NVIDIA GeForce GTX 1660 vs NVIDIA GeForce GTX 870M Comparison

The GeForce GTX 1660 and GTX 870M sit at opposite ends of the NVIDIA stack, separated by five years of architecture. The GTX 1660 is a desktop Turing part from the GeForce 16 series, a product designed for desktop PCIe 3.0 x16 systems. The GTX 870M is a Kepler mobile chip from the GeForce 800M family, built for MXM-B modules in laptops. The data shows one clear winner in compute performance, but the two cards target fundamentally different use cases. This analysis draws exclusively on the recorded benchmark results and specification fields in the database.

Head-to-Head Benchmarks

The only shared benchmark between the two GPUs in the database is Geekbench OpenCL. The GTX 1660 scores 47,850 points, while the GTX 870M scores 12,630 points. That is a 278.9% advantage for the GTX 1660, meaning the newer card delivers nearly four times the OpenCL compute throughput of the older mobile part. This is not a marginal gap; it is a generational leap that dwarfs typical year-over-year improvements.

To put that in context, the GTX 1660’s average benchmark score across all recorded tests is 11,680, placing it in the 51st percentile among all GPUs in the database. Its nearest rivals include the AMD Radeon RX 7800 XT (average score 11,627, a 0.5% difference), the AMD Radeon Pro 5500M (11,528, 1.3% ahead), and the AMD Radeon RX 6500 XT (11,842, 1.4% behind). The GTX 1660 sits within a tight cluster of these cards, effectively trading blows with them. The RX 6500 XT edges it out by 1.4%, while the RX 7800 XT is just 0.5% slower in the database’s average, which suggests the GTX 1660 is a well-balanced midrange performer rather than a standout outlier.

The GTX 870M, by contrast, has an average benchmark score of 9,959, placing it in the 48th percentile. Its nearest rivals are similarly modest: the AMD Radeon Pro 5300M (10,013, 0.5% faster), the NVIDIA Quadro K5100M (10,043, 0.8% faster), and the AMD Radeon R9 M375 (10,070, 1.1% faster). The GTX 870M is essentially on par with these cards, with deltas under 1.2% in either direction. The only other recorded benchmark for the GTX 870M is Geekbench Metal, where it scores 7,288, a test that the GTX 1660 does not appear in.

The head-to-head OpenCL result is the decisive data point. A 278.9% delta means the GTX 1660 is not merely faster; it is in a different performance class. For compute workloads that rely on OpenCL, the GTX 1660 leaves the GTX 870M in the dust. The record shows one win for the GTX 1660 and zero for the GTX 870M, so the head-to-head comparison is unambiguous.

Architecture Differences

The two GPUs come from different architectural eras. The GTX 1660 uses the TU116 chip, built on the Turing architecture, fabricated at TSMC on a 12 nm process. The GTX 870M uses the GK104 chip, based on the older Kepler architecture, also from TSMC but on a 28 nm node. The process shrink from 28 nm to 12 nm is a fundamental shift, enabling nearly double the transistor density. The GTX 1660 packs 6,600 million transistors into a die size of 284 mm², yielding a density of 23.2 million transistors per square millimeter. The GTX 870M has 3,540 million transistors on a slightly larger die of 294 mm², giving a density of just 12.0 million per square millimeter. The GTX 1660 fits almost twice as many transistors into a similar physical footprint, which explains its far higher compute throughput.

Clock speeds reflect this architectural advantage. The GTX 1660 runs at a base clock of 1530 MHz and boosts to 1785 MHz. The GTX 870M operates at a base of 941 MHz and boosts to 967 MHz. The GTX 1660’s boost clock is nearly double the GTX 870M’s boost, and it sustains that frequency while delivering more than double the FP32 performance. In terms of raw floating-point throughput, the GTX 1660 reaches 5.027 TFLOPS (FP32), while the GTX 870M manages 2.599 TFLOPS. The GTX 1660 also supports FP16 compute at 10.05 TFLOPS with a 2:1 ratio, a feature absent from the GTX 870M, which has no recorded FP16 capability.

Memory architecture differs as well. The GTX 1660 has 6 GB of GDDR5 memory on a 192-bit bus, delivering 192.1 GB/s of bandwidth. The GTX 870M has 3 GB of GDDR5 on the same 192-bit bus, but only achieves 120.0 GB/s due to a lower memory clock of 1250 MHz (5 Gbps effective) versus the GTX 1660’s 2001 MHz (8 Gbps effective). The GTX 1660 offers double the capacity and 60% more bandwidth, which matters for texture-heavy workloads and higher resolutions.

Shading resources tell a similar story. The GTX 1660 has 1,408 shading units, 88 texture mapping units (TMUs), and 48 raster operation units (ROPs). The GTX 870M has 1,344 shading units, 112 TMUs, and only 24 ROPs. While the GTX 870M has more TMUs, the GTX 1660 dominates in pixel throughput: 85.68 GPixel/s versus 27.08 GPixel/s, a 3.2x advantage. Texture rate also favors the GTX 1660 at 157.1 GTexel/s versus 108.3 GTexel/s. The TMU count alone does not determine texture performance; clock speed and architecture efficiency matter more.

API support reflects the newer architecture. The GTX 1660 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The GTX 870M supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The feature-level difference in DirectX (12_1 versus 11_0) means the GTX 1660 can handle more advanced rendering features. Neither card has ray tracing or tensor cores, so those fields are null for both.

Power and physical design diverge sharply. The GTX 1660 has a TDP of 120 W, requires a single 8-pin power connector, and a suggested power supply of 300 W. It is a dual-slot card measuring 229 mm in length, 111 mm in height, and 35 mm in width. The GTX 870M has a TDP of 100 W, uses no external power connectors (it draws power from the MXM slot), and has no suggested PSU rating. It is an MXM module, meaning its dimensions are not recorded in the database and its display outputs are described as "Portable Device Dependent," which is typical for laptop components. The GTX 870M is designed for mobility, while the GTX 1660 is a desktop part.

The Verdict

The data is clear for compute and raw graphics throughput: the GTX 1660 is the superior GPU. Its OpenCL score is 278.9% higher than the GTX 870M’s, its FP32 throughput is nearly double, its pixel rate is over three times higher, and its memory bandwidth is 60% greater. The GTX 1660 also has 6 GB of VRAM versus 3 GB, which is critical for modern textures and larger framebuffers. If the choice is between these two for any desktop application, the GTX 1660 wins on every recorded metric.

However, the GTX 870M is a mobile part. Its MXM form factor and portable-device-dependent outputs mean it only fits in laptops. The GTX 1660 cannot be installed in an MXM slot. So the verdict depends on the target platform. For a desktop system, the GTX 1660 is the only viable option, and it is a strong one, sitting in the 51st percentile of all GPUs. For an older laptop that supports MXM-B (3.0) modules, the GTX 870M is a drop-in upgrade path, though its 48th percentile ranking shows it is an average performer by modern standards.

The GTX 870M’s nearest rivals, such as the AMD Radeon R9 M375, are within 1.1% of its average score, so it is not a weak card for its era. But the GTX 1660’s nearest rivals are the RX 6500 XT and RX 7800 XT, which are much newer and faster parts. The performance gap between the two cards is not subtle; it is a generational chasm. Anyone choosing between these two for the same workload, assuming compatibility, should pick the GTX 1660 without hesitation. The only reason to choose the GTX 870M is if the system physically requires an MXM module.

Specification Differences

The key specification differences between the two cards are as follows:

• Process node: GTX 1660 uses 12 nm; GTX 870M uses 28 nm.
• Transistors: GTX 1660 has 6,600 million; GTX 870M has 3,540 million.
• Die size: GTX 1660 is 284 mm²; GTX 870M is 294 mm².
• Transistor density: GTX 1660 has 23.2M per mm²; GTX 870M has 12.0M per mm².
• Base clock: GTX 1660 runs at 1530 MHz; GTX 870M at 941 MHz.
• Boost clock: GTX 1660 boosts to 1785 MHz; GTX 870M to 967 MHz.
• Memory clock: GTX 1660 uses 2001 MHz (8 Gbps effective); GTX 870M uses 1250 MHz (5 Gbps effective).
• Memory size: GTX 1660 has 6 GB; GTX 870M has 3 GB.
• Memory bandwidth: GTX 1660 delivers 192.1 GB/s; GTX 870M delivers 120.0 GB/s.
• Shading units: GTX 1660 has 1,408; GTX 870M has 1,344.
• TMUs: GTX 1660 has 88; GTX 870M has 112.
• ROPs: GTX 1660 has 48; GTX 870M has 24.
• Pixel rate: GTX 1660 is 85.68 GPixel/s; GTX 870M is 27.08 GPixel/s.
• Texture rate: GTX 1660 is 157.1 GTexel/s; GTX 870M is 108.3 GTexel/s.
• FP32 performance: GTX 1660 is 5.027 TFLOPS; GTX 870M is 2.599 TFLOPS.
• FP16 performance: GTX 1660 is 10.05 TFLOPS (2:1); GTX 870M has none.
• TDP: GTX 1660 is 120 W; GTX 870M is 100 W.
• Slot width: GTX 1660 is dual-slot; GTX 870M is an MXM module.
• Power connectors: GTX 1660 uses 1x 8-pin; GTX 870M uses none.
• Suggested PSU: GTX 1660 recommends 300 W; GTX 870M has no suggestion.
• Bus interface: GTX 1660 uses PCIe 3.0 x16; GTX 870M uses MXM-B (3.0).
• Display outputs: GTX 1660 has DVI, HDMI 2.0, DisplayPort 1.4a; GTX 870M is portable-device dependent.
• DirectX support: GTX 1660 supports 12 (12_1); GTX 870M supports 12 (11_0).
• Vulkan support: GTX 1660 supports 1.4; GTX 870M supports 1.2.175.
• Release date: GTX 1660 launched in March 2019; GTX 870M launched in March 2014.
• Predecessor/successor: GTX 1660 follows GeForce 10 and precedes GeForce 20; GTX 870M follows GeForce 700M and precedes GeForce 900M.

FAQ

Q: Which GPU is faster in OpenCL compute?

A: The GTX 1660 scores 47,850 in Geekbench OpenCL, while the GTX 870M scores 12,630. That is a 278.9% advantage for the GTX 1660.

Q: How much VRAM does each card have?

A: The GTX 1660 has 6 GB of GDDR5 memory, while the GTX 870M has 3 GB of GDDR5 memory. Both use a 192-bit memory bus.

Q: What are the architecture generations for these GPUs?

A: The GTX 1660 uses the Turing architecture on a 12 nm process, while the GTX 870M uses the Kepler architecture on a 28 nm process.

Q: Which card has higher pixel fill rate?

A: The GTX 1660 has a pixel rate of 85.68 GPixel/s, compared to the GTX 870M’s 27.08 GPixel/s. The GTX 1660 is more than three times faster in this metric.

Q: Is the GTX 870M a desktop card?

A: No, the GTX 870M is an MXM module with portable-device-dependent display outputs, designed for laptops. The GTX 1660 is a dual-slot desktop card with DVI, HDMI, and DisplayPort outputs.

Q: What is the performance percentile ranking for each?

A: The GTX 1660 is in the 51st percentile of all GPUs in the database, while the GTX 870M is in the 48th percentile.