NVIDIA GeForce GTX 1660 SUPER vs NVIDIA GeForce RTX 2060 Comparison
NVIDIA GeForce GTX 1660 SUPER
GeForce RTX 2060
PERFORMANCE BENCHMARKS
Analysis: NVIDIA GeForce GTX 1660 SUPER vs NVIDIA GeForce RTX 2060
The NVIDIA GeForce RTX 2060 and the NVIDIA GeForce GTX 1660 SUPER are two Turing-based cards separated less by architecture than by feature set and silicon scale. The database shows the RTX 2060 winning eight of ten recorded head-to-head benchmarks, yet the GTX 1660 SUPER takes two results of its own, including a modern DirectX 12 test. Both share an identical memory subsystem, which makes the gaps in compute and graphics scores all the more interesting to unpack. What follows is a walk through the numbers, the silicon differences that explain them, and where each card earns its place.
Head-to-Head Benchmarks
The RTX 2060's single largest margin appears in passmark_directx_10, where it scores 98 against the GTX 1660 SUPER's 65, a 50.8 percent lead. Legacy API performance clearly favors the larger chip, and the pattern repeats elsewhere: 110 versus 104 in passmark_directx_11 (a 5.8 percent edge), 53 versus 50 in passmark_directx_12 (6 percent), and a near-dead heat in passmark_directx_9 at 190 versus 189, where the RTX 2060 squeaks ahead by 0.5 percent.
Compute-heavy workloads tilt even harder toward the RTX card. Geekbench OpenCL lands at 65014 for the RTX 2060 versus 52490 for the GTX 1660 SUPER, a 23.9 percent gap. Geekbench Vulkan shows 65846 versus 57102, a 15.3 percent advantage for the RTX 2060. PassMark GPU compute follows the same direction with 5488 versus 5076, an 8.1 percent difference. These are not marginal wins; the OpenCL result in particular suggests the 1660 SUPER leaves a substantial amount of throughput on the table in general-purpose workloads.
The GTX 1660 SUPER's two victories deserve scrutiny. In 3DMark Steel Nomad under DirectX 12, it scores 1278 against the RTX 2060's 1242, a 2.8 percent lead. It also wins passmark_g2d with 805 versus 745, a 7.5 percent margin. A modern ray-free DirectX 12 workload favoring the cheaper-positioned card raises a question the data cannot fully answer: does the 1660 SUPER's higher clocking offset its smaller shader count in certain engines? The aggregate numbers still favor the RTX 2060, with an average benchmark score of 15290 versus 12986, roughly an 18 percent advantage in the database average, and a percentile placement of 58 versus 53 against all GPUs.
Context from the rivals list sharpens the picture. The RTX 2060's average score sits essentially even with the NVIDIA GeForce GTX 580's 15283 (a 0 percent delta) and the AMD Radeon 680M's 15270 (0.1 percent), while edging the RTX 3050 OEM by 0.6 percent and the Radeon RX 7600 by 0.8 percent. The GTX 1660 SUPER clusters near the RTX 3050 Ti Mobile (0.4 percent delta), the Radeon RX 580 (0.4 percent), the Tesla M2090 (-0.7 percent), and the Radeon 740M (0.9 percent). In other words, the RTX 2060 keeps company with a slightly higher performance tier.
FAQ
Q: Which card wins more benchmarks overall?
A: The RTX 2060 wins 8 of the 10 head-to-head tests in the database, against 2 wins for the GTX 1660 SUPER.
Q: Does either card win a DirectX 12 test?
A: Yes. The GTX 1660 SUPER takes 3DMark Steel Nomad DX12 with 1278 versus 1242, a 2.8 percent lead, while the RTX 2060 wins PassMark DirectX 12 with 53 versus 50.
Q: How big is the compute gap?
A: In Geekbench OpenCL the RTX 2060 scores 65014 versus 52490, a 23.9 percent advantage. In Geekbench Vulkan it leads 65846 to 57102, 15.3 percent ahead.
Q: Do the cards differ in memory?
A: No. Both carry 6 GB of GDDR6 on a 192-bit bus with 336.0 GB/s of bandwidth and memory clocked at 1750 MHz (14 Gbps effective).
Q: Which card needs more power?
A: The RTX 2060 has a 160 W TDP with a 450 W suggested PSU; the GTX 1660 SUPER sits at 125 W with a 300 W suggested PSU. Both use one 8-pin connector.
Q: Where does each rank against all GPUs?
A: The RTX 2060 sits in the 58th percentile with an average benchmark score of 15290; the GTX 1660 SUPER sits in the 53rd percentile at 12986.
Architecture Differences
Both GPUs are Turing designs built on TSMC's 12 nm process, but they are very different slices of that architecture. The RTX 2060 uses the TU106 chip with 10,800 million transistors on a 445 mm² die, yielding a density of 24.3M per mm². The GTX 1660 SUPER uses the smaller TU116 with 6,600 million transistors on a 284 mm² die at 23.2M per mm². That is roughly 62 percent more transistors and a die more than half again as large for the RTX card.
The feature gap is the real story. The RTX 2060 includes 30 RT cores and 240 tensor cores; the GTX 1660 SUPER has neither. This maps directly onto API support: the RTX 2060 reports DirectX 12 Ultimate (feature level 12_2), while the GTX 1660 SUPER reports DirectX 12 (12_1). The compute benchmark wins for the RTX 2060 are consistent with its additional specialized hardware and much higher raw throughput: 6.451 TFLOPS FP32 versus 5.027 TFLOPS, and 12.90 TFLOPS FP16 versus 10.05 TFLOPS. Texture rate also favors the RTX card, 201.6 GTexel/s versus 157.1 GTexel/s.
Curiously, pixel fill rate goes the other way, 85.68 GPixel/s for the GTX 1660 SUPER against 80.64 GPixel/s for the RTX 2060, a consequence of the 1660 SUPER's higher boost clock. That may partly explain its passmark_g2d win and possibly its Steel Nomad result.
Specification Differences
The specification table splits cleanly. Where the two cards match: 6 GB GDDR6, 192-bit bus, 336.0 GB/s bandwidth, 48 ROPs, PCIe 3.0 x16, dual-slot width, one 8-pin connector, identical 229 mm length and 35 mm width, OpenGL 4.6, Vulkan 1.4, and end-of-life production status.
Where they differ:
- Chip and die: TU106, 445 mm², 10,800 million transistors versus TU116, 284 mm², 6,600 million transistors.
- Shading hardware: 1920 shading units and 120 TMUs versus 1408 shading units and 88 TMUs.
- Accelerators: 30 RT cores and 240 tensor cores versus none of either.
- Clocks: 1365 MHz base and 1680 MHz boost for the RTX 2060; 1530 MHz base and 1785 MHz boost for the GTX 1660 SUPER.
- Power: 160 W TDP and a 450 W suggested PSU versus 125 W TDP and a 300 W suggested PSU.
- Outputs: the RTX 2060 adds a USB Type-C output alongside DVI, HDMI 2.0, and two DisplayPort 1.4a; the GTX 1660 SUPER omits USB Type-C.
- DirectX support: 12 Ultimate (12_2) versus 12 (12_1).
- Height: 113 mm versus 111 mm.
- Series placement: the RTX 2060 belongs to the GeForce 20-series generation; the GTX 1660 SUPER is a GeForce 16 generation part whose successor is listed as GeForce 20.
- Launch MSRP: 349 USD for the RTX 2060, 229 USD for the GTX 1660 SUPER.
- Release dates: January 2019 for the RTX 2060, October 2019 for the GTX 1660 SUPER.
The Verdict
The recorded data points one direction for anyone chasing maximum performance: the RTX 2060 wins the database average by roughly 18 percent, sits five percentile points higher against all GPUs, and dominates OpenCL by 23.9 percent and Vulkan by 15.3 percent. It also brings RT cores, tensor cores, and DirectX 12 Ultimate support, features the GTX 1660 SUPER simply lacks, which matters for any workload or title that leans on those capabilities.
The GTX 1660 SUPER's case rests on efficiency and a narrower than expected gap in gaming-centric tests. It consumes 125 W against 160 W, pairs with a 300 W suggested PSU versus 450 W, and still manages to win 3DMark Steel Nomad DX12 by 2.8 percent and 2D graphics by 7.5 percent. If a workload matches its strengths and avoids heavy compute, the performance penalty shrinks considerably while the power draw drops by 35 W.
Notably, the memory subsystems are identical, so neither card holds a framebuffer or bandwidth advantage. Every performance difference here traces back to silicon scale, clocks, and dedicated accelerators.
Where Each One Wins
RTX 2060 territory: Compute and GPU processing workloads are its stronghold, evidenced by the 23.9 percent OpenCL lead and 8.1 percent PassMark compute lead. Anything requiring DirectX 12 Ultimate feature support, or the RT and tensor hardware, is exclusive to this card. Legacy DirectX 9 through 11 gaming also favors it, most dramatically in the 50.8 percent DirectX 10 margin. Users matching its performance tier, near the Radeon RX 7600 and RTX 3050 OEM, will find it the stronger all-rounder.
GTX 1660 SUPER territory: Power-sensitive builds are its natural home, with the lower TDP and much lower PSU recommendation. Modern DirectX 12 rasterization shows it competitive, even ahead in Steel Nomad, and its superior pixel fill rate plus 2D score suit general display work. Its peer group, the Radeon RX 580 and RTX 3050 Ti Mobile, reflects a card that holds its own a tier below the RTX 2060 while doing so with meaningfully less power.
The final tally, 8 wins to 2, favors the RTX 2060 decisively, but the two GTX 1660 SUPER wins land exactly where its design philosophy concentrates: efficiency and straightforward rasterization.