Tuesday, October 29th 2019
NVIDIA Releases GeForce GTX 1660 SUPER Graphics Card: GDDR6 Makes A World of Difference
NVIDIA today released the GeForce GTX 1660 SUPER graphics card at USD $229, just $10 more than the GTX 1660. The card has identical specifications to the GTX 1660, with 1,408 CUDA cores, 88 TMUs, and 48 ROPs. It even has the same GPU clock speeds, at 1530 MHz core with 1750 MHz GPU Boost. The not-so-secret sauce lending it a major performance boost is memory - GDDR6. Armed with 6 GB of 14 Gbps GDDR6 memory, faster than even what the GTX 1660 Ti ships with, NVIDIA is able to shore up performance of the GTX 1660 by near two figures. The GTX 1660 Super is probably designed to preempt AMD's Radeon RX 5500 series. A purely partner-driven launch, the GTX 1660 SUPER is available from today, in custom-design boards from nearly all NVIDIA partners.
We've reviewed five GeForce GTX 1660 Super cards today: ASUS Phoenix, Gigabyte Gaming OC, MSI Gaming X, Palit GamingPro OC and Zotac AMP.
We've reviewed five GeForce GTX 1660 Super cards today: ASUS Phoenix, Gigabyte Gaming OC, MSI Gaming X, Palit GamingPro OC and Zotac AMP.
32 Comments on NVIDIA Releases GeForce GTX 1660 SUPER Graphics Card: GDDR6 Makes A World of Difference
for Intel also; this should lower some prices some more. I'm just glad the over inflated "mining craze" gpu prices have gone close to normal again.
Price/performance boost - even a fairly minor one - is welcome though, I suppose.
The best part is that Nvidia knows exactly what the new AMD cards will bring and they are preemptively improving their performance/cost ratio in response. Should be good to have a little more competition.
Also, Turing is not exactly ancient tech. Leaving the RT and tensor cores aside, the shaders are quite different from Pascal (not redesigned from the ground up, but redesigned nonetheless): www.anandtech.com/show/13282/nvidia-turing-architecture-deep-dive/4
Other than not carrying our Windows conclusions about a GPU to Linux or the other way around, there's not much that we can meaningfully infer from those results.