Priced at $1100, the ASUS GeForce 4080 Super TUF OC features an improved cooling solution that delivers superior performance compared to the FE. The card comes equipped with a factory overclock and dual BIOS. When the quiet BIOS is active, it stands out as one of the quietest cards tested today.

Show full review

 

[P4-630]

So nvidia can stop producing 4080's non-supers now....Since obviously no one would buy those anymore now.....

 

[pat-roner]

Would love to have notes about coil whine in review. I feel that is a big issue from time to time recently

 

[Tigerfox]

So nvidia can stop producing 4080's non-supers now....Since obviously no one would buy those anymore now.....

They've done so months ago.

Why is everyone just repeating NVs white lie? AD103 has 84SM, not 80. It has 7 GPC and every Ada-GPU has 12 SM per GPC. The GPU-block diagramm NV published is nonsense, the GPU can't have 6 GPC with 12 SM each and one with 4. There was a rumour just before the release of the original 4080 stating that AD103 had some silicon flaw that made 4 SM useless. Seems NV couldn't sort that out or didn't bother.

That and the lack of faster memory are the reason why improvement of 4080S is so small.

 

[Deleted member 237269]

Thanks for the review @W1zzard
Should add in the negatives ”One of the biggest Chonky GPU”

 

[W1zzard]

Would love to have notes about coil whine in review. I feel that is a big issue from time to time recently

Coil whine is a function of the graphics card + the PSU + the motherboard. So my coil whine experience is different than yours

 

[xorbe]

Maybe this is nitpicking, but the "latest reviews" headliner on the main TPU page lacks the "OC" label, but this is the OC model (per the article title).

 

[MarsM4N]

Coil whine is a function of the graphics card + the PSU + the motherboard. So my coil whine experience is different than yours

Sure, coil whine depends on many variables, but I think it would still be valuable for users to share your personal experience. Esp. since you have access to the whole stack, and on equal conditions.

 

[W1zzard]

Maybe this is nitpicking, but the "latest reviews" headliner on the main TPU page lacks the "OC" label, but this is the OC model (per the article title).

Yeah the titles are shorted so they fit better on mobile, without wrapping into 3 lines

Sure, coil whine depends on many variables, but I think it would still be valuable for users to share your personal experience. Esp. since you have access to the whole stack, and on equal conditions.

Not easy to measure objectively, I'll look into it

 

[mama]

So, a price cut to the 4080? Is that the gist?

 

[Garrus]

$1701.28 CAD after taxes in Canada. I think people spend too much on video cards, letting NVidia rob everyone. I'm out

 

[MarsM4N]

Not easy to measure objectively, I'll look into it

I think the best would be to use a "Audio Spectrum Analyser" to make the sound signature graphical visible. Noise level doesn't tell the whole story, high pitch frequencies as coil whine are way more annoying than lower pitch frequencies like from the fans, and a spectrum analyser makes these frequencies visible. There are phone apps out there as well as some professional gear & professional software, but I think the Audacity solution (which has the best visualisation I've come across) would fit the task best. And it's FREE.

You just have to use a sound recording of let's say a benchmark run and import it to Audacity. Since coil whine is usually getting worse with higher FPS it would make sense to use a less demanding benchmark that pushes into the 1000FPS territory. Conducting the benchmarks with different frame limits (60, 120, 165, 240, 360, 1000FPS) would also make some practical sense. Coil whine with unlimited frames is pretty usual, but if it's still dominant with a frame limit it becomes a problem.

 

[pat-roner]

Yeah the titles are shorted so they fit better on mobile, without wrapping into 3 lines


Not easy to measure objectively, I'll look into it

I would just love to hear. "no noticeable coilwhine" on my system with this.

 

[mahanddeem]

A question about Far Cry 6 benchmark:
Are High Resolution Textures ON?
Is Ray Tracing (DXR Reflections) ON?
Thanks

 

[Buddha666]

Hello, is there any other difference between TUF 4080super OC and non-OC variant than the boost clock? Like a different cooler, PCB ... and so on?

 

[MarsM4N]

Hello, is there any other difference between TUF 4080super OC and non-OC variant than the boost clock? Like a different cooler, PCB ... and so on?

Looking at this comparison, the only difference should be slightly higher clocks on the "OC" version. Everything else should be the same.

RTX 4080 Super: 2550MHz (Stock-Profile), 2580MHz (OC-Profile) +30MHz Boost (OC-Profile)
RTX 4080 Super OC: 2610MHz (Stock-Profile), 2640MHz (OC-Profile) +90MHz Boost (OC-Profile), +60MHz Boost (Stock-Profile)

 

[Buddha666]

Looking at this comparison, the only difference should be slightly higher clocks on the "OC" version. Everything else should be the same.

RTX 4080 Super: 2550MHz (Stock-Profile), 2580MHz (OC-Profile) +30MHz Boost (OC-Profile)
RTX 4080 Super OC: 2610MHz (Stock-Profile), 2640MHz (OC-Profile) +90MHz Boost (OC-Profile), +60MHz Boost (Stock-Profile)

Thanks for a quick answer, then it makes no sense to pay 100 bucks more. I know the review says that you will get better cooling and dual bios for that raise, but they meant compared to founders I think. So with reference TUF gaming, I'd get the same benefits for the founder price.

 

[W1zzard]

Thanks for a quick answer, then it makes no sense to pay 100 bucks more. I know the review says that you will get better cooling and dual bios for that raise, but they meant compared to founders I think. So with reference TUF gaming, I'd get the same benefits for the founder price.

Correct, this is vs the FE. If you buy the card, let us know if the cooler looks the same (I'd assume so)

 

[DLAV]

Notice the poor judgement when they wrote the fan map for this model and other 4000 series tufs.

If you know the 0-30% fan speed range is not usable, as any value is still 0%, why have 31% at ~1000 RPM?

1000 is too high as a lowest rpm possible. You will always go from fan off straight to 1000, and it's noticeable if the rest of the system is silent. 3000 series tufs had the 1000 RPM map at ~53%, so you could have dead silent fan curves for light loads at 6-700 RPM, in the 40-50% range.

I wish nvidia would lower this 30% min fan speed a bit, if integrators can't have a smarter fan map around it.