AMD Radeon R7 350 vs NVIDIA GeForce GTX 950M Comparison

# The Verdict

The data presents a clear split decision between these two mobile and desktop graphics solutions. The NVIDIA GeForce GTX 950M wins the OpenCL compute benchmark with a score of 9,745 against the AMD Radeon R7 350's 7,792, a 25.1% advantage. However, the AMD Radeon R7 350 wins the Vulkan benchmark with 7,057 versus 6,525, a 7.5% lead. With each card claiming one benchmark victory, the choice hinges entirely on workload. The GTX 950M is the pick for general compute tasks and OpenCL-accelerated applications, while the R7 350 is the better option for Vulkan-based game engines and APIs. The GTX 950M's average benchmark score of 8,135 places it in the 42nd percentile of all GPUs, whereas the R7 350's average of 7,425 sits in the 40th percentile. The 9.6% gap in average scores favors the NVIDIA part, but the R7 350's superior Vulkan showing complicates any blanket recommendation. For users prioritizing raw compute throughput, the GTX 950M is the data-backed choice. For those targeting Vulkan-specific workloads, the R7 350's 7.5% edge is decisive.

# Architecture Differences

The two GPUs represent fundamentally different architectural philosophies from their respective manufacturers. The NVIDIA GeForce GTX 950M uses the GM107 chip built on the Maxwell architecture, part of the GeForce 900M generation. It is fabricated by TSMC on a 28 nm process node, containing 1,870 million transistors across a 148 mm² die, yielding a transistor density of 12.6 million per square millimeter. The AMD Radeon R7 350 employs the Cape Verde chip based on GCN 1.0 architecture, belonging to the Pirate Islands (R7 300) generation. It also uses TSMC's 28 nm process but packs 1,500 million transistors into a 123 mm² die, achieving a slightly lower transistor density of 12.2 million per square millimeter.

The GTX 950M features 640 shading units, 40 texture mapping units, and 16 raster output pipelines. In contrast, the R7 350 has 512 shading units, 32 TMUs, and the same 16 ROPs. This gives the NVIDIA chip a 20% advantage in both shading unit count and TMU count. The GTX 950M's pixel rate is 17.98 GPixel/s with a texture rate of 44.96 GTexel/s, while the R7 350 achieves 12.80 GPixel/s and 25.60 GTexel/s respectively. The FP32 compute throughput tells a similar story: the GTX 950M delivers 1,438.7 GFLOPS versus the R7 350's 819.2 GFLOPS.

API support differs meaningfully. The GTX 950M supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4. The R7 350 supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. While the NVIDIA part has a higher Vulkan version number, the benchmark results indicate the AMD hardware executes Vulkan workloads more efficiently in practice.

# Head-to-Head Benchmarks

The two available benchmark results tell a story of divergent strengths. In the Geekbench OpenCL test, the NVIDIA GeForce GTX 950M scores 9,745 against the AMD Radeon R7 350's 7,792. This represents a 25.1% delta in favor of the NVIDIA card. The GTX 950M's advantage in this test aligns with its substantial hardware lead in shading units, texture units, and raw FP32 compute. The 640 shading units and 1,438.7 GFLOPS of FP32 throughput provide a strong foundation for OpenCL workloads that scale with parallel compute resources.

The Vulkan benchmark reverses the outcome, with the R7 350 scoring 7,057 against the GTX 950M's 6,525. The AMD card leads by 7.5% in this test, despite having fewer shading units and lower theoretical compute throughput. This suggests the GCN 1.0 architecture's Vulkan driver implementation or hardware scheduling handles the workload more efficiently than Maxwell's. The R7 350's GDDR5 memory with 72.00 GB/s bandwidth versus the GTX 950M's DDR3 at 28.80 GB/s likely contributes to this Vulkan advantage, as memory bandwidth can be a limiting factor in graphics API workloads.

The average benchmark scores reflect the overall picture: the GTX 950M averages 8,135 across both tests, while the R7 350 averages 7,425. This 9.6% aggregate gap favors NVIDIA, but the R7 350's Vulkan win demonstrates that average scores can obscure meaningful API-specific differences.

# Specification Differences

The two cards differ across nearly every specification category. The GTX 950M carries 4 GB of DDR3 memory on a 128-bit bus with 28.80 GB/s bandwidth, while the R7 350 has 2 GB of GDDR5 memory on the same 128-bit bus but with 72.00 GB/s bandwidth — 2.5 times the memory bandwidth of the NVIDIA part. The GTX 950M's memory clock is 900 MHz (1,800 Mbps effective), whereas the R7 350's memory runs at 1,125 MHz (4.5 Gbps effective).

The GTX 950M has 640 shading units versus 512, 40 TMUs versus 32, and equal 16 ROPs. Its base clock is 993 MHz with a boost clock of 1,124 MHz; the R7 350's clocks are not specified in the data. Pixel rate favors NVIDIA at 17.98 GPixel/s versus 12.80 GPixel/s, as does texture rate at 44.96 GTexel/s versus 25.60 GTexel/s. FP32 compute is 1,438.7 GFLOPS versus 819.2 GFLOPS.

Power and physical specifications differ notably. The GTX 950M has a 75 W TDP and is an IGP form factor with no power connectors, while the R7 350 has a 55 W TDP, is single-slot, has no power connectors, and suggests a 250 W PSU. The bus interface differs: PCIe 3.0 x8 for the GTX 950M versus PCIe 3.0 x16 for the R7 350. Display outputs are "portable device dependent" for the NVIDIA part, while the AMD card offers 1x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2. The R7 350 has a physical length of 168 mm (6.6 inches). Transistor count and die size differ: 1,870M transistors on 148 mm² for NVIDIA versus 1,500M on 123 mm² for AMD.

# FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 950M has an average benchmark score of 8,135, compared to the AMD Radeon R7 350's 7,425, giving NVIDIA a 9.6% advantage.

Q: How do the two cards compare in Vulkan performance?

A: The AMD Radeon R7 350 wins the Geekbench Vulkan test with a score of 7,057 against the GTX 950M's 6,525, a 7.5% lead for AMD.

Q: Which GPU has more shading units?

A: The NVIDIA GeForce GTX 950M has 640 shading units, while the AMD Radeon R7 350 has 512 shading units, a 20% difference in NVIDIA's favor.

Q: What type of memory does each card use?

A: The GTX 950M uses 4 GB of DDR3 memory with 28.80 GB/s bandwidth, while the R7 350 uses 2 GB of GDDR5 memory with 72.00 GB/s bandwidth.

Q: Which GPU has the lower power draw?

A: The AMD Radeon R7 350 has a 55 W TDP, which is lower than the GTX 950M's 75 W TDP.

Q: What is the transistor density of each chip?

A: The GTX 950M's GM107 chip has a transistor density of 12.6 million per square millimeter, while the R7 350's Cape Verde chip has 12.2 million per square millimeter.

# Where Each One Wins

The NVIDIA GeForce GTX 950M wins decisively in OpenCL compute workloads, posting a 25.1% higher score than the R7 350. This makes it the preferred option for general-purpose GPU computing, OpenCL-accelerated applications, and any workload that leverages raw FP32 throughput. Its 1,438.7 GFLOPS of FP32 compute, combined with 640 shading units and 40 TMUs, provides substantial parallel processing headroom. The GTX 950M's 42nd percentile ranking and 8,135 average score place it ahead of the R7 350's 40th percentile and 7,425 average. Its higher pixel rate of 17.98 GPixel/s and texture rate of 44.96 GTexel/s also indicate stronger traditional rasterization throughput.

The AMD Radeon R7 350 wins the Vulkan benchmark by 7.5%, making it the data-backed choice for Vulkan-based game engines and applications. Its GDDR5 memory with 72.00 GB/s bandwidth — 2.5 times the GTX 950M's bandwidth — likely plays a significant role in this victory, as Vulkan workloads often stress memory throughput. The R7 350's lower 55 W TDP also makes it a more power-efficient option for sustained Vulkan workloads. Its PCIe 3.0 x16 interface provides twice the bus bandwidth of the GTX 950M's PCIe 3.0 x8 connection, which can benefit data transfer-heavy scenarios. The R7 350's 7,057 Vulkan score demonstrates that architectural efficiency can overcome raw compute disadvantages.

For users who need both OpenCL and Vulkan performance, the choice requires prioritizing one API over the other. The GTX 950M offers superior overall compute performance and a higher average score, while the R7 350 provides a more balanced power profile and a significant memory bandwidth advantage. The data shows the GTX 950M as the stronger all-around performer in aggregate benchmarks, but the R7 350's Vulkan win prevents a unanimous recommendation. The decision ultimately rests on which API the user's target applications rely on most heavily.