NVIDIA GeForce GTX 870M vs NVIDIA GeForce GTX 950M Comparison

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 870M leads with an average benchmark score of 9959, while the NVIDIA GeForce GTX 950M trails at 8135. That places the 870M in the 48th percentile of all GPUs, versus the 950M’s 42nd percentile.

Q: How do these two compare in the OpenCL benchmark?

A: The GTX 870M scores 12630 in Geekbench OpenCL, which is 29.6% ahead of the GTX 950M’s 9745. This is the only head-to-head benchmark recorded, and the 870M wins it outright.

Q: What are the closest rivals for each card according to the database?

A: The GTX 870M’s nearest rival is the AMD Radeon Pro 5300M, which scores 10013, a delta of -0.5%. The GTX 950M’s nearest rival is the AMD Radeon R9 M360, scoring 8129, a delta of 0.1%.

Q: Do both cards support DirectX 12?

A: Yes, both list DirectX 12 (11_0) support. They also both support OpenGL 4.6, though the GTX 950M lists Vulkan 1.4 while the GTX 870M lists Vulkan 1.2.175.

Q: Which GPU has the smaller die size?

A: The GTX 950M has a 148 mm² die, while the GTX 870M has a 294 mm² die. Despite the smaller die, the 950M has a slightly higher transistor density at 12.6M / mm² versus 12.0M / mm².

Q: What is the production status of both cards?

A: Both are listed as end-of-life. The GTX 870M was released on 2014-03-11, and the GTX 950M followed on 2015-03-12.

Architecture Differences

The two NVIDIA mobile GPUs represent different architectural generations despite sharing the same 28 nm TSMC process node. The GTX 870M is built on the GK104 chip with the Kepler architecture, while the GTX 950M uses the GM107 chip with the Maxwell architecture.

The transistor counts diverge sharply. The GTX 870M packs 3,540 million transistors on a 294 mm² die, yielding a density of 12.0M / mm². The GTX 950M contains 1,870 million transistors on a 148 mm² die, with a density of 12.6M / mm². The 870M’s larger die and higher transistor budget enable a substantially wider execution engine.

Shading unit counts underline this difference. The GTX 870M has 1344 shading units, 112 texture mapping units, and 24 ROPs. The GTX 950M has 640 shading units, 40 TMUs, and 16 ROPs. The 870M more than doubles the shading unit count and nearly triples the TMU count.

Memory configurations also differ fundamentally. The GTX 870M uses 3 GB of GDDR5 on a 192-bit bus, achieving 120.0 GB/s of bandwidth. The GTX 950M uses 4 GB of DDR3 on a 128-bit bus, achieving only 28.80 GB/s. The clock speeds tell a different story: the 870M runs at 941 MHz base and 967 MHz boost, while the 950M runs higher at 993 MHz base and 1124 MHz boost. However, the 870M’s memory clock is 1250 MHz with 5 Gbps effective, versus the 950M’s 900 MHz with 1800 Mbps effective.

The slot and interface designs differ as well. The GTX 870M is an MXM Module with an MXM-B (3.0) bus interface, while the GTX 950M is an IGP form factor with a PCIe 3.0 x8 interface. The 870M carries a 100 W TDP, while the 950M draws 75 W. Neither card requires external power connectors.

Head-to-Head Benchmarks

The recorded head-to-head data contains a single benchmark: Geekbench OpenCL. The GTX 870M delivers a score of 12630, while the GTX 950M scores 9745. That translates to a 29.6% advantage for the 870M, a decisive margin in a compute workload that stresses raw throughput.

This result aligns with the architectural disparity. The 870M’s 1344 shading units and 120.0 GB/s memory bandwidth provide a substantial compute advantage over the 950M’s 640 shading units and 28.80 GB/s bandwidth. The 950M’s higher boost clock of 1124 MHz does not compensate for the 870M’s wider execution resources and faster memory subsystem.

In the broader benchmark context, the GTX 870M’s average score of 9959 places it 1.1% above the NVIDIA Quadro 6000 (9846) and 1.1% below the AMD Radeon R9 M375 (10070). The GTX 950M’s average of 8135 sits 0.1% above the AMD Radeon R9 M360 (8129) and 0.4% above the NVIDIA GeForce 945M (8099). The 870M’s nearest rival, the AMD Radeon Pro 5300M, scores 10013, which is 0.5% higher than the 870M.

The delta between the two cards in OpenCL, 29.6%, is far larger than the deltas observed among their respective nearest rivals, none of which exceed 1.1%. This indicates that the performance gap between the 870M and 950M is not marginal; it is a clear tier separation.

The Verdict

The data points to a straightforward conclusion: the NVIDIA GeForce GTX 870M is the faster GPU. Its average benchmark score of 9959 versus 8135 for the GTX 950M represents a substantial lead. In the only direct head-to-head test, the 870M wins by 29.6% in OpenCL, a result consistent with its higher shading unit count, wider memory bus, and greater memory bandwidth.

The GTX 870M also holds a higher percentile ranking at 48 versus 42, meaning it outperforms a larger share of the GPU population. Its nearest rivals include the AMD Radeon Pro 5300M (10013, -0.5%) and the NVIDIA Quadro K5100M (10043, -0.8%), both of which are close to the 870M’s performance. The GTX 950M’s nearest rivals are all lower-tier parts, such as the AMD Radeon R9 M360 (8129, 0.1%) and the NVIDIA GeForce 945M (8099, 0.4%).

The GTX 950M does have advantages in specific areas. It runs at higher clock speeds, uses less power (75 W versus 100 W), and has a smaller die with slightly higher transistor density. It also offers 4 GB of memory versus 3 GB, and supports Vulkan 1.4 versus 1.2.175. But these advantages do not translate into benchmark superiority. The 950M’s DDR3 memory and narrower bus are severe limitations, and its average score reflects that.

For users prioritizing raw compute performance and higher overall GPU ranking, the GTX 870M is the clear choice. For users prioritizing lower power draw and a more compact IGP form factor, the GTX 950M is the only option of the two, though it comes with a measurable performance penalty.

Specification Differences

The two GPUs differ across nearly every major specification. The GTX 870M uses the GK104 chip with Kepler architecture, while the GTX 950M uses the GM107 chip with Maxwell architecture. The 870M belongs to the GeForce 800M generation, the 950M to the GeForce 900M generation.

Transistor count: 3,540 million versus 1,870 million. Die size: 294 mm² versus 148 mm². Transistor density: 12.0M / mm² versus 12.6M / mm².

Base clock: 941 MHz versus 993 MHz. Boost clock: 967 MHz versus 1124 MHz. Memory clock: 1250 MHz (5 Gbps effective) versus 900 MHz (1800 Mbps effective).

Memory size: 3 GB versus 4 GB. Memory type: GDDR5 versus DDR3. Bus width: 192 bit versus 128 bit. Bandwidth: 120.0 GB/s versus 28.80 GB/s.

Shading units: 1344 versus 640. TMUs: 112 versus 40. ROPs: 24 versus 16.

Pixel rate: 27.08 GPixel/s versus 17.98 GPixel/s. Texture rate: 108.3 GTexel/s versus 44.96 GTexel/s. FP32: 2.599 TFLOPS versus 1,438.7 GFLOPS.

TDP: 100 W versus 75 W. Slot width: MXM Module versus IGP. Bus interface: MXM-B (3.0) versus PCIe 3.0 x8.

Vulkan support: 1.2.175 versus 1.4.

Release date: 2014-03-11 versus 2015-03-12. Predecessor: GeForce 700M versus GeForce 800M. Successor: GeForce 900M versus GeForce 10 Mobile.

Both share the same process node (28 nm), foundry (TSMC), DirectX version (12 (11_0)), OpenGL version (4.6), and power connector requirement (None).

Where Each One Wins

The GTX 870M wins in every recorded benchmark category. Its OpenCL score of 12630 versus 9745 is a 29.6% advantage. Its average benchmark score of 9959 versus 8135 is a 22.4% advantage. Its percentile ranking of 48 versus 42 indicates it outperforms a larger fraction of all GPUs.

The 870M’s wins stem from its compute architecture. The 1344 shading units and 112 TMUs provide raw throughput that the 950M cannot match. The 120.0 GB/s memory bandwidth from GDDR5 on a 192-bit bus is over four times the 950M’s 28.80 GB/s, which is critical for memory-intensive workloads. The 870M also delivers higher pixel rate (27.08 GPixel/s versus 17.98 GPixel/s) and texture rate (108.3 GTexel/s versus 44.96 GTexel/s).

The GTX 950M wins in areas outside raw performance. It has a lower TDP of 75 W versus 100 W, making it more suitable for thinner, cooler-running laptops. Its IGP form factor and PCIe 3.0 x8 interface offer integration flexibility that the MXM Module of the 870M does not. It has 4 GB of memory versus 3 GB, which could benefit applications that require larger memory pools. It also supports a newer Vulkan version (1.4 versus 1.2.175). Its higher base and boost clocks (993 MHz and 1124 MHz versus 941 MHz and 967 MHz) and smaller die with higher transistor density (12.6M / mm² versus 12.0M / mm²) indicate a more efficient design per unit area.

The use-case split is clear. The GTX 870M is the choice for compute-heavy workloads, higher frame rates, and any task that leverages memory bandwidth. The GTX 950M is the choice for power-constrained systems, compact form factors, and scenarios where 4 GB of memory capacity outweighs the need for peak throughput. Benchmark data, however, consistently favors the 870M.