NVIDIA GeForce GTX 870M vs NVIDIA GeForce GTX 960 Comparison

The NVIDIA GeForce GTX 870M and the NVIDIA GeForce GTX 960 represent two distinct approaches to mobile and desktop graphics from consecutive generations. The GTX 870M is a Kepler-based MXM module from the GeForce 800M line, while the GTX 960 is a Maxwell 2.0 desktop card from the GeForce 900 series. Benchmark data shows a clear performance hierarchy, with the GTX 960 winning both head-to-head tests, but the specifications reveal a more nuanced story about architectural efficiency versus raw resource allocation.

FAQ

Q: Which GPU wins in the available head-to-head benchmarks?

A: The NVIDIA GeForce GTX 960 wins both head-to-head tests. It scores 8,773 in Geekbench Metal versus 7,288 for the GTX 870M, a delta of -16.9% from the GTX 960's perspective. In Geekbench OpenCL, the GTX 960 scores 18,925 versus 12,630, a -33.3% delta.

Q: How do the average benchmark scores compare between the two?

A: The GTX 870M has a higher average benchmark score of 9,959, placing it in the 48th percentile of all GPUs. The GTX 960 has an average score of 9,273, placing it in the 45th percentile. This is despite the GTX 960 winning both direct comparisons.

Q: What are the closest rivals to the GTX 870M?

A: The GTX 870M's nearest rivals include the AMD Radeon Pro 5300M with an average score of 10,013 (a -0.5% delta), the NVIDIA Quadro K5100M at 10,043 (-0.8%), and the AMD Radeon R9 M375 at 10,070 (-1.1%). The NVIDIA Quadro 6000 trails at 9,846, a +1.1% delta.

Q: What is the closest competitor to the GTX 960?

A: The GTX 960's nearest rival is the NVIDIA GeForce GTX 465 with an average score of 9,294, a -0.2% delta. Other close competitors include the NVIDIA GeForce GTX 850M at 9,302 (-0.3%) and the AMD Radeon R7 M380 at 9,313 (-0.4%). The AMD Radeon Vega 8 trails at 9,221, a +0.6% delta.

Q: Do both GPUs support modern graphics APIs like DirectX 12?

A: Yes, but at different feature levels. The GTX 870M supports DirectX 12 (11_0), while the GTX 960 supports DirectX 12 (12_1). Both support OpenGL 4.6, but the GTX 960 supports Vulkan 1.4, whereas the GTX 870M supports Vulkan 1.2.175.

Q: What is the launch MSRP of the GTX 960?

A: The GTX 960 has a launch MSRP of 199 USD. The GTX 870M has no recorded launch MSRP in the data.

Architecture Differences

The two GPUs are built on fundamentally different architectures despite sharing the same 28 nm process node from TSMC. The GTX 870M uses the GK104 chip based on Kepler architecture, while the GTX 960 uses the GM206 chip based on Maxwell 2.0. This generational shift explains most of the performance and efficiency differences.

The GTX 870M packs more raw hardware resources. It features 1,344 shading units, 112 texture mapping units, and 24 ROPs. In contrast, the GTX 960 has fewer shading units at 1,024, fewer TMUs at 64, but more ROPs at 32. The transistor counts tell a similar story: the GTX 870M has 3,540 million transistors on a 294 mm² die, while the GTX 960 has 2,940 million transistors on a smaller 228 mm² die. This gives the GTX 960 a higher transistor density of 12.9M per mm² versus 12.0M per mm² for the GTX 870M.

Clock speeds favor the GTX 960 significantly. The GTX 960 has a base clock of 1,127 MHz and a boost clock of 1,178 MHz. The GTX 870M operates at a base clock of 941 MHz and a boost clock of 967 MHz. This clock advantage, combined with the architectural improvements of Maxwell 2.0, allows the GTX 960 to achieve competitive performance despite having fewer cores.

Memory configurations also differ notably. The GTX 870M uses a 192-bit memory bus with 3 GB of GDDR5 memory, delivering 120.0 GB/s of bandwidth. The GTX 960 uses a narrower 128-bit bus with 2 GB of GDDR5 memory, but higher memory clocks of 1,753 MHz (7 Gbps effective) yield 112.2 GB/s. The GTX 870M's memory runs at 1,250 MHz (5 Gbps effective).

Compute throughput shows the Kepler card maintaining a slight edge in raw FP32: the GTX 870M delivers 2.599 TFLOPS versus 2.413 TFLOPS for the GTX 960. However, the GTX 960 achieves a higher pixel rate of 37.70 GPixel/s versus 27.08 GPixel/s, thanks to its 32 ROPs. The GTX 870M counters with a higher texture rate of 108.3 GTexel/s versus 75.39 GTexel/s for the GTX 960.

Power and physical design diverge sharply. The GTX 870M is an MXM module rated at 100 W TDP with no power connectors, designed for portable devices. The GTX 960 is a dual-slot desktop card with a 241 mm length, requiring a single 6-pin power connector and a 300 W suggested PSU, with a TDP of 120 W.

The Verdict

The data presents a clear choice based on platform and performance priorities. The GTX 960 wins both available head-to-head benchmarks decisively. In Geekbench Metal, it leads by 1,485 points, and in Geekbench OpenCL, the margin expands to 6,295 points. This represents a -16.9% and -33.3% delta respectively, indicating that the GTX 960 is substantially faster in compute-oriented workloads.

However, the GTX 870M holds a surprising advantage in aggregate benchmark scores. Its average benchmark score of 9,959 is higher than the GTX 960's 9,273, placing it in the 48th percentile versus the 45th percentile. This discrepancy suggests that the GTX 870M performs better across a broader range of tests not covered in the direct comparison, possibly due to its larger memory pool of 3 GB and wider 192-bit memory bus.

For users building or upgrading a desktop system, the GTX 960 is the logical choice from a pure performance standpoint. Its wins in both direct tests, coupled with its more modern DirectX 12 (12_1) support and newer Vulkan 1.4 API, make it the more future-proof option. The GTX 960's higher clock speeds and superior pixel rate also suggest better gaming performance in fill-rate-limited scenarios.

For those in the mobile or compact form-factor space, the GTX 870M presents the only viable option given its MXM module design and lack of power connectors. Its higher FP32 throughput of 2.599 TFLOPS and greater texture rate may benefit specific workloads. The 3 GB memory capacity also provides more headroom for memory-intensive applications, even if bandwidth is only marginally higher at 120.0 GB/s versus 112.2 GB/s.

The choice ultimately depends on whether a user prioritizes the GTX 960's clear wins in the tested benchmarks and newer API support, or the GTX 870M's higher average score, larger memory buffer, and mobile form factor. Desktop users should favor the GTX 960; users requiring a portable solution have only the GTX 870M.

Specification Differences

The two GPUs differ across nearly every major specification category. The GTX 870M uses the GK104 chip on Kepler architecture, while the GTX 960 uses the GM206 chip on Maxwell 2.0. Both are manufactured by TSMC on a 28 nm process, but the GTX 870M has a larger die at 294 mm² versus 228 mm², and a higher transistor count of 3,540 million versus 2,940 million.

Clock speeds show the GTX 960 running significantly faster: 1,127 MHz base and 1,178 MHz boost, compared to 941 MHz base and 967 MHz boost for the GTX 870M. Memory clocks also favor the GTX 960 at 1,753 MHz (7 Gbps effective) versus 1,250 MHz (5 Gbps effective) for the GTX 870M.

The GTX 870M has more shading units (1,344 vs 1,024), more TMUs (112 vs 64), but fewer ROPs (24 vs 32). Memory capacity and bus width favor the GTX 870M with 3 GB on a 192-bit bus, while the GTX 960 has 2 GB on a 128-bit bus. Bandwidth is close but slightly higher on the GTX 870M at 120.0 GB/s versus 112.2 GB/s.

Compute rates differ by metric: the GTX 870M has higher FP32 (2.599 TFLOPS vs 2.413 TFLOPS) and texture rate (108.3 GTexel/s vs 75.39 GTexel/s), but the GTX 960 has a higher pixel rate (37.70 GPixel/s vs 27.08 GPixel/s). TDP and form factor differ substantially, with the GTX 870M at 100 W in an MXM module and the GTX 960 at 120 W in a dual-slot card.

Interface and outputs diverge completely. The GTX 870M uses an MXM-B (3.0) interface with portable-device-dependent display outputs. The GTX 960 uses PCIe 3.0 x16 with 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2 outputs. API support shows the GTX 960 with DirectX 12 (12_1) versus the GTX 870M's DirectX 12 (11_0), and Vulkan 1.4 versus 1.2.175.

Head-to-Head Benchmarks

The direct comparison between the GTX 870M and GTX 960 yields two definitive wins for the GTX 960. The first test, Geekbench Metal, shows the GTX 960 scoring 8,773 against the GTX 870M's 7,288. This -16.9% delta represents a meaningful but not overwhelming advantage. The GTX 960's higher clock speeds and more efficient Maxwell architecture likely contribute to this performance gap in Apple's Metal API.

The second test, Geekbench OpenCL, shows a much more pronounced difference. The GTX 960 achieves 18,925 points, while the GTX 870M manages only 12,630 points. This -33.3% delta is substantial, indicating that the GTX 960 is nearly 50% faster in this compute workload. The result is striking given that the GTX 870M actually has a higher theoretical FP32 throughput of 2.599 TFLOPS versus 2.413 TFLOPS for the GTX 960.

The OpenCL result suggests that architectural efficiency matters more than raw core count or clock speed in certain workloads. The Maxwell 2.0 architecture in the GTX 960 appears to extract more usable performance from its 1,024 shading units than Kepler does from 1,344 units. The GTX 960's higher memory clock of 7 Gbps effective may also play a role, even with its narrower 128-bit bus.

In the aggregate, the GTX 870M still posts a higher average benchmark score of 9,959 versus 9,273 for the GTX 960. This means that while the GTX 960 wins the two tests available in the head-to-head data, the GTX 870M performs better across the broader benchmark suite used to calculate average scores. The GTX 870M's 3 GB memory capacity could provide an advantage in tests that are memory-capacity-sensitive.

The wins tally confirms the head-to-head dominance: the GTX 960 wins 2 tests, while the GTX 870M wins 0. Yet the percentile rankings tell a slightly different story, with the GTX 870M at the 48th percentile and the GTX 960 at the 45th percentile. This suggests that the GTX 870M's closest rivals are actually slightly faster than the GTX 960's closest rivals, as evidenced by the rival scores: the GTX 870M's rivals average around 10,000, while the GTX 960's rivals average around 9,280.

The data ultimately paints a picture of a desktop card that is faster in the specific tests measured, but a mobile chip that holds its own in the wider benchmark landscape. The GTX 960's OpenCL victory is particularly impressive, showing a 49.9% improvement over the GTX 870M in that test. This is the single largest performance delta observed across all available data points.