AMD Radeon R7 M370 vs NVIDIA GeForce GTX 670M Comparison

# AMD Radeon R7 M370 vs NVIDIA GeForce GTX 670M

The AMD Radeon R7 M370 and NVIDIA GeForce GTX 670M represent two distinct generations of mobile graphics, separated by roughly three years of architectural evolution. The data shows a narrow but decisive victory for the AMD part in the only head-to-head benchmark available, with the R7 M370 scoring 7063 to the GTX 670M's 6513 in Geekbench OpenCL — an 8.4% advantage. Both cards sit at the 38th percentile among all GPUs, indicating they occupy similar performance tiers despite their generational gap.

FAQ

Q: Which GPU wins in the available head-to-head benchmark?

A: The AMD Radeon R7 M370 wins the only direct comparison, the Geekbench OpenCL test, with a score of 7063 versus 6513 for the NVIDIA GeForce GTX 670M. This represents an 8.4% lead for the AMD card.

Q: How do these GPUs compare to their nearest rivals?

A: The R7 M370's average benchmark score of 6764 places it 1% behind the AMD FirePro M5100 (6830) and 2.6% behind the NVIDIA GeForce GTX 675M (6946), while running 1% ahead of the NVIDIA GeForce GT 1010 (6698). The GTX 670M's average of 6513 puts it 0.3% ahead of the NVIDIA GeForce GT 555M (6493) and 0.6% behind the Intel UHD Graphics P750 (6554).

Q: What are the memory configurations of these two cards?

A: The R7 M370 comes with 2 GB of GDDR5 memory on a 128-bit bus, delivering 57.60 GB/s of bandwidth. The GTX 670M has 1536 MB of GDDR5 on a wider 192-bit bus, achieving 72.00 GB/s — about 25% more memory bandwidth.

Q: Which GPU has the higher theoretical compute throughput?

A: Despite having fewer shading units (336 versus 384), the GTX 670M produces higher FP32 performance at 803.7 GFLOPS compared to the R7 M370's 737.3 GFLOPS. The NVIDIA card also leads in pixel rate (8.372 GPixel/s vs 7.680 GPixel/s) and texture rate (33.49 GTexel/s vs 23.04 GTexel/s).

Q: What are the architectural generations and process nodes?

A: The R7 M370 uses AMD's GCN 1.0 architecture on a 28 nm TSMC process, while the GTX 670M employs NVIDIA's Fermi 2.0 architecture on a larger 40 nm TSMC node. The AMD chip is significantly smaller at 77 mm² versus 332 mm² for the NVIDIA chip.

Q: Which card has better API support?

A: The R7 M370 supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The GTX 670M also supports DirectX 12 (11_0) and OpenGL 4.6, but has no Vulkan support listed.

Architecture Differences

The architectural divide between these two mobile GPUs is substantial. The AMD Radeon R7 M370 is built on GCN 1.0, AMD's Graphics Core Next architecture, manufactured on a 28 nm process at TSMC. The chip, codenamed Litho, packs 950 million transistors into a compact 77 mm² die, achieving a transistor density of 12.3 million transistors per square millimeter. This is a notably dense design for its era.

The NVIDIA GeForce GTX 670M, by contrast, uses the Fermi 2.0 architecture with the GF114 chip, fabricated on TSMC's older 40 nm process. The die is much larger at 332 mm² and contains 1,950 million transistors — more than double the AMD chip's count — but the transistor density drops to just 5.9 million per square millimeter due to the older process technology.

The R7 M370 belongs to AMD's "Gem System (R7 M300)" generation, with its predecessor listed as "Solar System" and successor as "Polaris Mobile." The GTX 670M is part of the GeForce 600M family, succeeding the GeForce 500M and followed by the GeForce 700M. The AMD card was released in May 2015, while the NVIDIA card came out in March 2012.

In terms of compute resources, the R7 M370 fields 384 shading units, 24 texture mapping units, and 8 ROPs. The GTX 670M has 336 shading units but compensates with 56 TMUs and 24 ROPs — substantially more texture and pixel processing capability. This disparity explains why the NVIDIA card achieves higher texture and pixel fill rates despite fewer shaders.

The memory subsystems also reflect different design philosophies. The R7 M370 uses a 128-bit bus with 2 GB of GDDR5 at 3.6 Gbps effective, yielding 57.60 GB/s. The GTX 670M uses a 192-bit bus with 1536 MB of GDDR5 at 3 Gbps effective, producing 72.00 GB/s. The NVIDIA card's wider bus gives it a bandwidth advantage, though the AMD card has more total memory capacity.

Where Each One Wins

The benchmark data shows the R7 M370 winning the only direct comparison, but the specification sheet reveals a more nuanced picture. The AMD card excels in scenarios that favor its newer architecture and higher shader count. Its 8.4% lead in Geekbench OpenCL suggests better compute performance in general-purpose workloads, which aligns with its GCN 1.0 design that was built with compute in mind.

The GTX 670M, despite losing the OpenCL test, has clear strengths in traditional graphics throughput. Its pixel rate of 8.372 GPixel/s exceeds the R7 M370's 7.680 GPixel/s by about 9%, and its texture rate of 33.49 GTexel/s is a substantial 45% higher than the AMD's 23.04 GTexel/s. This suggests the NVIDIA card would likely perform better in fill-rate-bound gaming scenarios at higher resolutions or with heavy texture filtering.

The R7 M370's 2 GB memory capacity is advantageous for modern games that exceed 1.5 GB of VRAM usage, while the GTX 670M's higher bandwidth might benefit texture streaming. The AMD card also has Vulkan support, which could provide an edge in Vulkan-based titles, while the NVIDIA card lacks this API entirely. The GTX 670M's 75 W TDP and MXM Module form factor indicate it was designed for larger laptops, whereas the R7 M370's unspecified power draw suggests a potentially more power-efficient design given its smaller die and newer process node.

Specification Differences

The two GPUs differ across nearly every major specification category:

• Process Node: The R7 M370 uses 28 nm, while the GTX 670M uses 40 nm.
• Transistors: 950 million (R7 M370) versus 1,950 million (GTX 670M).
• Die Size: 77 mm² (R7 M370) versus 332 mm² (GTX 670M).
• Transistor Density: 12.3M / mm² versus 5.9M / mm².
• Clock Speeds: The R7 M370 has a base clock of 875 MHz and boost clock of 960 MHz; the GTX 670M has no base or boost clocks listed.
• Memory Speed: 3.6 Gbps effective (R7 M370) versus 3 Gbps effective (GTX 670M).
• Memory Size: 2 GB versus 1536 MB.
• Memory Bus Width: 128 bit versus 192 bit.
• Memory Bandwidth: 57.60 GB/s versus 72.00 GB/s.
• Shading Units: 384 versus 336.
• TMUs: 24 versus 56.
• ROPs: 8 versus 24.
• Pixel Rate: 7.680 GPixel/s versus 8.372 GPixel/s.
• Texture Rate: 23.04 GTexel/s versus 33.49 GTexel/s.
• FP32 Performance: 737.3 GFLOPS versus 803.7 GFLOPS.
• TDP: The GTX 670M is rated at 75 W; the R7 M370 has no TDP listed.
• Slot Width: The GTX 670M is an MXM Module; the R7 M370 has no slot width listed.
• Power Connectors: The GTX 670M has none; the R7 M370 has none listed.
• Bus Interface: PCIe 3.0 x8 (R7 M370) versus MXM-B (3.0) (GTX 670M).
• Display Outputs: Portable Device Dependent (GTX 670M); none listed for R7 M370.
• Vulkan Support: R7 M370 supports 1.2.170; GTX 670M has no Vulkan support.
• DirectX Support: R7 M370 supports 12 (11_1); GTX 670M supports 12 (11_0).
• Release Date: May 2015 (R7 M370) versus March 2012 (GTX 670M).

Head-to-Head Benchmarks

The only direct benchmark comparison available is the Geekbench OpenCL test, where the AMD Radeon R7 M370 scores 7063 against the NVIDIA GeForce GTX 670M's 6513. The 8.4% delta in favor of the AMD card is meaningful but not overwhelming. This single data point suggests the R7 M370 has a modest compute advantage, likely stemming from its newer GCN architecture and higher shader count.

However, the specification differences paint a more complex picture. The GTX 670M's FP32 output of 803.7 GFLOPS is 9% higher than the R7 M370's 737.3 GFLOPS, yet the AMD card still wins the OpenCL benchmark. This indicates that raw theoretical throughput doesn't always translate directly to real-world compute performance — the architectural efficiency of GCN 1.0 appears to extract more useful work per FLOP than Fermi 2.0.

The GTX 670M's dominance in texture rate is striking: 33.49 GTexel/s versus 23.04 GTexel/s is a 45% advantage. With 56 TMUs versus 24, the NVIDIA card is clearly designed for texture-heavy workloads. Similarly, its 24 ROPs triple the AMD card's 8, giving it a theoretical pixel throughput advantage of 8.372 GPixel/s versus 7.680 GPixel/s.

In terms of average benchmark scores, the R7 M370's 6764 average sits 3.9% above the GTX 670M's 6513. The nearest rivals for each card reinforce this positioning: the R7 M370's closest competitor is the NVIDIA GeForce GTX 675M at 6946 (2.6% higher), while the GTX 670M's nearest rival is the NVIDIA GeForce GT 555M at 6493 (0.3% lower). This suggests the R7 M370 competes in a slightly higher performance tier than the GTX 670M.

The Verdict

The data points to the AMD Radeon R7 M370 as the better performer in compute-oriented workloads, winning the only head-to-head benchmark by 8.4% and holding a higher average benchmark score (6764 versus 6513). Its newer 28 nm process, smaller die, and superior transistor density indicate a more modern design that achieves competitive performance with significantly less silicon area.

The NVIDIA GeForce GTX 670M, however, retains advantages in traditional graphics metrics. Its higher pixel rate, texture rate, and FP32 throughput suggest it could outperform the R7 M370 in fill-rate-limited gaming scenarios, despite losing the compute benchmark. The wider memory bus and higher bandwidth (72.00 GB/s versus 57.60 GB/s) also favor the NVIDIA card for memory-intensive applications.

Users prioritizing modern API support, particularly Vulkan, should lean toward the R7 M370. Those who need higher raw graphics throughput or are constrained by a 75 W TDP might prefer the GTX 670M. The R7 M370's 2 GB memory capacity is more future-proof than the GTX 670M's 1536 MB, but the NVIDIA card's bandwidth advantage may matter more in specific workloads.

Ultimately, the R7 M370 is the pragmatic choice for most scenarios given its benchmark victory and newer architecture. The GTX 670M remains a capable option for users who understand its fill-rate strengths and can work within its older API limitations. Both cards are end-of-life products, so availability and driver support should factor into any decision.