ASUS GeForce GTX 1660 Ti STRIX OC 6 GB

[W1zzard]

The ASUS GTX 1660 Ti STRIX OC is the company's flagship GTX 1660 Ti variant. It comes with a large triple-fan, triple-slot thermal solution that delivers better temperature and noise levels than any other GTX 1660 Ti we have tested so far. Also included is a backplate, RGB illumination, dual BIOS, and fan headers that synchronize with the GPU's fan speed.

Show full review

 

[jabbadap]

Not a bad card... But I would rather buy rtx2060FE for $20 more, which have game bundle too.

 

[Animalpak]

This card is the 1 billion time confirm that people around the world will still play at 1080p for a looooooong loooooong time.

 

[chaosmassive]

this card so expensive that RTX 2060 has 107% more perf/dollar, lmao

 

[M2B]

This card is the 1 billion time confirm that people around the world will still play at 1080p for a looooooong loooooong time.

If ray tracing is going to be a serious thing in game industry then it might actually take even longer than a looooooong time to move from 1080p.
but honestly, 1080p with proper Anti-Aliasing is good enough for PC Gaming; Considering the average size of monitors.

 

[kastriot]

This is great card for 200$ but until nvidia gets some hammering competition they will always be overpriced.

 

[thebluebumblebee]

What's with all the 3's?
$330
300 mm (really!?)
3 fans (seriously?)
3 slots

 

[Auxois]

W1zzard, it's time to ditch the Heaven Benchmark, especially using it to benchmark the effects of overclocking.

The benchmark has been flawed for years. Kepler performed way better than Maxwell, so you'd have 780's outperforming 970's and the same goes with Pascal. And yet again Pascal is outperforming Turing - as can be seen in the performance summary, where the 1070 is behind the 1660 Ti Strix over a vast number of games, yet even overclocked the 1660 Ti Strix is struggling to keep up with a stock 1070 in Heaven.

I've used this wonderful site for years and always used the performance summary charts to compare GPU's, but the use of Heaven has got to stop, as it's obviously very outdated and doesn't provide reliable data, rendering the OC part of the review useless if not factually wrong. You'd find that even 3D Mark Timespy is more modern and a much better benchmark overall, being DX12 and all.

Just registered to post this and I hope that you'll take it into consideration

 

[Deleted member 158293]

Nice PCB, but that's a lot of money for not much GPU...

 

[W1zzard]

W1zzard, it's time to ditch the Heaven Benchmark, especially using it to benchmark the effects of overclocking.

Just using it to test relative performance gained from overclocking seems fine? The testing is not about the actual performance delivered.

I'm also not testing the full run, but only a subset that puts proper load on the GPU

Just to clarify: the Heaven scores are not included in performance summary

Edit: I modded Heaven to detect OC instability, which makes my life MUCH easier Saves about 1 hour for each review

 

[FreedomEclipse]

Pitty no SLi

 

[Voluman]

Thanks for the review, great as always.

At first page the table says 1536 cuda cores, but in the text there is 1408 cuda cores.
"NVIDIA carved the GTX 1660 out of the "TU116" silicon by disabling 2 out of 24 streaming multiprocessors, resulting in a CUDA core count of 1,408 and 88 TMUs, which is still higher than what the "Pascal" based GTX 1060 6 GB packs. With 48 ROPs and a 192-bit GDDR5 memory bus driving 6 GB of memory, the rendering and memory subsystem is practically carried over."

I guess you paste the text from wrong article (1660), not from the 1660Ti-s.