AMD Radeon Vega 8 vs NVIDIA GeForce GTX 870M Comparison

The NVIDIA GeForce GTX 870M and AMD Radeon Vega 8 are evenly matched in aggregate, with each winning one benchmark in direct comparison. The GTX 870M takes a commanding lead in OpenCL compute, while the Vega 8 dominates Metal performance. Overall average scores favor the GTX 870M by 8.0%, placing it in the 48th percentile of all GPUs versus the Vega 8’s 45th percentile, but the Vega 8’s newer architecture and API support make it the more versatile option for modern workloads.

Head-to-Head Benchmarks

The most decisive result comes from Geekbench OpenCL, where the NVIDIA GeForce GTX 870M scores 12,630 against the AMD Radeon Vega 8’s 8,822. That is a 43.2% advantage for NVIDIA, representing the largest performance gap in any shared test. This margin reflects the GTX 870M’s dedicated GDDR5 memory and far higher raw shading throughput rather than any architectural elegance — the older Kepler design simply has more execution resources to throw at compute tasks.

The tables turn completely in Geekbench Metal, where the AMD Radeon Vega 8 posts 10,706 versus the GTX 870M’s 7,288. The delta here is 31.9% in AMD’s favor, a substantial reversal. Metal is a low-level API that rewards modern hardware features, and the Vega 8’s GCN 5.0 architecture with full DirectX 12_1 support clearly extracts more efficiency from the same class of workload. The GTX 870M’s Kepler core, limited to DirectX 11_0 feature levels, struggles to keep pace despite its brute-force compute advantage.

Looking at the broader benchmark landscape, the GTX 870M’s average score of 9,959 sits just 0.5% below the AMD Radeon Pro 5300M and 0.8% below the NVIDIA Quadro K5100M, while running 1.1% ahead of the NVIDIA Quadro 6000. The Vega 8’s 9,221 average is virtually tied with the AMD Radeon 890M (0.1% ahead) and trails the NVIDIA GeForce GTX 960 by 0.6% and the GTX 465 by 0.8%. Notably, the Vega 8 sits just 0.9% behind the NVIDIA GeForce GTX 850M, a dedicated mobile GPU from the same era as the 870M — which puts the integrated part’s achievement into perspective.

Architecture Differences

The GTX 870M uses the GK104 chip on a 28 nm TSMC process, packing 3,540 million transistors into a 294 mm² die. The Vega 8 employs the Raven chip on GlobalFoundries’ 14 nm node, with 4,940 million transistors across a smaller 210 mm² die. This yields a transistor density of 23.5 million per mm² for AMD versus 12.0 million per mm² for NVIDIA — the Vega 8 crams nearly twice the transistors into roughly two-thirds the silicon area.

Core configurations differ starkly. The GTX 870M fields 1,344 shading units, 112 texture mapping units, and 24 ROPs, clocked at 941 MHz base and 967 MHz boost. The Vega 8 operates with just 512 shading units, 32 TMUs, and 8 ROPs, but boosts to 1,100 MHz from a 300 MHz base — a 3.7× clock ramp that partially compensates for the fewer units. Pixel rate tells the story: the GTX 870M outputs 27.08 GPixel/s versus 8.80 GPixel/s for the Vega 8, and texture rate is 108.3 GTexel/s versus 35.20 GTexel/s. FP32 compute lands at 2.599 TFLOPS for NVIDIA and 1,126.4 GFLOPS for AMD.

Memory is the biggest structural differentiator. The GTX 870M has 3 GB of dedicated GDDR5 on a 192-bit bus, delivering 120.0 GB/s of bandwidth at 5 Gbps effective. The Vega 8 uses System Shared memory, with bandwidth described as System Dependent and no dedicated VRAM. This means the Vega 8’s performance scales with the host system’s RAM configuration, while the GTX 870M offers predictable, consistent bandwidth. The GTX 870M’s TDP is 100 W, whereas the Vega 8 draws just 25 W — a 4× power gap that reflects the integrated versus discrete design philosophy.

The Vega 8 supports DirectX 12_1, while the GTX 870M is limited to DirectX 12 (11_0). Both support OpenGL 4.6, but the Vega 8 carries Vulkan 1.3 versus the GTX 870M’s Vulkan 1.2.175. The GTX 870M also supports FP16 at 2.253 TFLOPS (2:1 rate), a feature absent from the older Kepler part. The GTX 870M ships as an MXM Module with an MXM-B (3.0) interface, while the Vega 8 is an IGP with no bus interface beyond the motherboard itself.

Where Each One Wins

The GTX 870M is the clear winner for OpenCL compute tasks. Its 43.2% lead over the Vega 8 in that specific benchmark demonstrates superior raw throughput for general-purpose GPU workloads, scientific simulations, and any application that leverages OpenCL for acceleration. The dedicated 120.0 GB/s of memory bandwidth and 2.599 TFLOPS of FP32 performance give it a decisive edge in memory-bound and compute-heavy scenarios. The 28 nm Kepler architecture, despite its age, compensates for its older design through sheer resource count — 1,344 shading units versus 512.

The Vega 8 wins in Metal environments, where its 31.9% advantage over the GTX 870M indicates better optimization for Apple’s low-level graphics API. This makes it the superior choice for Metal-based rendering, video processing, and creative applications that target that API. The Vega 8’s 14 nm process and GCN 5.0 architecture also provide better power efficiency, with a 25 W TDP versus 100 W, meaning it can sustain performance in thermally constrained systems where the GTX 870M would throttle or require active cooling.

For gaming, the picture is mixed based on available data. The Vega 8’s Vulkan 1.3 support gives it access to the latest graphics features, while the GTX 870M’s older Vulkan 1.2.175 limits compatibility with bleeding-edge titles. However, the GTX 870M’s higher fill rates and dedicated memory make it likely more capable in traditional rasterized workloads that favor bandwidth and pixel throughput. The Vega 8’s System Shared memory is a double-edged sword — it can leverage fast system RAM but suffers when system memory is slow or shared with CPU demands.

The Verdict

Choose the NVIDIA GeForce GTX 870M if your priority is raw compute performance in OpenCL-heavy applications. The data shows it is 43.2% faster than the Vega 8 in that benchmark, and its average score of 9,959 places it 8.0% ahead of the AMD part’s 9,221. The 3 GB of dedicated GDDR5 memory with 120.0 GB/s bandwidth provides consistent performance that does not depend on host system configuration. This GPU is the safer pick for users running CUDA-adjacent workloads, legacy OpenCL software, or any task where consistent memory bandwidth is critical.

Choose the AMD Radeon Vega 8 if Metal performance or power efficiency is your priority. Its 31.9% lead over the GTX 870M in Geekbench Metal is substantial, and its 25 W TDP makes it viable in thin-and-light systems where the GTX 870M’s 100 W requirement would be prohibitive. The Vega 8 also offers superior API support — DirectX 12_1 and Vulkan 1.3 — which future-proofs it for newer software titles. The trade-off is clear: you sacrifice 43.2% in OpenCL speed and 8.0% in average benchmark score, but you gain Metal performance, modern API coverage, and a fraction of the power draw.

For most users today, the Vega 8 is the more practical choice due to its architectural recency and API support, but the GTX 870M remains a capable dedicated solution for compute-oriented tasks. The GTX 870M’s end-of-life status and 2014 release date mean it lacks the feature set of the Vega 8, which launched in 2018 and benefits from four additional years of architectural evolution.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA GeForce GTX 870M averages 9,959 points, which is 8.0% higher than the AMD Radeon Vega 8’s 9,221 points.

Q: How large is the GTX 870M’s lead in OpenCL performance?

A: The GTX 870M scores 12,630 in Geekbench OpenCL versus 8,822 for the Vega 8, a 43.2% advantage for NVIDIA.

Q: Does the AMD Radeon Vega 8 support newer graphics APIs?

A: Yes, the Vega 8 supports DirectX 12_1 and Vulkan 1.3, while the GTX 870M is limited to DirectX 12 (11_0) and Vulkan 1.2.175.

Q: What is the memory configuration difference between the two?

A: The GTX 870M has 3 GB of dedicated GDDR5 on a 192-bit bus with 120.0 GB/s bandwidth, while the Vega 8 uses System Shared memory with System Dependent bandwidth.

Q: How do their power requirements compare?

A: The GTX 870M has a 100 W TDP, while the Vega 8 draws only 25 W — a 4× difference in favor of AMD.

Q: Which GPU has more shading units?

A: The GTX 870M has 1,344 shading units, compared to 512 on the Vega 8, though the Vega 8 boosts to a higher 1,100 MHz clock versus 967 MHz for NVIDIA.

Specification Differences

| Specification | NVIDIA GeForce GTX 870M | AMD Radeon Vega 8 |

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

| Architecture | Kepler | GCN 5.0 |

| Process Node | 28 nm | 14 nm |

| Foundry | TSMC | GlobalFoundries |

| Transistors | 3,540 million | 4,940 million |

| Die Size | 294 mm² | 210 mm² |

| Transistor Density | 12.0M / mm² | 23.5M / mm² |

| Base Clock | 941 MHz | 300 MHz |

| Boost Clock | 967 MHz | 1100 MHz |

| Memory Size | 3 GB | System Shared |

| Memory Type | GDDR5 | System Shared |

| Memory Bus | 192 bit | System Shared |

| Memory Bandwidth | 120.0 GB/s | System Dependent |

| Shading Units | 1344 | 512 |

| TMUs | 112 | 32 |

| ROPs | 24 | 8 |

| Pixel Rate | 27.08 GPixel/s | 8.800 GPixel/s |

| Texture Rate | 108.3 GTexel/s | 35.20 GTexel/s |

| FP32 Performance | 2.599 TFLOPS | 1,126.4 GFLOPS |

| FP16 Performance | N/A | 2.253 TFLOPS (2:1) |

| TDP | 100 W | 25 W |

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

| Vulkan Support | 1.2.175 | 1.3 |

| Bus Interface | MXM-B (3.0) | IGP |

| Release Date | 2014-03-11 | 2018-02-11 |