AMD Radeon HD 8870M vs NVIDIA GeForce GTX 675MX Comparison

The AMD Radeon HD 8870M and NVIDIA GeForce GTX 675MX are both end-of-life mobile graphics solutions from the 28 nm era, but they deliver different performance profiles. Based on the benchmark data, the NVIDIA GeForce GTX 675MX is the clear winner in raw compute, leading the AMD Radeon HD 8870M by 21.1% in the Geekbench OpenCL test. However, the AMD card is not without merit; it offers a more modern feature set and higher architectural efficiency per transistor, making it a compelling choice for specific workloads. The data shows that while NVIDIA wins on sheer performance, AMD counters with architectural advantages that may matter more to some users.

Where Each One Wins

The NVIDIA GeForce GTX 675MX wins the only direct head-to-head benchmark available. In Geekbench OpenCL, it scores 10723 against the AMD Radeon HD 8870M’s 8462, a decisive 21.1% advantage. This makes the GTX 675MX the superior choice for any compute-heavy task that leverages OpenCL, such as general-purpose GPU computing. The margin is substantial enough that the GTX 675MX is not merely faster; it is in a different performance class for this workload.

The AMD Radeon HD 8870M, despite losing the compute benchmark, wins in architectural sophistication. It is built on GCN 1.0 architecture, which supports DirectX 12 (11_1), whereas the GTX 675MX’s Kepler architecture only supports DirectX 12 (11_0). For applications that can exploit the newer API features, the AMD card holds a functional advantage. Additionally, the HD 8870M achieves a higher transistor density (12.2M / mm²) compared to the GTX 675MX (12.0M / mm²), indicating a more efficient use of silicon area.

The AMD card also wins on connectivity and API support. It uses a PCIe 3.0 x16 interface, while the GTX 675MX relies on an MXM-B (3.0) module interface. For users with a system that supports standard PCIe slots, the HD 8870M is more flexible. Furthermore, the HD 8870M supports Vulkan 1.2.170, while the GTX 675MX supports Vulkan 1.2.175; the difference is minor, but the AMD card’s support is more recent in the version numbering scheme.

Architecture Differences

The two GPUs are built on fundamentally different architectures. The AMD Radeon HD 8870M uses the Venus chip based on GCN 1.0, while the NVIDIA GeForce GTX 675MX uses the GK104 chip based on Kepler. Both are manufactured on a 28 nm process at TSMC, but the similarities end there. The AMD chip packs 1,500 million transistors onto a 123 mm² die, yielding a transistor density of 12.2M / mm². The NVIDIA chip is much larger, with 3,540 million transistors on a 294 mm² die, giving a slightly lower density of 12.0M / mm².

The shading unit count differs significantly. The GTX 675MX has 960 shading units, 80 texture mapping units (TMUs), and 32 render output units (ROPs). The HD 8870M has only 640 shading units, 40 TMUs, and 16 ROPs. This means the NVIDIA card has 50% more shading units, double the TMUs, and double the ROPs. The memory subsystem also diverges: the GTX 675MX uses a 256-bit bus with 115.2 GB/s bandwidth, while the HD 8870M uses a 128-bit bus with 72.00 GB/s bandwidth. Both have 2 GB of GDDR5 memory.

Clock speeds are reported differently. The HD 8870M has a base clock of 725 MHz and a boost clock of 775 MHz, with memory running at 1125 MHz (4.5 Gbps effective). The GTX 675MX does not list base or boost clocks, but its memory runs at 900 MHz (3.6 Gbps effective). The NVIDIA card’s higher pixel rate (13.08 GPixel/s vs 12.40 GPixel/s) and texture rate (52.32 GTexel/s vs 31.00 GTexel/s) directly reflect its larger resource pool.

Head-to-Head Benchmarks

The sole head-to-head benchmark is Geekbench OpenCL, and the results are unambiguous. The NVIDIA GeForce GTX 675MX scores 10723, while the AMD Radeon HD 8870M scores 8462. The deltaPct is -21.1% from the perspective of the AMD card, meaning the GTX 675MX is 21.1% faster. This is the largest performance gap in either card’s nearest rival list. For context, the HD 8870M’s nearest rival is the NVIDIA GeForce MX330, which scores 8458 (a 0.1% difference), and the GTX 675MX’s nearest rival is the AMD Radeon 880M, scoring 8436 (a -0.1% difference from NVIDIA’s perspective). The GTX 675MX also beats the AMD Radeon R9 M375X by 1.2%, but that margin is trivial compared to its lead over the HD 8870M.

The benchmark data also reveals that the GTX 675MX has a second benchmark result: Geekbench Metal, where it scores 6131. The HD 8870M has no Metal benchmark, so no comparison is possible there. However, the OpenCL result is sufficient to establish the GTX 675MX’s dominance in compute tasks. The HD 8870M’s average benchmark score is 8462, while the GTX 675MX’s average is 8427, but this average is skewed because the NVIDIA card’s Metal score is lower than its OpenCL score. When isolating OpenCL, the GTX 675MX is clearly superior.

FAQ

Q: Which GPU is faster in OpenCL compute?

A: The NVIDIA GeForce GTX 675MX is 21.1% faster in Geekbench OpenCL, scoring 10723 versus the AMD Radeon HD 8870M’s 8462.

Q: Do both GPUs support DirectX 12?

A: Yes, both support DirectX 12, but the AMD Radeon HD 8870M supports version 11_1, while the NVIDIA GeForce GTX 675MX supports the older 11_0 feature level.

Q: Which GPU has more memory bandwidth?

A: The NVIDIA GeForce GTX 675MX has 115.2 GB/s bandwidth, compared to the AMD Radeon HD 8870M’s 72.00 GB/s.

Q: Are these GPUs still in production?

A: No. Both are end-of-life products, with the AMD card releasing in 2013 and the NVIDIA card releasing in 2012.

Q: Which GPU has a higher shading unit count?

A: The NVIDIA GeForce GTX 675MX has 960 shading units, while the AMD Radeon HD 8870M has 640.

Q: What is the transistor density difference?

A: The AMD Radeon HD 8870M has a slightly higher density at 12.2M / mm², versus the NVIDIA GeForce GTX 675MX’s 12.0M / mm².

Specification Differences

| Specification | AMD Radeon HD 8870M | NVIDIA GeForce GTX 675MX |

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

| Architecture | GCN 1.0 | Kepler |

| Chip | Venus | GK104 |

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

| Die Size | 123 mm² | 294 mm² |

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

| Base Clock | 725 MHz | Not specified |

| Boost Clock | 775 MHz | Not specified |

| Memory Clock | 1125 MHz (4.5 Gbps effective) | 900 MHz (3.6 Gbps effective) |

| Memory Bus Width | 128 bit | 256 bit |

| Memory Bandwidth | 72.00 GB/s | 115.2 GB/s |

| Shading Units | 640 | 960 |

| TMUs | 40 | 80 |

| ROPs | 16 | 32 |

| Pixel Rate | 12.40 GPixel/s | 13.08 GPixel/s |

| Texture Rate | 31.00 GTexel/s | 52.32 GTexel/s |

| FP32 Performance | 992.0 GFLOPS | 1,255.7 GFLOPS |

| TDP | Not specified | 100 W |

| Slot Width | Not specified | MXM Module |

| Power Connectors | Not specified | None |

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

| Display Outputs | Not specified | Portable Device Dependent |

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

| Vulkan Support | 1.2.170 | 1.2.175 |

| Release Date | 2013-03-31 | 2012-09-30 |

| Predecessor | London | GeForce 500M |

| Successor | Gem System | GeForce 700M |

The Verdict

The data is clear: the NVIDIA GeForce GTX 675MX is the more powerful GPU. Its 21.1% lead in Geekbench OpenCL over the AMD Radeon HD 8870M is decisive, and it also holds advantages in nearly every raw specification, including shading units (960 vs 640), TMUs (80 vs 40), ROPs (32 vs 16), memory bandwidth (115.2 GB/s vs 72.00 GB/s), and FP32 performance (1,255.7 GFLOPS vs 992.0 GFLOPS). For any user prioritizing compute performance, the GTX 675MX is the unequivocal choice.

However, the AMD Radeon HD 8870M is not without its strengths. It features a more modern DirectX 12 feature level (11_1 vs 11_0), a higher transistor density, and a standard PCIe 3.0 x16 interface. It also has a higher boost clock (775 MHz) than its base clock, which the NVIDIA card lacks entirely. For users with older systems that support PCIe but not MXM modules, the HD 8870M is the only viable option. Its lower die size (123 mm² vs 294 mm²) suggests better thermal characteristics, though the TDP is not specified for the AMD card.

The GTX 675MX also holds a slight Vulkan version advantage (1.2.175 vs 1.2.170), but this is negligible in practice. The AMD card’s release date is later (2013 vs 2012), which may imply better driver maturity for newer APIs. Ultimately, the choice depends on the use case: the GTX 675MX wins on performance, while the HD 8870M wins on architectural flexibility and interface compatibility. Benchmark results indicate that for OpenCL compute, the GTX 675MX is the superior product, and its 43rd percentile ranking among all GPUs matches the HD 8870M’s, but with a higher ceiling in the specific test where they compete.