NVIDIA GeForce GTX 850M vs NVIDIA GeForce GTX 950M Comparison

The NVIDIA GeForce GTX 850M and GTX 950M are both Maxwell-based mobile GPUs from NVIDIA, but they belong to different generations: the 850M is part of the GeForce 800M series, while the 950M is from the GeForce 900M series. Despite sharing the same GM107 chip, the same 28 nm process node, and identical core counts, benchmark results show a nuanced picture. The recorded data from the database reveals that the GTX 850M actually wins both head-to-head tests, though the margin in one test is substantial and in the other is minimal. The GTX 950M, however, offers a higher average clock profile and more memory, which may matter for certain workloads even if aggregate benchmark scores do not favor it.

Head-to-Head Benchmarks

The database includes two direct comparisons between these GPUs: Geekbench OpenCL and Geekbench Vulkan. In the OpenCL test, the GTX 850M scores 9821 points, while the GTX 950M scores 9745 points. The delta is only 0.8% in favor of the 850M, which is effectively a statistical tie. For practical purposes, OpenCL compute performance is nearly identical between the two, and users should not expect a meaningful difference in applications that rely on this API.

The Vulkan test tells a very different story. The GTX 850M scores 8782 points, while the GTX 950M scores 6525 points. That is a 34.6% advantage for the 850M, a massive gap that dominates the overall comparison. This is surprising given that the 950M has a higher base clock of 993 MHz and a boost clock of 1124 MHz, compared to the 850M’s unspecified base and boost clocks. The 850M’s memory runs at 1001 MHz (2 Gbps effective), while the 950M’s memory runs at 900 MHz (1800 Mbps effective), so the 850M also has a slight memory clock advantage. Still, a 34.6% Vulkan lead is far larger than what clock differences alone would suggest, indicating that the 850M’s drivers or memory subsystem handle Vulkan workloads more efficiently in the recorded tests.

When looking at aggregate average benchmark scores, the situation reverses. The GTX 850M has an average benchmark score of 9302, while the GTX 950M has an average of 8135. That is a 14.3% difference in favor of the 850M, driven almost entirely by the Vulkan result. Interestingly, the 950M’s average score is pulled down by its weak Vulkan showing, despite comparable OpenCL performance. The 850M also ranks higher percentile-wise, sitting at the 46th percentile of all GPUs, while the 950M sits at the 42nd percentile. This means the 850M outperforms a larger share of the GPU population in the database.

In terms of nearest rivals, the 850M’s average score of 9302 is close to the NVIDIA GeForce GTX 465 (9294, delta 0.1%), the AMD Radeon R7 M380 (9313, delta -0.1%), the NVIDIA GeForce GTX 960 (9273, delta 0.3%), and the AMD Radeon Vega 8 (9221, delta 0.9%). These deltas are all under 1%, meaning the 850M sits in a tight cluster of similar-performing GPUs. The 950M’s average of 8135 is comparable to the AMD Radeon R9 M360 (8129, delta 0.1%), the NVIDIA GeForce GTX 980 (8167, delta -0.4%), the NVIDIA GeForce 945M (8099, delta 0.4%), and the NVIDIA GRID K2 (8080, delta 0.7%). The 950M is thus in a lower performance tier than the 850M, despite being a newer generation.

Where Each One Wins

The GTX 850M wins decisively in Vulkan-based workloads. The 34.6% advantage in Geekbench Vulkan suggests that applications using Vulkan for rendering or compute will run noticeably smoother on the 850M. This is the clearest use-case split in the data. For users who play Vulkan titles or use Vulkan-based benchmarks, the 850M is the stronger choice by a wide margin.

The GTX 950M does not win any of the recorded head-to-head benchmarks, but it does have advantages that are not captured in the two tests. It has 4 GB of DDR3 memory, double the 850M’s 2 GB. In scenarios where memory capacity matters, such as running games with high-resolution textures or multitasking with GPU memory-heavy applications, the 950M could hold an edge. The 950M also has a higher pixel rate of 17.98 GPixel/s, compared to 14.43 GPixel/s on the 850M, and a higher texture rate of 44.96 GTexel/s versus 36.08 GTexel/s. These rates indicate that the 950M can fill pixels and process textures faster in raw throughput terms, which may translate to better performance in rasterization-bound tasks, even if the synthetic benchmarks do not reflect it.

The 950M’s FP32 performance is also higher: 1438.7 GFLOPS versus 1154.6 GFLOPS on the 850M. That is a 24.6% advantage in theoretical compute throughput. Again, the Vulkan benchmark contradicts this, but for applications that are well-optimized for the 950M’s clock speeds, the extra compute headroom could be useful.

The 850M wins in memory bandwidth, with 32.03 GB/s versus 28.80 GB/s on the 950M. This is a 11.2% advantage, which partly explains why the 850M performs better in the OpenCL test despite lower clock rates. In memory-intensive compute tasks, the 850M’s faster memory can compensate for its lower core clock.

Architecture Differences

Both GPUs are built on the same GM107 chip, using the Maxwell architecture. The process node is 28 nm, fabricated by TSMC, with 1,870 million transistors on a 148 mm² die. The transistor density is 12.6 million transistors per mm². These are identical figures, so the fundamental silicon is the same. Neither GPU has ray tracing cores or tensor cores, as both predate those features in NVIDIA’s lineup.

The shading unit count is 640 for both, with 40 texture mapping units and 16 raster output units. Again, identical. The differences lie in clock speeds and memory configuration. The 850M specifies a memory clock of 1001 MHz (2 Gbps effective), while the 950M has a memory clock of 900 MHz (1800 Mbps effective). The 850M’s memory bus is 128-bit, and the 950M’s is also 128-bit, so the bandwidth difference comes purely from the clock speed.

The 950M has explicit base and boost clocks of 993 MHz and 1124 MHz, respectively. The 850M’s base and boost clocks are not listed in the database, which makes direct clock comparison incomplete. However, the 950M’s higher throughput rates (pixel, texture, and FP32) indicate that its core clocks are indeed higher in practice. The 850M compensates with faster memory.

In terms of interface, the 850M uses PCIe 3.0 x16, while the 950M uses PCIe 3.0 x8. This is a notable difference: the 850M has double the lane width for host communication. For most gaming workloads, PCIe bandwidth is not a bottleneck, but for GPU compute tasks that transfer large amounts of data between CPU and GPU, the x16 interface could be advantageous.

The 850M has a TDP of 45 W, while the 950M has a TDP of 75 W. This is a significant power difference. The 950M draws 66.7% more power, which means it will generate more heat and require better cooling in a laptop chassis. Both are integrated into the motherboard (slot width is listed as IGP) and use no external power connectors, relying on the system’s power delivery. The 950M’s higher TDP may limit its use in thinner laptops, while the 850M is more power-efficient.

API support is identical: DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4. Both have the same display outputs: portable device dependent. Neither has a launch MSRP listed in the database.

The Verdict

Based strictly on the recorded data, the GTX 850M is the better performer in benchmark terms. It wins both head-to-head tests, with a 0.8% edge in OpenCL and a 34.6% edge in Vulkan. Its average benchmark score is 14.3% higher, and it ranks at the 46th percentile compared to the 950M’s 42nd percentile. For anyone prioritizing synthetic benchmark results, the 850M is the clear choice.

The GTX 950M’s case rests on features not reflected in those scores. It has twice the memory (4 GB vs 2 GB), which is crucial for modern games that exceed 2 GB of VRAM usage. It also has higher pixel, texture, and FP32 rates, suggesting it can handle rasterization and compute workloads that are not memory-bandwidth-limited. The 950M’s higher TDP of 75 W indicates it has more power headroom, which could allow sustained clocks under load if the cooling system is adequate.

Users who play Vulkan-based games or run Vulkan compute should choose the 850M without hesitation. The 34.6% delta is too large to ignore. Users who need more memory capacity for texture-heavy games or who prefer the theoretical compute throughput of the 950M might lean the other way, but they should be aware that the benchmark data does not support that choice in the tests recorded.

The 850M is also more efficient in terms of bandwidth per watt, given its 32.03 GB/s bandwidth on a 45 W TDP, versus 28.80 GB/s on a 75 W TDP. For laptop buyers, the 850M likely offers better battery life and lower heat, though the database does not record thermal or battery performance directly.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The GTX 850M has an average benchmark score of 9302, while the GTX 950M has an average of 8135. The 850M is 14.3% higher.

Q: How much faster is the GTX 850M in Vulkan?

A: In Geekbench Vulkan, the GTX 850M scores 8782 versus 6525 for the GTX 950M, a 34.6% difference in favor of the 850M.

Q: Does the GTX 950M have more memory?

A: Yes, the GTX 950M has 4 GB of DDR3 memory, while the GTX 850M has 2 GB. Both use a 128-bit memory bus.

Q: Are the core counts identical?

A: Yes, both GPUs have 640 shading units, 40 TMUs, and 16 ROPs. The chip is the same GM107 with 1,870 million transistors.

Q: What is the TDP difference?

A: The GTX 850M has a TDP of 45 W, while the GTX 950M has a TDP of 75 W. The 950M draws more power.

Q: Which GPU has a higher memory bandwidth?

A: The GTX 850M has a bandwidth of 32.03 GB/s, while the GTX 950M has 28.80 GB/s. The 850M is 11.2% higher.

Specification Differences

The main differences between the two GPUs are as follows. The GTX 850M belongs to the GeForce 800M generation, while the GTX 950M belongs to the GeForce 900M generation. The 850M’s memory clock is 1001 MHz (2 Gbps effective), whereas the 950M’s is 900 MHz (1800 Mbps effective). The 850M has 2 GB of DDR3 memory, while the 950M has 4 GB. The 850M’s bandwidth is 32.03 GB/s, and the 950M’s is 28.80 GB/s.

The 850M’s pixel rate is 14.43 GPixel/s, compared to 17.98 GPixel/s on the 950M. Its texture rate is 36.08 GTexel/s, versus 44.96 GTexel/s. FP32 performance is 1154.6 GFLOPS on the 850M and 1438.7 GFLOPS on the 950M. The TDP is 45 W for the 850M and 75 W for the 950M. The bus interface is PCIe 3.0 x16 on the 850M and PCIe 3.0 x8 on the 950M. The 850M was released on 2014-03-11, and the 950M on 2015-03-12. The 850M’s predecessor is GeForce 700M, and its successor is GeForce 900M. The 950M’s predecessor is GeForce 800M, and its successor is GeForce 10 Mobile. Both have the same process node, die size, transistor count, architecture, and API support.