GTX 1660 Super Overview and Testing in Mining Crypt Currency

Released in November 2019, the new product from NVidia GTX 1660 Super should replace the previous model GTX 1660. In fact, it is a minor upgrade of the previous model, with the replacement of the outdated GDDR5 memory by the high-speed GDDR6. The graphics card is based on TU116 chip with 1400 CUDA cores and 22 multiprocessor modules. The frequency of the core remained at the level of the younger model – 1500/1800 MHz, but the memory speed reached as much as 14 GB / sec, which is even higher than the GTX 1660 Ti – 12 GB / sec. Such a high-speed memory should have a positive impact on the result of 1660 Super in the mining algorithms that load video memory. So far, NVidia is going to sell GTX 1660, 1660 Super and Ti version at the same time, so a decrease in prices for the models of this line is not expected. Before we consider the GTX 1660 Super Mining results, we will compare the novelty with other models of the ruler.

GTX 1660 Super Mining

Comparison of the GTX 1660 ruler

GTX 1660 Super Mining

Based on the comparison table, we can see that the memory bandwidth of the novelty is even higher than GTX 1660 TI and catching up in this component RTX 2060 with a price tag of 350$. We can assume that in the algorithms that load 1660 Super Mining video memory performance will exceed the level of GTX 1660 Ti and inferior to it in the GPU dependent algorithms. Improved memory subsystem performance should also have a good impact on overall adapter performance. NVidia claims a 20% improvement over base 1660. Let’s start testing GeForce GTX 1660 Super Mining in popular algorithms. Also, you can find out the results of mining tests for gtx 1660 and 1660 ti here.

GTX 1660 Super Mining Results

It should be noted at once that for the purity of the experiment all the tests were performed at the factory settings of memory and kernel frequencies.

Algorithm

Mainer

Speed

BeamV2

GminerCuda9.0

25.064 Sol/s

Zhash

GminerCuda9.0

37.240 Sol/s

(Grin) Cuckarood29

GminerCuda9.0

5.204 G/s

X16r

ZEnemy

13.454 MH/s

X16rV2

ZEnemy

14.652 MH/s

Lyra2Rev3

CCminerTpruvot

34.537 MH/s

CuckooCicle

NBminer

4.309 G/s

DaggerHashimito+Blake2s

ClaymoreDual

26.031MH/s

BeamV2

LolMinerBeam

19.437 Sol/s

Lyra2Z

Trex

2.557 MH/s

As 1660 Super Mining tests showed, our assumptions were correct and in memory-dependent algorithms the novelty bypasses not only GTX 1660, but also 1660 Ti. For example, in Ethash 1660 Super achieves 26.3 MH/s, which is more than in 1660 Ti, gaining 24 MH/s. Also note that in the Beam HashV2 algorithm, the Super version almost catches up with TI, gaining 25.7 MH/s against 26.8 MH/s. In the algorithm (Grin) Cuckarood29 NVidia 1660 super mining shows the same 1660 Ti result.

Despite the use of a truncated TU116 chip, GTX 1660 Super shows similar results to 1660 Ti in GPU dependent algorithms. Why is it that a $230 graphics card is not inferior to a $280 adapter? Perhaps NVidia has optimized the chip for GDDR6 memory and this has improved overall performance. To be honest, we are only glad that the GTX 1660 Super test in mining showed such results.

A small result

Video card GTX 1660 Super pleasantly surprised us with its price / performance ratio. The adapter costs $50 cheaper than its older version, while showing similar results. At the time of writing, rtx 1066 super in mining becomes the optimal solution when building a budget farm to work with VRam dependent algorithms. Based on the current course, GTX 1066 Super will generate $14 income per month, some of which will go to pay for electricity. Knowing the average electricity price and video card cost, the payback will come in 22-23 months. This figure may change depending on the rate of the crypt currency. It is also possible to accelerate production by 7-10% by accelerating.