GPU Comparison

The AMD FirePro M6100 and NVIDIA GeForce GTX 950 are two end-of-life GPUs from different eras and market segments. The FirePro M6100 is a professional mobile workstation part built on GCN 2.0, while the GTX 950 is a consumer desktop graphics card based on Maxwell 2.0. Despite their different origins, benchmark data places them close in overall average score, making for a nuanced comparison.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD FirePro M6100 has an average benchmark score of 13354, while the NVIDIA GeForce GTX 950 has an average of 13189. The FirePro M6100 holds a slight edge of about 1.2% in this metric.

Q: How do they compare in the Geekbench OpenCL benchmark?

A: The GTX 950 scores 15662 in Geekbench OpenCL, which is 14.7% higher than the FirePro M6100's score of 13354. This is the only head-to-head benchmark available, and it favors the NVIDIA card.

Q: What is the performance percentile for each GPU relative to all GPUs?

A: The FirePro M6100 sits at the 54th percentile, slightly above the GTX 950's 53rd percentile. Both are near the middle of the performance distribution, indicating they are comparable mid-range parts.

Q: Which GPU has a higher transistor count and larger die size?

A: The GTX 950 uses 2,940 million transistors on a 228 mm² die, while the FirePro M6100 uses 2,080 million transistors on a 160 mm² die. The NVIDIA chip is substantially larger in both respects.

Q: Are there differences in API support between the two?

A: Yes. The FirePro M6100 supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.2.170. The GTX 950 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, giving it a higher DirectX feature level and a newer Vulkan version.

Q: Which GPU has more shading units and texture mapping units?

A: The FirePro M6100 has 896 shading units and 56 TMUs, compared to the GTX 950's 768 shading units and 48 TMUs. The AMD card has a 16.7% advantage in both counts.

Architecture Differences

The two GPUs represent fundamentally different architectural approaches from their respective manufacturers. The AMD FirePro M6100 is built on GCN 2.0 architecture, using the Emerald chip. The NVIDIA GeForce GTX 950 uses Maxwell 2.0 architecture with the GM206 chip. Both are manufactured by TSMC on the same 28 nm process node, but their designs diverge significantly from there.

The transistor budgets tell the story clearly. The GTX 950 packs 2,940 million transistors onto a 228 mm² die, yielding a transistor density of 12.9M per mm². The FirePro M6100 is more modest, with 2,080 million transistors on a 160 mm² die, achieving a slightly higher density of 13.0M per mm². Despite having fewer transistors overall, the AMD chip's smaller size means its density is nearly identical.

In terms of compute resources, the FirePro M6100 allocates its transistors toward a wider array of processing units. It features 896 shading units, 56 texture mapping units, and 16 raster operation pipelines. The GTX 950, by contrast, has 768 shading units, 48 TMUs, but doubles the ROP count to 32. This suggests NVIDIA optimized for pixel throughput while AMD focused on raw shader and texture processing.

Clock behavior also differs. The GTX 950 has explicit base and boost clocks of 1024 MHz and 1188 MHz respectively, while the FirePro M6100's base and boost clocks are not specified in the data. The NVIDIA card's memory runs at 1653 MHz (6.6 Gbps effective), whereas the AMD part's memory clock is 1500 MHz (6 Gbps effective).

The interface and form factor reflect their intended environments. The FirePro M6100 is an MXM Module with an MXM-B (3.0) bus interface and no power connectors, designed for laptops. The GTX 950 is a dual-slot desktop card with a PCIe 3.0 x16 interface, a single 6-pin power connector, and a suggested PSU of 250 W. The GTX 950 has a 90 W TDP, while the FirePro M6100's TDP is not listed.

Head-to-Head Benchmarks

The only direct benchmark comparison available is the Geekbench OpenCL test, and it produces a clear winner. The NVIDIA GeForce GTX 950 scores 15662, while the AMD FirePro M6100 scores 13354. This represents a 14.7% advantage for the GTX 950, a substantial margin in raw compute performance.

This result is interesting given the FirePro M6100's higher shading unit count and higher theoretical FP32 throughput. The AMD card is rated at 1.971 TFLOPS, while the GTX 950 is rated at 1.825 TFLOPS. Despite the AMD card's 8% advantage in raw floating-point capability, the NVIDIA card wins decisively in the actual OpenCL workload, suggesting better architectural efficiency or driver optimization in this test.

The GTX 950 also benefits from higher memory bandwidth. Its 105.8 GB/s bandwidth exceeds the FirePro M6100's 96.00 GB/s by roughly 10%. Both cards use 2 GB of GDDR5 memory on a 128-bit bus, but the NVIDIA card's faster memory clock makes the difference.

Looking at the nearest rivals for context, the FirePro M6100's average score of 13354 is nearly identical to the AMD Radeon RX 5500M's 13356 (0% delta) and slightly below the AMD Radeon Pro 555's 13407 (-0.4% delta). The GTX 950's average of 13189 sits between the NVIDIA Tesla M2090's 13075 (0.9% delta) and the AMD Radeon Pro 555X's 13321 (-1% delta). These rival positions confirm that both cards occupy the same performance tier, despite the GTX 950's strong showing in the single OpenCL test.

Specification Differences

The two GPUs differ across nearly every major specification category. Starting with memory, both have 2 GB of GDDR5 on a 128-bit bus, but the GTX 950's memory runs at 1653 MHz with 6.6 Gbps effective speed, yielding 105.8 GB/s bandwidth. The FirePro M6100's memory runs at 1500 MHz with 6 Gbps effective speed, producing 96.00 GB/s bandwidth. The NVIDIA card has a 10.2% bandwidth advantage.

Pixel and texture rates diverge significantly. The GTX 950 achieves 38.02 GPixel/s, more than double the FirePro M6100's 17.60 GPixel/s. This is a direct result of the NVIDIA card's 32 ROPs versus AMD's 16. In texture rate, the situation reverses: the FirePro M6100 hits 61.60 GTexel/s, slightly ahead of the GTX 950's 57.02 GTexel/s, thanks to its 56 TMUs versus 48.

The FP32 compute ratings are close but favor AMD. The FirePro M6100 is rated at 1.971 TFLOPS, while the GTX 950 is rated at 1.825 TFLOPS. Neither card supports FP16 in the data provided.

Architectural and interface differences are substantial. The FirePro M6100 is a 28 nm GCN 2.0 part with an MXM-B (3.0) bus interface and portable-device-dependent display outputs. The GTX 950 is a 28 nm Maxwell 2.0 part with PCIe 3.0 x16 and multiple display outputs: 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2. The FirePro M6100's slot width is listed as "MXM Module" with no power connectors, while the GTX 950 is dual-slot with a 1x 6-pin connector and a 250 W suggested PSU.

Release timing also differs. The FirePro M6100 launched on May 26, 2014, while the GTX 950 launched on August 19, 2015. Both are end-of-life products. The GTX 950 has a launch MSRP of 159 USD.

The Verdict

The data presents a split decision. In the single available head-to-head benchmark, the NVIDIA GeForce GTX 950 dominates, scoring 14.7% higher than the AMD FirePro M6100 in Geekbench OpenCL. This is a significant margin that suggests NVIDIA's Maxwell architecture extracts more real-world performance from its hardware, despite having fewer shading units and lower theoretical FP32 throughput.

However, the overall average benchmark score tells a different story. The FirePro M6100's average of 13354 edges out the GTX 950's 13189, a difference of about 1.2%. This means that across multiple workloads represented in the average, the AMD card is marginally ahead, even though it loses the specific OpenCL test.

The choice between these two depends heavily on the intended use case and platform. The FirePro M6100 is an MXM module designed for mobile workstations. It has no power connectors, no specified TDP, and its display outputs are dependent on the portable device. This is a laptop part through and through. The GTX 950 is a desktop card requiring a PCIe 3.0 x16 slot, a 6-pin power connector, and a 250 W PSU. It is physically larger at 202 mm (8 inches) and occupies two slots.

For raw compute performance in the OpenCL workload tested, the GTX 950 is the clear choice. Its 15662 score versus 13354 is not a marginal difference; it is a decisive victory. The NVIDIA card also has better API support with DirectX 12_1 and Vulkan 1.4, which may matter for forward-looking software compatibility.

For overall balanced performance as measured by the average score, the FirePro M6100 holds a slim lead. Its higher shading unit count, greater texture rate, and superior FP32 rating contribute to this, even if they do not translate to OpenCL dominance. The AMD card also has a slightly higher percentile ranking at 54 versus 53.

Where Each One Wins

The NVIDIA GeForce GTX 950 wins in compute-intensive benchmarks like Geekbench OpenCL, where it outperforms the FirePro M6100 by 14.7%. It also wins on pixel throughput, delivering 38.02 GPixel/s compared to 17.60 GPixel/s, making it the better choice for fill-rate-bound workloads. Its higher memory bandwidth of 105.8 GB/s versus 96.00 GB/s provides an edge in memory-heavy tasks. The GTX 950 also offers a more complete API feature set with DirectX 12_1 and Vulkan 1.4. Its desktop form factor with multiple display outputs makes it suitable for multi-monitor setups.

The AMD FirePro M6100 wins in overall average benchmark score, maintaining a 1.2% lead over the GTX 950. It wins on texture processing with 61.60 GTexel/s versus 57.02 GTexel/s, and on raw compute with 1.971 TFLOPS versus 1.825 TFLOPS FP32. Its higher shading unit count of 896 versus 768 gives it an advantage in shader-bound workloads. The FirePro M6100 also wins on transistor density at 13.0M per mm² versus 12.9M per mm², though this is marginal. Its MXM form factor and lack of power connectors make it the only viable option for mobile workstation platforms where the GTX 950 cannot physically be installed.

The choice ultimately comes down to platform and workload. Desktop users with access to a PCIe slot and power connector should favor the GTX 950 for its OpenCL dominance and modern API support. Mobile users constrained to an MXM module have no choice but the FirePro M6100, which offers respectable performance and a slight edge in average score. The GTX 950 is the stronger performer in the tested workload, but the FirePro M6100 holds its own in the broader performance picture.