The GeForce GTX 1060 launched way back in the Summer of 2016, as part of NVIDIA's bid to extend its then squeaky new "Pascal" GPU architecture to the value-conscious sub-$300 market. Since then, it has become one of the company's best selling graphics cards targeting the vast majority of PC gamers who just want to play at Full HD (1080p) resolution with high-to-ultra details. NVIDIA has refreshed the 6 GB variant of GTX 1060 from time to time. When AMD introduced the Radeon RX 580 in Spring 2017, NVIDIA responded with the GTX 1060 6 GB 9 Gbps, a limited-edition SKU with faster memory. AMD's recent introduction of the Radeon RX 590 to rake in Holiday 2018 sales has triggered another response from NVIDIA.
If you've been keeping up with our tech news surrounding NVIDIA's inventories post the crypto-mining craze and "Turing" launch, you'll know that the company has been grappling with a pile of unsold 16 nm "GP104" inventories. With the introduction of GDDR6, there could be a similar impact on the slightly older GDDR5X memory chip inventories, and as the sole implementer of GDDR5X, it's natural to imagine the company must have landed itself vast quantities of unsold memory chips. The combination of the two results in a less publicized new SKU, the GeForce GTX 1060 6 GB G5X.
NVIDIA carved the GTX 1060 6 GB G5X by cutting down the "GP104" silicon. Half of its 20 streaming multiprocessors are disabled, resulting in a CUDA core count of 1,280, which is unchanged from the "GP106" based original. The memory bus width, too, is unchanged at 192-bit. What's more, while the GDDR5X memory controllers of the "GP104" are capable of stock memory frequencies as high as 11 Gbps (as seen on the GTX 1080 11 Gbps, another oddball SKU), NVIDIA decided to clock the memory at just 8 Gbps, lower than even last year's GTX 1060 6 GB 9 Gbps. On paper, the GTX 1060 6 GB G5X has the same specifications as the standard GTX 1060 6 GB, with the exception of GDDR5X memory. In theory, it should also have the same exact performance.
So what's the X-factor, you ask? NVIDIA probably gambles that "GDDR5X" looks nice on a box, particularly for the X-mas gift/upgrade buyer, and the company gets to dump its "GP104" and GDDR5X chip inventories in one fell swoop. Buyers who know a thing or two about the underlying technology (eg: you, our esteemed reader), probably imagine that since NVIDIA is using 10 Gbps-rated GDDR5X memory chips, the memory can be overclocked to kingdom come, as well as the GPU, to achieve more performance than an overclocked RX 590 would ever achieve. In this review, we put both theories to test.
We talked to several board vendors, requesting samples of the GTX 1060 6 GB GDDR5X, but none were willing to provide any, probably because this SKU isn't that sexy nowadays. So, we just went out and bought the first GTX 1060 GDDR5X that became available, which is the KFA2 GeForce GTX 1060 6 GB GDDR5X.
The KFA2 card comes with a small overclock out of the box, giving it a base clock of 1518 MHz and a rated boost clock of 1733 MHz, which is just 33 MHz higher than the NVIDIA reference clocks for the GTX 1060 6 GB GDDR5X.
We paid around €250 for our card, which should roughly be equal to US$250 without taxes.
GTX 1060 GDDR5X Market Segment Analysis
Price
Shader Units
ROPs
Core Clock
Boost Clock
Memory Clock
GPU
Transistors
Memory
GTX 970
$235
1664
56
1051 MHz
1178 MHz
1750 MHz
GM204
5200M
4 GB, GDDR5, 256-bit
RX 480
$230
2304
32
1120 MHz
1266 MHz
2000 MHz
Ellesmere
5700M
8 GB, GDDR5, 256-bit
RX 580
$200
2304
32
1257 MHz
1340 MHz
2000 MHz
Ellesmere
5700M
8 GB, GDDR5, 256-bit
RX 590
$280
2304
32
1469 MHz
1545 MHz
2000 MHz
Polaris 30
5700M
8 GB, GDDR5, 256-bit
GTX 1060 3 GB
$200
1152
48
1506 MHz
1708 MHz
2002 MHz
GP106
4400M
3 GB, GDDR5, 192-bit
GTX 1060
$230
1280
48
1506 MHz
1708 MHz
2002 MHz
GP106
4400M
6 GB, GDDR5, 192-bit
GTX 1060 GDDR5X
$250
1280
48
1506 MHz
1708 MHz
1001 MHz
GP104
7200M
6 GB, GDDR5X, 192-bit
KFA2 GTX 1060 GDDR5X OC
$250
1280
48
1518 MHz
1733 MHz
1001 MHz
GP104
7200M
6 GB, GDDR5X, 192-bit
GTX 980 Ti
$390
2816
96
1000 MHz
1075 MHz
1750 MHz
GM200
8000M
6 GB, GDDR5, 384-bit
R9 Fury X
$380
4096
64
1050 MHz
N/A
500 MHz
Fiji
8900M
4 GB, HBM, 4096-bit
GTX 1070
$360
1920
64
1506 MHz
1683 MHz
2002 MHz
GP104
7200M
8 GB, GDDR5, 256-bit
RX Vega 56
$350
3584
64
1156 MHz
1471 MHz
800 MHz
Vega 10
12500M
8 GB, HBM2, 2048-bit
GTX 1070 Ti
$380
2432
64
1607 MHz
1683 MHz
2000 MHz
GP104
7200M
8 GB, GDDR5, 256-bit
Packaging and Contents
You will receive:
Graphics card
Documentation
PCIe power adapter
The Card
The KFA2 card comes with a nice shroud that imitates a high-quality metal look very well even though the shroud is made out of plastic. A backplate is not available. Dimensions of the card are 23.5 x 12.0 cm.
Installation requires two slots in your system.
Display connectivity options include a DisplayPort, an HDMI port, and a DVI port.
Compared to other GTX 1060 cards, the number of outputs is quite low, probably due to cost cutting by KFA2 and not lesser capability of the GP104 GPU. Other GTX 1060 GDDR5X cards do have more outputs.
The board uses one 6-pin power connector. This input configuration is specified for up to 150 watts of power draw.
The GeForce GTX 1060 does not support SLI.
On the back of the PCB, KFA2 has placed some measuring points for voltmodders to get easy access to the voltages for monitoring.
We shine the light from a self-leveling line laser on to the card, which shows no sagging.
Disassembly
The KFA2 heatsink uses two copper heatpipes.
On the next page, we dive deep into the PCB layout and VRM configuration.