NVIDIA GeForce GTX 950M vs NVIDIA GRID K2 Comparison

The NVIDIA GeForce GTX 950M and NVIDIA GRID K2 are both end-of-life NVIDIA parts built on a 28 nm TSMC process, yet they target completely different segments: the GTX 950M is a mobile GPU for laptops, while the GRID K2 is a dual-slot datacenter card with no display outputs. Their average benchmark scores are nearly identical—the GTX 950M sits at 8,135 and the GRID K2 at 8,080—but the underlying hardware and performance profile diverge sharply. The GRID K2 wins the only head-to-head benchmark, yet the GTX 950M holds its own in raw compute tasks, making this comparison a study in how architecture and memory type shape real-world results.

Head-to-Head Benchmarks

The sole direct benchmark comparison available is Geekbench OpenCL, and here the NVIDIA GRID K2 takes the win. The GRID K2 scores 10,602 against the GTX 950M's 9,745, a delta of -8.1% for the GTX 950M. That margin is meaningful—nearly a tenth of the GTX 950M's score—and it reflects the GRID K2's substantially larger hardware resources. The GRID K2 packs 1,536 shading units, 128 texture mapping units, and 32 ROPs, compared to the GTX 950M's 640 shading units, 40 TMUs, and 16 ROPs. In raw shading throughput, the GRID K2 delivers 2.289 TFLOPS of FP32 compute against the GTX 950M's 1,438.7 GFLOPS, a difference of roughly 59% in favor of the GRID K2. The texture rate tells a similar story: 95.36 GTexel/s for the GRID K2 versus 44.96 GTexel/s for the GTX 950M.

However, the GTX 950M is not entirely outclassed. Its average benchmark score of 8,135 actually edges out the GRID K2's 8,080 by 0.7%, per the nearestRivals data. That is a narrow margin, but it indicates that in aggregate workloads—not just the single OpenCL test—the GTX 950M can match or slightly exceed the GRID K2. The Geekbench Vulkan score for the GTX 950M is 6,525, though no Vulkan score exists for the GRID K2 (it only lists a Metal score of 5,557, which is not directly comparable). The OpenCL result is the only apples-to-apples measurement, and it favors the GRID K2 by that 8.1% margin.

Digging into the percentile rankings, both GPUs sit at the 42nd percentile against all GPUs, reinforcing that they are mid-pack performers in the broader landscape. The nearest rivals for the GTX 950M include the AMD Radeon R9 M360 (average score 8,129, delta 0.1%), the NVIDIA GeForce GTX 980 (average score 8,167, delta -0.4%), and the NVIDIA GeForce 945M (average score 8,099, delta 0.4%). The GRID K2's nearest rivals include the GeForce GTX 650 Ti Boost (8,067, delta 0.2%), the GeForce 945M (8,099, delta -0.2%), and the GeForce GTX 880M (8,040, delta 0.5%). Both parts are clustered within a 1% band of their peers, meaning neither is a standout in its class—but the GRID K2's OpenCL superiority is the clearest differentiator.

FAQ

Q: Which GPU wins the head-to-head OpenCL benchmark?

A: The NVIDIA GRID K2 wins Geekbench OpenCL with a score of 10,602, beating the GTX 950M's 9,745 by 8.1%.

Q: How do their average benchmark scores compare?

A: The GTX 950M has a slightly higher average benchmark score of 8,135, while the GRID K2 sits at 8,080, a 0.7% difference in favor of the GTX 950M.

Q: What is the memory configuration difference?

A: Both have 4 GB of memory, but the GTX 950M uses DDR3 on a 128-bit bus with 28.80 GB/s bandwidth, while the GRID K2 uses GDDR5 on a 256-bit bus with 160.0 GB/s bandwidth.

Q: Do both GPUs support the same APIs?

A: Both support DirectX 12 (11_0) and OpenGL 4.6. The GTX 950M supports Vulkan 1.4, while the GRID K2 supports Vulkan 1.2.175.

Q: What is the physical form factor difference?

A: The GTX 950M is an IGP (integrated graphics processor) with no power connectors and portable-device-dependent display outputs. The GRID K2 is a dual-slot card, 267 mm (10.5 inches) long, requiring a 1x 6-pin plus 1x 8-pin power connector and a 550 W suggested PSU, with no display outputs.

Q: Which GPU has more shading units?

A: The GRID K2 has 1,536 shading units, more than double the GTX 950M's 640 shading units.

Architecture Differences

The two GPUs come from different NVIDIA architectures. The GTX 950M is built on the Maxwell architecture with the GM107 chip, while the GRID K2 uses the Kepler architecture with the GK104 chip. This architectural split explains many of their performance characteristics. Maxwell was designed for power efficiency in mobile parts, which is why the GTX 950M has a TDP of just 75 W, while the GRID K2, aimed at datacenter workloads, has a TDP of 225 W.

The transistor counts differ substantially: the GRID K2's GK104 packs 3,540 million transistors on a 294 mm² die, whereas the GTX 950M's GM107 has 1,870 million transistors on a 148 mm² die. Transistor density is comparable—12.6M per mm² for the GTX 950M versus 12.0M per mm² for the GRID K2—but the GRID K2 has nearly twice the die area and nearly twice the transistor budget. That hardware advantage translates into the GRID K2's higher pixel rate (23.84 GPixel/s versus 17.98 GPixel/s) and texture rate (95.36 GTexel/s versus 44.96 GTexel/s).

Memory architecture is another key divider. The GTX 950M uses DDR3 with a 128-bit bus, yielding 28.80 GB/s of bandwidth, while the GRID K2 uses GDDR5 with a 256-bit bus, delivering 160.0 GB/s—over 5.5 times the bandwidth. This is critical for compute-heavy workloads that thrash memory, even if the GTX 950M's lower bandwidth is adequate for typical laptop gaming. The GRID K2 also has a wider PCIe interface (PCIe 3.0 x16 versus PCIe 3.0 x8), which further favors it in data-transfer-intensive scenarios.

Clock speeds tell a nuanced story. The GTX 950M lists a base clock of 993 MHz and a boost clock of 1124 MHz, while the GRID K2 has no base or boost clocks listed in the data, only a memory clock of 1250 MHz (5 Gbps effective). The GTX 950M's memory clock is 900 MHz (1800 Mbps effective). The GRID K2's lack of published core clocks suggests it may be constrained by power or thermal limits in its intended server environment, but the sheer number of shaders compensates.

The Verdict

The data points to a clear split: the NVIDIA GRID K2 is the superior compute performer in the one benchmark where both are measured, while the GTX 950M holds a marginal edge in average benchmark score. For anyone prioritizing raw OpenCL throughput, the GRID K2 is the pick—its 10,602 score versus 9,745 is a decisive 8.1% advantage, backed by roughly 59% more FP32 compute (2.289 TFLOPS versus 1,438.7 GFLOPS) and 5.5 times the memory bandwidth (160.0 GB/s versus 28.80 GB/s). The GRID K2's 1,536 shading units and 128 TMUs make it a heavyweight for parallel workloads that scale with core count.

However, the GTX 950M's higher average benchmark score (8,135 versus 8,080) and its 0.7% delta in its favor suggest that in mixed or real-world tasks, the Maxwell architecture's efficiency narrows the gap. The GTX 950M also offers Vulkan 1.4 support versus the GRID K2's Vulkan 1.2.175, which could matter for newer applications. The GTX 950M's 75 W TDP and IGP form factor mean it can fit into laptops with no external power connectors, whereas the GRID K2 demands a dual-slot card with 1x 6-pin and 1x 8-pin connectors, a 550 W PSU, and has no display outputs—it is strictly a compute or virtualization card.

In terms of lifecycle, the GTX 950M was released on 2015-03-12, while the GRID K2 came earlier on 2013-05-10. The GTX 950M has a predecessor (GeForce 800M) and successor (GeForce 10 Mobile), while the GRID K2 has neither listed. The GRID K2's launch MSRP was 5,199 USD, a figure that underscores its enterprise positioning, but this comparison should not be read as a value assessment—the performance gap in OpenCL is real and measurable.

Specification Differences

The two GPUs differ in nearly every measurable hardware category except memory size (both 4 GB) and process node (both 28 nm, both TSMC). Key divergences include:

• Chip and architecture: GTX 950M uses GM107 (Maxwell); GRID K2 uses GK104 (Kepler)
• Transistors: GTX 950M has 1,870 million; GRID K2 has 3,540 million
• Die size: GTX 950M is 148 mm²; GRID K2 is 294 mm²
• Shading units: GTX 950M has 640; GRID K2 has 1,536
• TMUs: GTX 950M has 40; GRID K2 has 128
• ROPs: GTX 950M has 16; GRID K2 has 32
• Pixel rate: GTX 950M is 17.98 GPixel/s; GRID K2 is 23.84 GPixel/s
• Texture rate: GTX 950M is 44.96 GTexel/s; GRID K2 is 95.36 GTexel/s
• FP32: GTX 950M is 1,438.7 GFLOPS; GRID K2 is 2.289 TFLOPS
• Memory type: GTX 950M is DDR3; GRID K2 is GDDR5
• Memory bus: GTX 950M is 128-bit; GRID K2 is 256-bit
• Memory bandwidth: GTX 950M is 28.80 GB/s; GRID K2 is 160.0 GB/s
• Memory clock: GTX 950M is 900 MHz (1800 Mbps effective); GRID K2 is 1250 MHz (5 Gbps effective)
• TDP: GTX 950M is 75 W; GRID K2 is 225 W
• Slot width: GTX 950M is IGP; GRID K2 is dual-slot
• Power connectors: GTX 950M has none; GRID K2 has 1x 6-pin + 1x 8-pin
• Suggested PSU: Not listed for GTX 950M; GRID K2 requires 550 W
• Bus interface: GTX 950M is PCIe 3.0 x8; GRID K2 is PCIe 3.0 x16
• Display outputs: GTX 950M is portable-device-dependent; GRID K2 has no outputs
• Vulkan version: GTX 950M supports 1.4; GRID K2 supports 1.2.175
• Release date: GTX 950M is 2015-03-12; GRID K2 is 2013-05-10
• Launch MSRP: Not available for GTX 950M; GRID K2 was 5,199 USD

Where Each One Wins

NVIDIA GRID K2 wins in:

• Raw compute throughput: OpenCL score of 10,602 versus 9,745, an 8.1% lead
• Shader and texture processing: 1,536 shading units and 128 TMUs enable 95.36 GTexel/s, over double the GTX 950M's 44.96 GTexel/s
• Memory bandwidth: 160.0 GB/s is critical for large datasets, and the GDDR5 type and 256-bit bus outperform the GTX 950M's DDR3 and 128-bit bus
• Pixel fill rate: 23.84 GPixel/s versus 17.98 GPixel/s
• Datacenter integration: PCIe 3.0 x16 interface and no display outputs suit server workloads

NVIDIA GeForce GTX 950M wins in:

• Average benchmark score: 8,135 versus 8,080, a 0.7% edge that suggests better balance across varied tasks
• Power efficiency: 75 W TDP versus 225 W means far lower thermal and power demands
• Form factor: IGP design with no power connectors fits mobile devices; the GRID K2 requires a dual-slot card and external power
• API support: Vulkan 1.4 versus 1.2.175 offers broader compatibility with newer software
• Portability: Portable-device-dependent display outputs allow for direct display connection, unlike the GRID K2's lack of outputs

The GRID K2 is the clear choice for compute-heavy, memory-intensive workloads where its massive shader count and bandwidth pay off. The GTX 950M is better suited for mobile scenarios where power constraints and display output matter, and where its slightly higher aggregate benchmark score indicates it does not sacrifice overall performance. Neither GPU is a top-tier part—both sit at the 42nd percentile—but the GRID K2's OpenCL dominance is the standout statistical fact, while the GTX 950M's efficiency and API currency make it the more versatile everyday option.