AMD Radeon HD 8870M vs NVIDIA GeForce GTX 870M Comparison

Where Each One Wins

The benchmark data splits cleanly between these two mobile graphics processors. The NVIDIA GeForce GTX 870M claims the only recorded head-to-head victory, taking the Geekbench OpenCL test with a score of 12,630 against 8,462 for the AMD Radeon HD 8870M. That is a decisive 49.3% margin in favor of the NVIDIA part. The AMD card has no benchmark wins in the database, meaning every recorded measurement favors the GeForce GTX 870M.

Looking at the broader database standings, the GTX 870M sits at the 48th percentile among all GPUs, while the Radeon HD 8870M lands at the 43rd percentile. The average benchmark score tells the same story: 9,959 for the NVIDIA part versus 8,462 for the AMD part. The delta of roughly 1,500 points represents a substantial performance gap, not a marginal one. When the database ranks these two against their nearest rivals, the GTX 870M trades blows with cards like the AMD Radeon Pro 5300M (0.5% behind), the NVIDIA Quadro K5100M (0.8% behind), and the AMD Radeon R9 M375 (1.1% behind), while running 1.1% ahead of the NVIDIA Quadro 6000. The Radeon HD 8870M, by contrast, sits within a narrow band of its own rivals: 0.1% ahead of the GeForce MX330, 0.3% ahead of the Radeon 880M, 0.4% ahead of the GeForce GTX 675MX, and 1.1% behind the Intel Arc A380.

The data suggests the GTX 870M is the clear compute winner in OpenCL workloads. The Radeon HD 8870M does not hold a single recorded advantage. For any application that relies on OpenCL acceleration, the NVIDIA part is the safer choice based on these measurements.

Architecture Differences

The two chips come from different architectural families and different generations. NVIDIA's GeForce GTX 870M uses the GK104 chip built on the Kepler architecture, part of the GeForce 800M generation. AMD's Radeon HD 8870M uses the Venus chip built on GCN 1.0, part of the Solar System (HD 8800M) generation. Both are fabricated by TSMC on a 28 nm process, and their transistor densities are nearly identical: 12.0 million transistors per square millimeter for the NVIDIA chip versus 12.2 million for the AMD chip. The similarity ends there.

The GK104 die is substantially larger and more complex. It packs 3,540 million transistors on a 294 mm² die, while the Venus chip contains 1,500 million transistors on a 123 mm² die. That is more than double the transistor count and more than double the physical area. The NVIDIA part fields 1,344 shading units, 112 texture mapping units, and 24 render output units. The AMD part counters with 640 shading units, 40 TMUs, and 16 ROPs. In every functional unit category, the GTX 870M has the higher count.

Clock speeds favor the NVIDIA chip as well. The GTX 870M runs at a 941 MHz base clock with a 967 MHz boost, while the Radeon HD 8870M operates at 725 MHz base and 775 MHz boost. Memory clocks follow the same pattern: the GTX 870M uses 1250 MHz memory (5 Gbps effective) versus 1125 MHz (4.5 Gbps effective) for the AMD part. The NVIDIA GPU also carries 3 GB of GDDR5 on a 192-bit bus, yielding 120.0 GB/s of bandwidth. The AMD GPU has 2 GB of GDDR5 on a 128-bit bus, yielding 72.00 GB/s. That is a 66.7% bandwidth advantage for the GTX 870M, which matters in memory-heavy workloads.

Compute throughput figures reflect the architectural gap. The GTX 870M achieves 2.599 TFLOPS of FP32 performance, a pixel rate of 27.08 GPixel/s, and a texture rate of 108.3 GTexel/s. The Radeon HD 8870M delivers 992.0 GFLOPS of FP32, 12.40 GPixel/s, and 31.00 GTexel/s. The NVIDIA part is roughly 2.6 times faster in FP32, more than twice as fast in pixel throughput, and about 3.5 times faster in texture fill.

Power delivery and interface also differ. The GTX 870M is an MXM module with no power connectors and a 100 W TDP, using an MXM-B (3.0) bus interface. The Radeon HD 8870M uses a PCIe 3.0 x16 interface, with no TDP, slot width, or power connector data recorded. Both support DirectX 12, but the NVIDIA part is rated for DX 12 (11_0) while the AMD part supports DX 12 (11_1). OpenGL support is identical at 4.6. Vulkan support is close: 1.2.175 for NVIDIA, 1.2.170 for AMD.

Head-to-Head Benchmarks

The database records exactly one head-to-head benchmark between these two GPUs, and it is the Geekbench OpenCL test. The NVIDIA GeForce GTX 870M scores 12,630 against 8,462 for the AMD Radeon HD 8870M, a 49.3% advantage. That margin is significant enough to place the two cards in different performance tiers, despite their similar mobile positioning.

Context from the rivals list reinforces this gap. The GTX 870M's average benchmark score of 9,959 places it in a cluster with the AMD Radeon Pro 5300M (10,013, only 0.5% higher), the NVIDIA Quadro K5100M (10,043, 0.8% higher), and the AMD Radeon R9 M375 (10,070, 1.1% higher). It also edges past the NVIDIA Quadro 6000 (9,846) by 1.1%. The Radeon HD 8870M's average of 8,462 sits among lighter competition: the GeForce MX330 (8,458, 0.1% lower), the Radeon 880M (8,436, 0.3% lower), and the GeForce GTX 675MX (8,427, 0.4% lower). The Intel Arc A380 (8,558) is 1.1% ahead of the AMD card.

These cluster positions matter. The GTX 870M competes with cards that are roughly 10,000-point performers, while the HD 8870M competes with cards in the 8,400 to 8,600 range. The 1,497-point gap in average scores between the two units is larger than the entire spread within either card's rival group. In the recorded OpenCL measurement, the NVIDIA part is not just ahead; it is ahead by nearly half of the AMD card's own score.

The Verdict

The data supports one clear conclusion: the NVIDIA GeForce GTX 870M is the stronger GPU in every recorded metric. It wins the only head-to-head benchmark by 49.3%, holds a higher average benchmark score (9,959 versus 8,462), and ranks higher in the overall database at the 48th percentile versus the 43rd percentile. Its architectural resources, including 1,344 shading units, 112 TMUs, 24 ROPs, and 120.0 GB/s of memory bandwidth, all exceed the AMD part's corresponding specifications.

Buyers or system builders choosing between these two should pick the GTX 870M if OpenCL compute performance is a priority. The 12,630 OpenCL score is in a different league from the 8,462 posted by the Radeon HD 8870M. The GTX 870M also offers more memory (3 GB versus 2 GB) and a wider memory bus (192-bit versus 128-bit), which helps in texture-heavy and high-resolution workloads.

The Radeon HD 8870M is not without merit. It uses a smaller die (123 mm² versus 294 mm²) and fewer transistors (1,500 million versus 3,540 million), which historically implies lower manufacturing cost and potentially lower power draw, though no TDP is recorded for the AMD part. Its DirectX 12 (11_1) support is slightly newer than the NVIDIA card's DirectX 12 (11_0). But these are secondary considerations. In raw performance, the database shows no category where the AMD card leads.

For a laptop buyer facing this choice, the GTX 870M is the performance pick. The HD 8870M would only make sense if the system integration, power envelope, or other non-benchmark factors outweighed the substantial compute deficit, but the recorded data cannot justify that trade on performance grounds alone.

FAQ

Q: Which GPU wins the recorded head-to-head benchmark?

A: The NVIDIA GeForce GTX 870M wins the Geekbench OpenCL test with a score of 12,630 versus 8,462 for the AMD Radeon HD 8870M, a 49.3% margin.

Q: How do their average benchmark scores compare?

A: The GTX 870M has an average benchmark score of 9,959, while the Radeon HD 8870M averages 8,462. The NVIDIA part ranks at the 48th percentile among all GPUs, and the AMD part ranks at the 43rd percentile.

Q: What are the memory specifications for each card?

A: The GTX 870M has 3 GB of GDDR5 on a 192-bit bus with 120.0 GB/s bandwidth. The Radeon HD 8870M has 2 GB of GDDR5 on a 128-bit bus with 72.00 GB/s bandwidth.

Q: Which card has more shading units?

A: The GTX 870M has 1,344 shading units, 112 TMUs, and 24 ROPs. The Radeon HD 8870M has 640 shading units, 40 TMUs, and 16 ROPs.

Q: What are the clock speeds of the two GPUs?

A: The GTX 870M runs at a 941 MHz base clock and 967 MHz boost. The Radeon HD 8870M runs at a 725 MHz base clock and 775 MHz boost.

Q: How does the GTX 870M compare to its nearest rivals?

A: The GTX 870M is 0.5% behind the AMD Radeon Pro 5300M, 0.8% behind the NVIDIA Quadro K5100M, 1.1% behind the AMD Radeon R9 M375, and 1.1% ahead of the NVIDIA Quadro 6000.

Specification Differences

| Specification | NVIDIA GeForce GTX 870M | AMD Radeon HD 8870M |

|---|---|---|

| Chip | GK104 | Venus |

| Architecture | Kepler | GCN 1.0 |

| Generation | GeForce 800M | Solar System (HD 8800M) |

| Process Node | 28 nm | 28 nm |

| Transistors | 3,540 million | 1,500 million |

| Die Size | 294 mm² | 123 mm² |

| Transistor Density | 12.0M / mm² | 12.2M / mm² |

| Base Clock | 941 MHz | 725 MHz |

| Boost Clock | 967 MHz | 775 MHz |

| Memory Clock | 1250 MHz, 5 Gbps effective | 1125 MHz, 4.5 Gbps effective |

| Memory Size | 3 GB | 2 GB |

| Memory Type | GDDR5 | GDDR5 |

| Memory Bus Width | 192 bit | 128 bit |

| Memory Bandwidth | 120.0 GB/s | 72.00 GB/s |

| Shading Units | 1344 | 640 |

| TMUs | 112 | 40 |

| ROPs | 24 | 16 |

| Pixel Rate | 27.08 GPixel/s | 12.40 GPixel/s |

| Texture Rate | 108.3 GTexel/s | 31.00 GTexel/s |

| FP32 Performance | 2.599 TFLOPS | 992.0 GFLOPS |

| TDP | 100 W | Not recorded |

| Bus Interface | MXM-B (3.0) | PCIe 3.0 x16 |

| DirectX Support | 12 (11_0) | 12 (11_1) |

| Vulkan Support | 1.2.175 | 1.2.170 |

| Release Date | 2014-03-11 | 2013-03-31 |

| Predecessor | GeForce 700M | London |

| Successor | GeForce 900M | Gem System |