AMD Radeon HD 8870M vs NVIDIA GeForce GTX 1660 Ti Comparison

The AMD Radeon HD 8870M and the NVIDIA GeForce GTX 1660 Ti represent two very different eras of mobile graphics, and the benchmark data reflects that gulf clearly. While the HD 8870M was a capable part in its day, the GTX 1660 Ti is in an entirely different performance class, as the single head-to-head result and supporting scores make plain.

Head-to-Head Benchmarks

The only direct comparison available between these two cards is the Geekbench OpenCL test, and the result is not close. The AMD Radeon HD 8870M scores 8,462 points, while the NVIDIA GeForce GTX 1660 Ti scores 51,029 points. That gives the GTX 1660 Ti a massive 83.4% advantage, meaning it delivers roughly six times the raw compute performance in this workload. This is not a marginal generational improvement; it is a complete overhaul of what the hardware can do.

Looking at the broader benchmark picture reinforces that gap. The HD 8870M’s average benchmark score is 8,462, placing it at the 43rd percentile of all GPUs. Its nearest rivals in the database include the NVIDIA GeForce MX330 at 8,458 (0.1% slower), the AMD Radeon 880M at 8,436 (0.3% slower), and the NVIDIA GeForce GTX 675MX at 8,427 (0.4% slower). The only listed rival that beats it is the Intel Arc A380 at 8,558, which is 1.1% ahead. In short, the HD 8870M sits in a tight cluster of low-end and older parts, all trading blows within a couple of percent of each other.

The GTX 1660 Ti, by contrast, has a much wider benchmark portfolio. Its Geekbench OpenCL score of 51,029 is the standout, but it also shows strong results elsewhere. In PassMark G3D, it scores 12,871; in PassMark G2D, it scores 804; in PassMark GPU Compute, it scores 5,226. It even has a 3DMark Steel Nomad DX12 score of 1,061 and a Geekbench Vulkan score of 5,827. Interestingly, its average benchmark score across all tests is 7,723, which is lower than its OpenCL score alone might suggest. That is because the PassMark DX9, DX10, DX11, and DX12 tests return much smaller numbers (190, 66, 103, and 51 respectively), dragging the average down. This puts the GTX 1660 Ti at the 41st percentile overall, with nearest rivals like the AMD Radeon 540 at 7,673 (0.7% slower), the AMD Radeon Pro WX 3100 at 7,580 (1.9% slower), and the AMD Radeon R7 250 at 7,557 (2.2% slower).

The takeaway is straightforward: the GTX 1660 Ti wins the only shared test by a landslide, and its other scores confirm it is operating in a completely different league. The HD 8870M has no counterargument in the data.

The Verdict

The data points to one clear conclusion: the NVIDIA GeForce GTX 1660 Ti is the superior product in every measurable way. If you are choosing between these two for a system you plan to use for any modern workload, the GTX 1660 Ti is the only sensible pick. The 83.4% lead in OpenCL compute is decisive, and its additional scores in DirectX and Vulkan tests show versatility that the HD 8870M simply cannot match.

The AMD Radeon HD 8870M has no wins in any head-to-head benchmark, and its percentile ranking is only marginally better than the GTX 1660 Ti’s (43rd vs 41st). That slight percentile edge is meaningless in practice, given that the HD 8870M’s nearest rivals are all low-end or legacy parts, while the GTX 1660 Ti’s nearest rivals include the Radeon 540 and Radeon Pro WX 3100, which are also older or entry-level cards. The GTX 1660 Ti’s raw scores are simply in a higher tier.

For a builder today, the GTX 1660 Ti is the obvious choice for anyone who needs actual graphics performance, whether for gaming, content creation, or compute tasks. The HD 8870M is a legacy part, end-of-life, and its performance profile places it firmly in the territory of basic display output or very old games. There is no scenario in the data where the HD 8870M comes out ahead.

Architecture Differences

The two GPUs are built on fundamentally different architectures and processes, which explains much of the performance gap. The AMD Radeon HD 8870M uses the Venus chip, based on GCN 1.0 architecture, manufactured on a 28 nm process at TSMC. It packs 1,500 million transistors on a 123 mm² die, giving it a transistor density of 12.2M per mm². This is an older design, released in early 2013 as part of the Solar System generation (HD 8800M series).

The NVIDIA GeForce GTX 1660 Ti uses the TU116 chip, based on Turing architecture, manufactured on a 12 nm process, also at TSMC. It contains 6,600 million transistors on a 284 mm² die, for a transistor density of 23.2M per mm². That is more than four times the transistor count on a die that is only about 2.3 times larger, reflecting the denser, more modern manufacturing process. The GTX 1660 Ti launched in early 2019 as part of the GeForce 16 series.

The difference in raw specifications is stark. The HD 8870M has 640 shading units, 40 texture mapping units, and 16 ROPs. The GTX 1660 Ti has 1,536 shading units, 96 TMUs, and 48 ROPs. That is 2.4 times the shading units, 2.4 times the TMUs, and 3 times the ROPs. Clock speeds also differ significantly: the HD 8870M runs at a 725 MHz base and 775 MHz boost, while the GTX 1660 Ti runs at 1,500 MHz base and 1,770 MHz boost. The GTX 1660 Ti’s boost clock is more than double the HD 8870M’s base clock.

These architectural differences translate directly into throughput figures. The GTX 1660 Ti achieves 84.96 GPixel/s pixel rate and 169.9 GTexel/s texture rate, versus 12.40 GPixel/s and 31.00 GTexel/s for the HD 8870M. FP32 performance is 5.437 TFLOPS for the GTX 1660 Ti versus 992.0 GFLOPS for the HD 8870M. The GTX 1660 Ti also supports FP16 at 10.87 TFLOPS (2:1 rate), which the HD 8870M does not list at all.

FAQ

Q: Which GPU has higher raw compute performance?

A: The NVIDIA GeForce GTX 1660 Ti is dramatically ahead. In the Geekbench OpenCL test, it scores 51,029 versus 8,462 for the AMD Radeon HD 8870M, an 83.4% difference. Its FP32 throughput is 5.437 TFLOPS compared to 992.0 GFLOPS.

Q: How do their memory subsystems compare?

A: The GTX 1660 Ti has 6 GB of GDDR6 memory on a 192-bit bus with 288.0 GB/s bandwidth and 12 Gbps effective speed. The HD 8870M has 2 GB of GDDR5 memory on a 128-bit bus with 72.00 GB/s bandwidth and 4.5 Gbps effective speed.

Q: Are both GPUs still in production?

A: No. Both are listed as end-of-life products. The AMD Radeon HD 8870M was released in early 2013, and the NVIDIA GeForce GTX 1660 Ti was released in early 2019.

Q: What is the power consumption of each card?

A: The GTX 1660 Ti has a listed TDP of 120 W and a suggested PSU of 300 W. The HD 8870M has no TDP listed in the data.

Q: Which card has better API support?

A: The GTX 1660 Ti supports DirectX 12 (12_1) and Vulkan 1.4, while the HD 8870M supports DirectX 12 (11_1) and Vulkan 1.2.170. Both support OpenGL 4.6.

Q: How does the GTX 1660 Ti compare to its own nearest rivals?

A: In average benchmark score, the GTX 1660 Ti (7,723) leads the AMD Radeon 540 (7,673) by 0.7%, the AMD Radeon Pro WX 3100 (7,580) by 1.9%, the AMD Radeon R7 250 (7,557) by 2.2%, and the Intel Arc A310 (7,550) by 2.3%.

Where Each One Wins

The NVIDIA GeForce GTX 1660 Ti wins in essentially every category the data covers. For compute-heavy tasks like OpenCL workloads, it is in a different stratosphere. For DirectX gaming, its PassMark DX11 score of 103 and DX12 score of 51, while modest in absolute terms, are far above anything the HD 8870M can produce. The GTX 1660 Ti also has Vulkan support at version 1.4 and a Geekbench Vulkan score of 5,827, making it suitable for modern cross-platform titles. Its 6 GB of GDDR6 memory with 288.0 GB/s bandwidth is sufficient for higher-resolution textures and more demanding scenes.

The AMD Radeon HD 8870M wins nowhere in the head-to-head data. Its only benchmark is the Geekbench OpenCL score, where it loses by 83.4%. That said, there is a narrow context where it could be considered: its 43rd percentile ranking is slightly higher than the GTX 1660 Ti’s 41st percentile, and its nearest rivals are all within 1.1% of its score. This suggests that within its own performance class, it is a competitive part. If you have a system that requires the HD 8870M’s specific 28 nm GCN architecture, or you are building a period-correct machine from the early 2010s, it will function as expected. But there is no modern use case where the data supports choosing it over the GTX 1660 Ti.

Specification Differences

The most obvious differences are in core configuration and memory. The GTX 1660 Ti has 1,536 shading units, 96 TMUs, and 48 ROPs, versus 640 shading units, 40 TMUs, and 16 ROPs for the HD 8870M. Memory is 6 GB GDDR6 on a 192-bit bus versus 2 GB GDDR5 on a 128-bit bus, with bandwidth of 288.0 GB/s versus 72.00 GB/s. Clock speeds are also far apart: the GTX 1660 Ti boosts to 1,770 MHz from a 1,500 MHz base, while the HD 8870M boosts to 775 MHz from a 725 MHz base.

The process node and die details differ substantially. The GTX 1660 Ti uses a 12 nm TSMC process with 6,600 million transistors on a 284 mm² die. The HD 8870M uses a 28 nm TSMC process with 1,500 million transistors on a 123 mm² die. Transistor density is 23.2M per mm² versus 12.2M per mm². The GTX 1660 Ti is a dual-slot card with a 229 mm length, 111 mm height, and 35 mm width, requiring a 1x 8-pin power connector and a 300 W suggested PSU. The HD 8870M has no physical dimensions, power connector, or TDP listed.

API support is another differentiator. The GTX 1660 Ti supports DirectX 12 (12_1) and Vulkan 1.4, while the HD 8870M supports DirectX 12 (11_1) and Vulkan 1.2.170. Display outputs also differ: the GTX 1660 Ti offers 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a, while the HD 8870M has no display outputs listed. Both use a PCIe 3.0 x16 bus interface, and both are end-of-life products with no current production status.