NVIDIA GeForce GTX 850M vs NVIDIA GeForce GTX 960M Comparison

The NVIDIA GeForce GTX 960M and NVIDIA GeForce GTX 850M are both Maxwell-based mobile GPUs from NVIDIA, but the data shows they occupy distinct performance tiers despite sharing a chip and core configuration. The 960M, part of the GeForce 900M generation, posts a higher average benchmark score of 9645 compared to the 850M's 9302, placing the 960M ahead by 3.7% overall. However, the benchmark results are not uniformly in favor of the newer part, and the differences in memory architecture and power envelope create a more nuanced picture than the averages alone suggest.

Head-to-Head Benchmarks

The head-to-head data reveals a split decision between the two GPUs. In the Geekbench OpenCL test, the GTX 960M delivers a score of 11045, decisively outperforming the GTX 850M's 9821. This represents a 12.5% advantage for the 960M, which aligns with its higher clock speeds and significantly faster memory subsystem. The 960M's boost clock of 1176 MHz, combined with its 5 Gbps effective memory speed, allows it to pull ahead in compute-heavy workloads that stress raw throughput. The 850M's 2 Gbps effective memory speed is a substantial bottleneck in comparison, and the OpenCL result reflects that gap clearly.

Conversely, in the Geekbench Vulkan test, the tables turn. The GTX 850M scores 8782, beating the GTX 960M's 8245 by 6.1%. This is a notable reversal, as the 960M is the newer and generally faster part on paper. The Vulkan result suggests that the 850M's lower power consumption and potentially different driver optimization for that API contribute to its advantage. It is importantly the 850M's architecture, while sharing the same GM107 chip, is configured with a lower thermal design power (45 W versus 75 W), which might influence how the GPU manages its clock behavior under API-specific workloads. The data does not specify why this inversion occurs, but the benchmark result is unambiguous: the 850M wins this round.

Looking at the broader context, the 960M's OpenCL victory is more substantial in percentage terms (12.5%) than the 850M's Vulkan win (6.1%), which explains the 960M's higher average benchmark score. In the average score calculation, the 960M's 9645 places it 0.1% ahead of the NVIDIA Quadro K5000 and 0.1% behind the AMD Radeon Pro WX 2100, showing it sits in a tightly packed cluster of mid-range GPUs. The 850M's 9302 average places it 0.1% ahead of the NVIDIA GeForce GTX 465 and 0.3% ahead of the NVIDIA GeForce GTX 960, indicating it holds its own against older desktop parts. When comparing the two directly, the 960M's average is 3.7% higher than the 850M's, a margin that is meaningful but not overwhelming.

The Verdict

From the data, the choice between these two GPUs depends heavily on the workload. For users prioritizing OpenCL compute performance, the GTX 960M is the clear winner, offering a 12.5% advantage in that specific benchmark. Its higher memory bandwidth (80.19 GB/s versus 32.03 GB/s) and faster boost clock make it the superior option for tasks that rely on memory-intensive parallel processing. The 960M's average benchmark score of 9645, which places it in the 46th percentile of all GPUs, is also higher than the 850M's 9302, reinforcing its overall performance lead.

However, for those whose primary use case involves Vulkan-based applications, the GTX 850M's 6.1% edge in that benchmark is a compelling counterpoint. The 850M's lower 45 W TDP also makes it a more power-efficient option for thin-and-light laptops, where thermal constraints are critical. The data shows the 850M achieves its Vulkan win despite having a lower memory clock and half the memory capacity (2 GB versus 4 GB), which suggests its driver stack and power management are well-tuned for that API. Ultimately, the 960M is the better all-around performer for compute-heavy tasks, while the 850M offers a specific Vulkan advantage and lower power draw. A user who values raw OpenCL throughput should pick the 960M; a user who prioritizes Vulkan performance or battery life should consider the 850M.

Architecture Differences

Both GPUs are built on the same GM107 chip and use the Maxwell architecture, fabricated on a 28 nm process at TSMC. The transistor count is identical at 1,870 million, and the die size is the same at 148 mm², resulting in a matching transistor density of 12.6M / mm². The core configuration is also identical: 640 shading units, 40 texture mapping units, and 16 render output units. This means the architecture's raw compute resources are the same on both chips.

The key architectural difference lies in the memory subsystem and the clock behavior. The GTX 960M uses 4 GB of GDDR5 memory with a 128-bit bus, achieving a bandwidth of 80.19 GB/s. The GTX 850M, in contrast, uses 2 GB of DDR3 memory with the same 128-bit bus, but its bandwidth drops to 32.03 GB/s. This is a dramatic difference, as the 960M's GDDR5 memory provides more than double the bandwidth of the 850M's DDR3. The memory clock also differs: the 960M runs at 1253 MHz (5 Gbps effective), while the 850M runs at 1001 MHz (2 Gbps effective). This memory configuration is the primary driver of the 960M's superior OpenCL performance.

The pixel rate and texture rate further illustrate the performance gap. The 960M achieves 18.82 GPixel/s and 47.04 GTexel/s, while the 850M manages 14.43 GPixel/s and 36.08 GTexel/s. These differences stem from the 960M's higher clock speeds; the 960M has a base clock of 1097 MHz and a boost of 1176 MHz, whereas the 850M's base and boost clocks are not specified in the data. The fp32 compute performance is also higher on the 960M at 1.505 TFLOPS versus 1,154.6 GFLOPS on the 850M, reflecting the clock advantage.

Specification Differences

The specification sheets reveal several distinct differences between the two GPUs. The most obvious is the memory configuration: the 960M offers 4 GB of GDDR5, while the 850M offers 2 GB of DDR3. This affects both capacity and bandwidth, as detailed above. The 960M's memory bandwidth is 80.19 GB/s, which is 2.5 times higher than the 850M's 32.03 GB/s. The memory clock also differs, with the 960M at 1253 MHz (5 Gbps effective) versus the 850M at 1001 MHz (2 Gbps effective).

The power envelope is another significant differentiator. The 960M has a TDP of 75 W, while the 850M draws only 45 W. This 30 W difference has implications for laptop cooling and battery life, though the data does not quantify the real-world impact. The slot width also differs: the 960M uses an MXM Module form factor with an MXM-B (3.0) bus interface, whereas the 850M is an IGP (integrated graphics processor) with a PCIe 3.0 x16 interface. This suggests the 960M is designed as a replaceable module, while the 850M is soldered onto the motherboard.

The release dates are also different: the 850M launched on 2014-03-11, and the 960M launched on 2015-03-12, exactly one year later. Their generational placement differs accordingly, with the 850M in the GeForce 800M series and the 960M in the GeForce 900M series. The 850M's predecessor is GeForce 700M and its successor is GeForce 900M, while the 960M's predecessor is GeForce 800M and its successor is GeForce 10 Mobile. Both GPUs share the same API support, including DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4, and both have portable-device-dependent display outputs.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA GeForce GTX 960M has an average benchmark score of 9645, which is 3.7% higher than the GTX 850M's average of 9302.

Q: How do the two GPUs compare in the Geekbench OpenCL test?

A: The GTX 960M scores 11045 in Geekbench OpenCL, outperforming the GTX 850M's 9821 by a margin of 12.5%.

Q: Does the GTX 850M win any benchmark against the GTX 960M?

A: Yes, in the Geekbench Vulkan test, the GTX 850M scores 8782, which is 6.1% higher than the GTX 960M's 8245.

Q: What is the difference in memory bandwidth between the two GPUs?

A: The GTX 960M has a memory bandwidth of 80.19 GB/s using GDDR5 memory, while the GTX 850M has a bandwidth of 32.03 GB/s using DDR3 memory.

Q: Are the core configurations of the two GPUs identical?

A: Yes, both the GTX 960M and GTX 850M have 640 shading units, 40 TMUs, and 16 ROPs, based on the same GM107 chip with 1,870 million transistors.

Q: What is the TDP difference between the GTX 960M and GTX 850M?

A: The GTX 960M has a TDP of 75 W, while the GTX 850M has a lower TDP of 45 W, making the 850M more power-efficient on paper.