AMD Radeon R7 350 vs NVIDIA GeForce GTX 675M Comparison

AMD Radeon R7 350 vs NVIDIA GeForce GTX 675M is a comparison between two end-of-life mobile and desktop graphics solutions separated by roughly four years of architectural evolution. The data places the AMD Radeon R7 350 in the 40th percentile of all GPUs, while the NVIDIA GeForce GTX 675M sits just one percentile lower at 39th. Their average benchmark scores are close—7425 for the AMD part versus 6946 for the NVIDIA part—but the head-to-head results reveal a clear, if modest, performance gap in favor of the newer AMD architecture.

Head-to-Head Benchmarks

The only direct benchmark comparison available is Geekbench OpenCL, where the AMD Radeon R7 350 scores 7792 against the NVIDIA GeForce GTX 675M's 6946. That represents a 12.2% advantage for the AMD card, which is a substantial margin for two GPUs that land within one percentile of each other in the overall distribution. The AMD card's win count in head-to-head tests is 1, while the NVIDIA card wins 0, so the data is unambiguous: the R7 350 is the faster part in compute workloads.

Context from the nearest rivals reinforces this positioning. The R7 350's average score of 7425 is only 0.2% behind the Intel UHD Graphics 750 (7441) and 0.3% behind the AMD Radeon HD 8850M (7447). It trails the NVIDIA GeForce GTX 1650 by 0.6% and the Intel Arc A310 by 1.7%. These are all extremely tight margins, meaning the R7 350 is essentially performance-equivalent to a modern integrated GPU and a low-end discrete card from a more recent generation. The GTX 675M, by contrast, sits 0.4% behind the AMD Radeon R5 M240 (6975), 1.1% behind the NVIDIA GeForce GTX 680M (7023), and 1.3% behind the NVIDIA T600 (7035). Its only win in the rival group is against the AMD FirePro M5100, which it leads by 1.7%. So while the GTX 675M is competitive within its own immediate peer group, the R7 350's 12.2% head-to-head lead over it is the largest delta between any two directly compared cards here.

The OpenCL result is the only measurable performance data point, but its implications are broad. A 12.2% lead in a compute API like OpenCL often translates to similar gains in other GPU-accelerated tasks, though gaming performance can vary based on driver optimizations and API support. The R7 350 also has a Vulkan score of 7057 from Geekbench, which is not directly comparable to the GTX 675M since that card has no Vulkan benchmark result in the data.

Architecture Differences

The architectural gap between these two GPUs is significant and explains much of the performance delta. The AMD Radeon R7 350 uses the Cape Verde chip built on GCN 1.0 architecture, manufactured on a 28 nm process at TSMC. It packs 1,500 million transistors into a die size of 123 mm², yielding a transistor density of 12.2 million per square millimeter. The NVIDIA GeForce GTX 675M uses the GF114 chip based on Fermi 2.0 architecture, also built by TSMC but on a larger 40 nm process. It contains 1,950 million transistors across a much larger 332 mm² die, giving a density of just 5.9 million per square millimeter. The AMD card is therefore more than twice as transistor-dense, which is a direct consequence of the smaller process node.

Core configuration differs substantially. The R7 350 has 512 shading units, 32 texture mapping units, and 16 raster operations pipelines. The GTX 675M has fewer shading units at 384, but more TMUs at 64 and more ROPs at 32. This creates an interesting split: the AMD card has 33% more shaders, while the NVIDIA card has double the texture units and double the ROPs. The pixel rate tells the story of ROP efficiency: the R7 350 achieves 12.80 GPixel/s versus the GTX 675M's 9.920 GPixel/s, meaning the AMD card's 16 ROPs are more effective per unit. Texture rate goes the other way, with the GTX 675M hitting 39.68 GTexel/s against the R7 350's 25.60 GTexel/s, reflecting its 64 TMUs.

Memory subsystems also diverge. Both cards have 2 GB of GDDR5, but the R7 350 uses a 128-bit bus with memory clocked at 1125 MHz (4.5 Gbps effective), yielding 72.00 GB/s of bandwidth. The GTX 675M uses a 256-bit bus with memory at 750 MHz (3 Gbps effective), producing 96.00 GB/s. The NVIDIA card has a third more memory bandwidth, which is notable given its lower compute scores. Floating-point performance is close: the R7 350 delivers 819.2 GFLOPS FP32, while the GTX 675M delivers 952.3 GFLOPS, a 16% advantage for the older card. This suggests the GTX 675M's architecture is less efficient at converting raw throughput into benchmark results.

API support differentiates them further. The R7 350 supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The GTX 675M supports DirectX 12 (11_0) and OpenGL 4.6, but has no Vulkan support listed. Power draw favors the AMD card at 55 W TDP versus 100 W for the NVIDIA part, and the R7 350 is a single-slot card with no power connectors, while the GTX 675M is an MXM module. The R7 350 uses PCIe 3.0 x16, and the GTX 675M uses MXM-B (3.0), reflecting its mobile origins.

The Verdict

The data clearly favors the AMD Radeon R7 350 for general compute performance. Its 12.2% lead in Geekbench OpenCL is the only head-to-head result, and it is decisive. The R7 350 also has a higher percentile ranking (40th vs 39th) and a higher average benchmark score (7425 vs 6946). For users running OpenCL workloads or Vulkan applications, the R7 350 is the rational choice, especially given its lower power draw and modern process node.

However, the GTX 675M is not without merit in specific scenarios. Its 256-bit memory bus and 96 GB/s bandwidth are superior, which could benefit memory-bound tasks. Its higher FP32 throughput (952.3 GFLOPS) and double the texture units suggest it may excel in certain fill-rate-limited or compute-heavy scenarios that are not captured by the single OpenCL benchmark. The GTX 675M also has a higher transistor count (1,950 million vs 1,500 million), though on an older process.

For gaming, the picture is less clear. The R7 350's higher pixel rate (12.80 GPixel/s) and shading unit count suggest better raw pixel pushing, but the GTX 675M's superior texture rate and memory bandwidth could matter in texture-heavy titles. Without gaming benchmarks in the data, either card could be preferable depending on the specific workload. The R7 350's Vulkan support gives it access to modern APIs that the GTX 675M lacks, which is a significant long-term advantage.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon R7 350 has an average benchmark score of 7425, compared to 6946 for the NVIDIA GeForce GTX 675M.

Q: What is the performance difference in the only direct head-to-head test?

A: In Geekbench OpenCL, the R7 350 scores 7792 against the GTX 675M's 6946, giving the AMD card a 12.2% advantage.

Q: How do their memory bandwidths compare?

A: The GTX 675M has 96.00 GB/s of bandwidth via a 256-bit bus, while the R7 350 has 72.00 GB/s via a 128-bit bus, making the NVIDIA card 33% faster in this metric.

Q: Does the NVIDIA card support Vulkan?

A: No, the GTX 675M has no Vulkan support listed. The R7 350 supports Vulkan 1.2.170.

Q: Which card has more shading units?

A: The R7 350 has 512 shading units, while the GTX 675M has 384, a 33% advantage for the AMD card.

Q: What are the power consumption figures?

A: The R7 350 has a TDP of 55 W, while the GTX 675M has a TDP of 100 W.

Where Each One Wins

The AMD Radeon R7 350 wins in the only direct benchmark comparison, taking Geekbench OpenCL with a 12.2% margin. It also wins on pixel rate (12.80 GPixel/s vs 9.920 GPixel/s), shading units (512 vs 384), and transistor density (12.2M/mm² vs 5.9M/mm²). Its 28 nm process node is two generations ahead of the NVIDIA card's 40 nm node, and it offers Vulkan support where the GTX 675M does not. The R7 350 also draws nearly half the power (55 W vs 100 W) and has a smaller die (123 mm² vs 332 mm²).

The NVIDIA GeForce GTX 675M wins on raw memory bandwidth (96.00 GB/s vs 72.00 GB/s), texture rate (39.68 GTexel/s vs 25.60 GTexel/s), and FP32 compute (952.3 GFLOPS vs 819.2 GFLOPS). It has more TMUs (64 vs 32), more ROPs (32 vs 16), and a larger memory bus (256-bit vs 128-bit). Its transistor count is higher at 1,950 million versus 1,500 million, and it uses an MXM-B interface suited for notebook implementations.

Specification Differences

The two cards differ in nearly every specification category. The R7 350 uses a 28 nm process at TSMC with a 123 mm² die, while the GTX 675M uses 40 nm with a 332 mm² die. Transistor counts are 1,500 million versus 1,950 million, and density is 12.2M/mm² versus 5.9M/mm². Memory configuration is identical in size (2 GB GDDR5) but differs in bus width (128-bit vs 256-bit), clock speed (1125 MHz vs 750 MHz), and bandwidth (72.00 GB/s vs 96.00 GB/s).

Compute resources diverge: 512 shading units, 32 TMUs, and 16 ROPs for the R7 350 versus 384 shading units, 64 TMUs, and 32 ROPs for the GTX 675M. Pixel rates are 12.80 GPixel/s versus 9.920 GPixel/s, and texture rates are 25.60 GTexel/s versus 39.68 GTexel/s. FP32 performance is 819.2 GFLOPS versus 952.3 GFLOPS. The R7 350 has a TDP of 55 W and is single-slot with no power connectors, while the GTX 675M has a TDP of 100 W and uses an MXM module form factor.

API support differs in DirectX version (12 (11_1) vs 12 (11_0)), with the R7 350 also supporting Vulkan 1.2.170 and the GTX 675M lacking Vulkan entirely. Bus interfaces are PCIe 3.0 x16 versus MXM-B (3.0). Display outputs are 1x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2 for the R7 350, while the GTX 675M's outputs are listed as portable device dependent. The R7 350 was released in 2016-07-05, while the GTX 675M came out on 2012-03-21.