AMD Radeon R7 M350 vs NVIDIA GeForce GTX 675M Comparison
AMD Radeon R7 M350
GeForce GTX 675M
PERFORMANCE BENCHMARKS
Analysis: AMD Radeon R7 M350 vs NVIDIA GeForce GTX 675M
The NVIDIA GeForce GTX 675M and AMD Radeon R7 M350 are two end-of-life mobile graphics solutions from different eras, and the benchmark data shows a surprisingly close contest in raw compute. The GTX 675M, a 2012 Fermi part, and the R7 M350, a 2015 GCN 3.0 part, land within a fraction of a percent of each other in OpenCL performance. The data suggests that for legacy applications relying on OpenCL, the choice is nearly a coin flip, while the AMD card offers more modern API support and double the memory capacity.
The Verdict
The benchmark results indicate a near-tie in overall performance, but the decision hinges on what you prioritize. If your workload is strictly bound to OpenCL compute, the AMD Radeon R7 M350 holds a razor-thin edge, scoring 6991 versus the GTX 675M’s 6946 — a 0.6% difference that is effectively negligible in real-world terms. This makes the R7 M350 the safer pick for general compute tasks where the slightly higher score and the presence of Vulkan support (a score of 5662 in geekbench_vulkan) provide more flexibility for modern applications.
However, the NVIDIA GeForce GTX 675M is not obsolete in this comparison. Its architecture delivers substantially higher pixel and texture throughput, with a pixel rate of 9.920 GPixel/s and a texture rate of 39.68 GTexel/s, compared to the R7 M350’s 8.120 GPixel/s and 24.36 GTexel/s. For older games or applications that are fill-rate limited, the GTX 675M will likely perform better despite its lower raw compute score. Choose the GTX 675M if you care about traditional rasterization performance and are working with a system that uses an MXM module. Choose the R7 M350 if you need more memory (4 GB vs 2 GB), modern API support like Vulkan, and a slightly higher compute score.
Architecture Differences
The two GPUs come from fundamentally different design philosophies. The NVIDIA GeForce GTX 675M is built on the Fermi 2.0 architecture using a 40 nm process at TSMC. It packs 1,950 million transistors into a die size of 332 mm², which results in a transistor density of 5.9M per mm². The chip is designated GF114. In contrast, the AMD Radeon R7 M350 uses the GCN 3.0 architecture on a more advanced 28 nm process, also from TSMC. It contains 1,550 million transistors on a much smaller die of 125 mm², yielding a higher transistor density of 12.4M per mm².
The core configurations differ significantly. Both GPUs have 384 shading units, but the GTX 675M has 64 texture mapping units (TMUs) and 32 render output units (ROPs), while the R7 M350 has only 24 TMUs and 8 ROPs. This explains the large gap in texture and pixel rates. The memory subsystems are also distinct: the GTX 675M uses 2 GB of GDDR5 on a 256-bit bus for a bandwidth of 96.00 GB/s, whereas the R7 M350 uses 4 GB of DDR3 on a 64-bit bus, resulting in just 16.00 GB/s of bandwidth. The R7 M350 has a base clock of 1000 MHz with a boost of 1015 MHz; the GTX 675M lists no base or boost clocks but runs its memory at 750 MHz (3 Gbps effective). The GTX 675M’s FP32 compute is 952.3 GFLOPS, ahead of the R7 M350’s 779.5 GFLOPS. The R7 M350 also supports FP16 at a 1:1 ratio (779.5 GFLOPS), which the GTX 675M does not list.
Where Each One Wins
The data shows a clear split between compute-centric and fill-rate-centric tasks. The AMD Radeon R7 M350 wins in the only head-to-head benchmark recorded, the geekbench_opencl test, with a score of 6991 against 6946. This 0.6% advantage, while small, suggests that the GCN 3.0 architecture is more efficient at general compute workloads. The R7 M350 also has a significant advantage in memory capacity, offering 4 GB versus the GTX 675M’s 2 GB, which can be crucial for workloads that need larger datasets in memory. Its support for Vulkan 1.2.170 and DirectX 12 (12_0) also gives it a software feature edge.
The NVIDIA GeForce GTX 675M wins in the raw throughput metrics. Its pixel rate of 9.920 GPixel/s is 22% higher than the R7 M350’s 8.120 GPixel/s, and its texture rate of 39.68 GTexel/s is 63% higher than the R7 M350’s 24.36 GTexel/s. These figures indicate that for rasterization-heavy tasks — such as traditional game rendering at lower resolutions — the GTX 675M should deliver smoother performance. The GTX 675M also has a higher FP32 rating (952.3 GFLOPS vs 779.5 GFLOPS) and a much larger memory bus (256-bit vs 64-bit), which helps in bandwidth-sensitive scenarios despite its overall bandwidth being higher at 96.00 GB/s versus 16.00 GB/s.
FAQ
Q: Which GPU has a higher average benchmark score?
A: The NVIDIA GeForce GTX 675M has a higher average benchmark score of 6946, while the AMD Radeon R7 M350 has an average score of 6327. However, the R7 M350’s average is pulled down by its Vulkan score of 5662; its OpenCL score of 6991 is actually higher than the GTX 675M’s single OpenCL score of 6946.
Q: Do these GPUs support modern APIs?
A: The AMD Radeon R7 M350 supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.2.170. The NVIDIA GeForce GTX 675M supports DirectX 12 (11_0) and OpenGL 4.6, but has no listed Vulkan support.
Q: How do they compare in memory bandwidth?
A: The NVIDIA GeForce GTX 675M has a memory bandwidth of 96.00 GB/s, which is vastly higher than the AMD Radeon R7 M350’s 16.00 GB/s. This is due to the GTX 675M’s 256-bit bus and GDDR5 memory, compared to the R7 M350’s 64-bit bus and DDR3 memory.
Q: Which GPU has more shading units?
A: Both GPUs have exactly 384 shading units. The difference lies in other parts of the pipeline: the GTX 675M has 64 TMUs and 32 ROPs, while the R7 M350 has 24 TMUs and 8 ROPs.
Q: What are the process nodes for each GPU?
A: The NVIDIA GeForce GTX 675M is fabricated on a 40 nm process, while the AMD Radeon R7 M350 uses a smaller 28 nm process. Both are manufactured by TSMC.
Q: Which GPU has a higher pixel rate?
A: The NVIDIA GeForce GTX 675M has a pixel rate of 9.920 GPixel/s, which is higher than the AMD Radeon R7 M350’s 8.120 GPixel/s.
Head-to-Head Benchmarks
The only direct benchmark comparison in the data is the geekbench_opencl test. In this test, the AMD Radeon R7 M350 wins with a score of 6991 against the NVIDIA GeForce GTX 675M’s 6946. The delta percentage is -0.6% from the perspective of the GTX 675M, meaning the R7 M350 is 0.6% faster. This is a marginal win, but it is the decisive one in the head-to-head record, giving the R7 M350 a 1-0 win tally.
Beyond that single test, the data shows significant wins for the GTX 675M in theoretical throughput. The GTX 675M’s texture rate of 39.68 GTexel/s is a clear victory over the R7 M350’s 24.36 GTexel/s. Similarly, the pixel rate of 9.920 GPixel/s for the GTX 675M beats the R7 M350’s 8.120 GPixel/s. In terms of memory bandwidth, the GTX 675M’s 96.00 GB/s dwarfs the R7 M350’s 16.00 GB/s. The GTX 675M also has a higher FP32 compute rating at 952.3 GFLOPS versus 779.5 GFLOPS. The R7 M350 counters with a higher transistor density (12.4M per mm² vs 5.9M per mm²) and a more modern feature set, but in the field of raw speed, the GTX 675M holds the lead in almost every measured metric except the OpenCL benchmark score.
Specification Differences
The two GPUs differ across nearly every major specification. The process node is 40 nm for the GTX 675M versus 28 nm for the R7 M350. The transistor count is 1,950 million for NVIDIA versus 1,550 million for AMD. The die size is 332 mm² for NVIDIA versus 125 mm² for AMD, leading to a transistor density of 5.9M per mm² versus 12.4M per mm². The memory configuration is starkly different: the GTX 675M has 2 GB of GDDR5 on a 256-bit bus, while the R7 M350 has 4 GB of DDR3 on a 64-bit bus. This results in memory bandwidth of 96.00 GB/s versus 16.00 GB/s. The TMU count is 64 for NVIDIA versus 24 for AMD, and the ROP count is 32 versus 8. The pixel rate is 9.920 GPixel/s versus 8.120 GPixel/s, and the texture rate is 39.68 GTexel/s versus 24.36 GTexel/s. The FP32 performance is 952.3 GFLOPS versus 779.5 GFLOPS; the R7 M350 also lists FP16 at 779.5 GFLOPS, while the GTX 675M has no FP16 figure. The base and boost clocks are listed only for the R7 M350 (1000 MHz and 1015 MHz), while the GTX 675M lists only a memory clock of 750 MHz. The TDP is 100 W for the GTX 675M, with no TDP listed for the R7 M350. The bus interface is MXM-B (3.0) for NVIDIA versus PCIe 3.0 x8 for AMD. The API support differs: DirectX 12 (11_0) for NVIDIA versus DirectX 12 (12_0) for AMD, and Vulkan is only listed for AMD (1.2.170). Finally, the release dates are 2012-03-21 for the GTX 675M and 2015-05-04 for the R7 M350, with the GTX 675M belonging to the GeForce 600M generation and the R7 M350 to the Gem System (R7 M300) generation.