NVIDIA GeForce GTX 1060 6 GB vs NVIDIA GeForce RTX 2070 SUPER Comparison

The NVIDIA GeForce GTX 1060 6 GB and the NVIDIA GeForce RTX 2070 SUPER represent two distinct generations of NVIDIA GPU design, separated by a three-year gap in release timing. The data places the GTX 1060 in the 67th percentile of all GPUs, while the RTX 2070 SUPER sits at the 65th percentile, despite the latter delivering dramatically higher raw performance in the shared benchmarks. This apparent contradiction is explained by the benchmark sets used: the GTX 1060's average score of 21,856 is derived from three Geekbench tests, whereas the RTX 2070 SUPER's average of 20,282 includes a broader mix of DirectX and Passmark workloads, some of which are heavily weighted toward older APIs. The head-to-head comparison, however, is unambiguous: the RTX 2070 SUPER wins both shared tests decisively, with the GTX 1060 failing to secure a single victory.

Head-to-Head Benchmarks

The two cards share exactly two benchmark results in the data: Geekbench OpenCL and Geekbench Vulkan. In the OpenCL test, the GTX 1060 scores 32,322, while the RTX 2070 SUPER reaches 83,358. This represents a delta of -61.2% for the older card, meaning the RTX 2070 SUPER delivers roughly 2.6 times the OpenCL compute performance. The Vulkan gap is even more pronounced: the GTX 1060 manages 18,659, while the RTX 2070 SUPER posts 90,637, a delta of -79.4%. In absolute terms, the RTX 2070 SUPER is about 4.9 times faster in this workload. The Vulkan result is particularly telling because it reflects the newer card's ability to leverage its architectural advantages through a low-level API, whereas the OpenCL gap, while large, is less extreme.

Interpreting these deltas requires context. The -61.2% figure in OpenCL means the GTX 1060 scores 61.2% lower than the RTX 2070 SUPER, not that it is 61.2% as fast. The data shows the RTX 2070 SUPER's OpenCL score is 2.58x higher, and its Vulkan score is 4.86x higher. No other benchmarks are shared between the two cards, so these two tests constitute the entirety of the direct comparison. The RTX 2070 SUPER's nearest rivals in the overall database—the NVIDIA Quadro M4000M, NVIDIA GeForce RTX 3070 Mobile, Intel Arc B570, and Intel Arc A750—all sit within a 1.5% range of its average score, indicating that its average is pulled down by legacy DirectX benchmarks. The GTX 1060's rivals, including the NVIDIA Quadro RTX 5000 and AMD Radeon RX 6700S, cluster within 2.2% of its average, showing that its three Geekbench results are internally consistent.

Architecture Differences

The GTX 1060 is built on the Pascal architecture using the GP106 chip, fabricated on a 16 nm process at TSMC. It contains 4,400 million transistors on a 200 mm² die, yielding a transistor density of 22.0M per mm². The RTX 2070 SUPER uses the Turing architecture with the TU104 chip, also from TSMC but on a 12 nm process. This chip packs 13,600 million transistors into a 545 mm² die, for a density of 25.0M per mm². The transistor count is over three times higher on the newer card, and the die is 2.7 times larger. The process node shrink from 16 nm to 12 nm allows for this increase in complexity while keeping power consumption manageable.

The shading resources differ substantially. The GTX 1060 has 1,280 shading units, 80 texture mapping units, and 48 ROPs. The RTX 2070 SUPER doubles the shading units to 2,560, doubles the TMUs to 160, and increases ROPs to 64. The RTX 2070 SUPER also introduces dedicated hardware absent from the GTX 1060: 40 RT cores for ray tracing and 320 tensor cores for AI acceleration. Neither the RT cores nor tensor cores have direct analogues in the Pascal chip, as the GTX 1060 lists null values for both. This is a generational leap in feature set, not just a scaling of existing resources.

Clock speeds are higher on the RTX 2070 SUPER, with a base of 1605 MHz and boost of 1770 MHz, compared to the GTX 1060's 1506 MHz base and 1709 MHz boost. The memory subsystem also diverges: the GTX 1060 uses 6 GB of GDDR5 on a 192-bit bus, while the RTX 2070 SUPER uses 8 GB of GDDR6 on a 256-bit bus. Memory bandwidth jumps from 192.2 GB/s to 448.0 GB/s, a 2.33x increase. The effective memory speed is 8 Gbps on the older card versus 14 Gbps on the newer one. Pixel rate rises from 82.03 GPixel/s to 113.3 GPixel/s, and texture rate from 136.7 GTexel/s to 283.2 GTexel/s. FP32 compute nearly doubles from 4.375 TFLOPS to 9.062 TFLOPS, while FP16 performance is dramatically different: the GTX 1060 offers 68.36 GFLOPS at a 1:64 ratio, whereas the RTX 2070 SUPER offers 18.12 TFLOPS at a 2:1 ratio.

FAQ

Q: Which card has higher memory bandwidth?

A: The RTX 2070 SUPER has 448.0 GB/s of bandwidth, which is 2.33 times the GTX 1060's 192.2 GB/s. The newer card also uses GDDR6 memory instead of GDDR5.

Q: Does the GTX 1060 support ray tracing?

A: No. The GTX 1060 has no RT cores listed, while the RTX 2070 SUPER includes 40 RT cores. The GTX 1060 is a Pascal-generation card without dedicated ray tracing hardware.

Q: What is the FP32 compute difference?

A: The RTX 2070 SUPER delivers 9.062 TFLOPS of FP32 compute, while the GTX 1060 delivers 4.375 TFLOPS. This is approximately a 2.07x advantage for the newer card.

Q: How do the cards compare in Vulkan performance?

A: In the Geekbench Vulkan test, the RTX 2070 SUPER scores 90,637 versus the GTX 1060's 18,659, a delta of -79.4% for the older card. The RTX 2070 SUPER is roughly 4.86 times faster.

Q: Which card has a higher transistor density?

A: The RTX 2070 SUPER has a density of 25.0M transistors per mm², compared to the GTX 1060's 22.0M per mm². This is despite the newer card's larger die and more advanced 12 nm process.

Q: What is the DirectX feature level of each card?

A: The GTX 1060 supports DirectX 12 (12_1), while the RTX 2070 SUPER supports DirectX 12 Ultimate (12_2). The newer card also has a higher Vulkan API version provision, though both list Vulkan 1.4.

Specification Differences

The two cards differ across nearly every major specification field. The process node is 16 nm for the GTX 1060 and 12 nm for the RTX 2070 SUPER, both from TSMC. Transistor count is 4,400 million versus 13,600 million, and die size is 200 mm² versus 545 mm². Base clocks are 1506 MHz versus 1605 MHz, and boost clocks are 1709 MHz versus 1770 MHz. Memory size is 6 GB versus 8 GB, with GDDR5 versus GDDR6 type. Bus width is 192-bit versus 256-bit, and bandwidth is 192.2 GB/s versus 448.0 GB/s. Shading units are 1,280 versus 2,560, TMUs are 80 versus 160, and ROPs are 48 versus 64. The RTX 2070 SUPER has 40 RT cores and 320 tensor cores, while the GTX 1060 lists none. Pixel rate is 82.03 versus 113.3 GPixel/s, texture rate is 136.7 versus 283.2 GTexel/s, and FP32 is 4.375 versus 9.062 TFLOPS. FP16 performance differs by orders of magnitude: 68.36 GFLOPS (1:64) versus 18.12 TFLOPS (2:1). TDP is 120 W versus 215 W, with the GTX 1060 requiring a single 6-pin connector and a 300 W power supply, while the RTX 2070 SUPER needs a 6-pin plus 8-pin and a 550 W PSU. Display outputs are 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.4a on the GTX 1060, versus 1x HDMI 2.0, 3x DisplayPort 1.4a, and 1x USB Type-C on the RTX 2070 SUPER. The GTX 1060 is 250 mm long and 111 mm tall, while the RTX 2070 SUPER is 267 mm long, 116 mm tall, and 35 mm wide. The GTX 1060 was released on 2016-07-18, and the RTX 2070 SUPER on 2019-07-08.

The Verdict

The benchmark data shows a clear generational hierarchy. In the only two directly comparable tests, the RTX 2070 SUPER wins both with deltas of -61.2% and -79.4% against the GTX 1060. The newer card offers more than double the shading units, TMUs, and FP32 compute, along with a 2.33x memory bandwidth advantage. It also introduces RT and tensor cores, which the GTX 1060 lacks entirely. The RTX 2070 SUPER's average benchmark score of 20,282 is slightly lower than the GTX 1060's 21,856, but this is an artifact of the benchmark mix: the RTX 2070 SUPER's scores include low-value Passmark DirectX 9 (223), DirectX 10 (132), DirectX 11 (151), and DirectX 12 (67) results, which drag down its average. The GTX 1060's average is based solely on Geekbench Metal, OpenCL, and Vulkan, all of which are compute-heavy and favor its architecture. For any modern workload using Vulkan or OpenCL, the RTX 2070 SUPER is the superior choice by a wide margin. The GTX 1060 retains relevance only in legacy DirectX scenarios, where its higher percentile ranking (67 versus 65) reflects its relative strength in those specific tests.

Where Each One Wins

The RTX 2070 SUPER wins decisively in every shared benchmark and in all modern compute scenarios. Its Vulkan score of 90,637 versus 18,659 indicates a 4.86x advantage in low-level API workloads, which are increasingly common in current games and applications. Its OpenCL score of 83,358 versus 32,322 shows a 2.58x edge in general-purpose GPU compute. The RTX 2070 SUPER also supports DirectX 12 Ultimate (12_2), enabling features like ray tracing and variable rate shading that the GTX 1060's DirectX 12 (12_1) cannot handle. For users running 3DMark Steel Nomad or other modern DirectX 12 Ultimate titles, the RTX 2070 SUPER is the only viable option of the two. Its higher memory capacity (8 GB versus 6 GB) and bandwidth (448.0 GB/s versus 192.2 GB/s) provide headroom for texture-heavy workloads.

The GTX 1060's wins are confined to its average benchmark percentile. Its 67th percentile ranking versus the RTX 2070 SUPER's 65th reflects the composition of its benchmark set, which is entirely Geekbench-based and does not include the legacy DirectX tests that penalize the RTX 2070 SUPER. In the specific Passmark DirectX 9, 10, 11, and 12 tests, the RTX 2070 SUPER scores 223, 132, 151, and 67 respectively, which are low absolute numbers that pull its average down. The GTX 1060 has no corresponding Passmark results in the data, so a direct comparison in those workloads is impossible. However, the GTX 1060's lower TDP of 120 W versus 215 W means it draws significantly less power, and its single 6-pin connector and 300 W PSU requirement make it easier to integrate into systems with modest power supplies. Its smaller physical footprint (250 mm versus 267 mm length) also aids compatibility in compact cases. For users constrained by power or space, or those running primarily legacy DirectX 9/10/11 applications, the GTX 1060 remains a functional option—but the data shows no performance scenario where it beats the RTX 2070 SUPER in a shared, modern benchmark.