AMD Radeon R7 M370 vs NVIDIA GeForce GTX 670MX Comparison

Where Each One Wins

The recorded data splits cleanly between these two mobile GPUs. The AMD Radeon R7 M370 wins both head-to-head benchmark tests, making it the stronger performer in compute-oriented workloads. In geekbench_opencl, the AMD part scores 7063 against the NVIDIA GeForce GTX 670MX's 6125, a 15.3% advantage. The gap widens in geekbench_vulkan, where the AMD card reaches 6465 versus 5316 for NVIDIA, a 21.6% lead.

The NVIDIA GeForce GTX 670MX does not win either recorded benchmark. However, the database places it in a different competitive neighborhood. Its nearest rivals include the NVIDIA GeForce GTX 550 Ti at 5731, the Intel Iris Pro Graphics P6300 at 5712, and the AMD Radeon HD 8790M at 5691. The AMD Radeon R7 M370 sits alongside the NVIDIA GeForce GTX 675M at 6946, the AMD FirePro M5100 at 6830, and the NVIDIA GeForce GT 1010 at 6698. This means the AMD card competes in a higher performance tier overall, while the NVIDIA card sits closer to entry-level desktop and older mobile parts.

The percentile ranking reinforces this split. The AMD Radeon R7 M370 lands at the 38th percentile among all GPUs in the database, while the NVIDIA GeForce GTX 670MX sits at the 33rd. The average benchmark score tells the same story: 6764 for AMD versus 5721 for NVIDIA. For users focused on OpenCL and Vulkan compute workloads, the AMD part is the clear pick. For users who only need modest graphics acceleration in an older notebook, the NVIDIA part still has a place, but the data shows it is measurably slower in both recorded tests.

Architecture Differences

The two GPUs come from different architectural generations and different design philosophies. The AMD Radeon R7 M370 uses the GCN 1.0 architecture on a chip codenamed Litho. It is built on a 28 nm process at TSMC with 950 million transistors on a 77 mm² die. The transistor density works out to 12.3 million transistors per square millimeter. This is a small, efficient chip designed for thin-and-light laptops.

The NVIDIA GeForce GTX 670MX uses the Kepler architecture on a GK104 chip, also built on a 28 nm process at TSMC. The difference is scale: NVIDIA packs 3,540 million transistors onto a 294 mm² die, giving a density of 12.0 million transistors per square millimeter. The GK104 is a much larger chip, roughly four times the transistor count and nearly four times the die area of the AMD part. This reflects its origin as a desktop-derived GPU scaled down for mobile use.

The compute resources differ dramatically. The AMD Radeon R7 M370 has 384 shading units, 24 texture mapping units, and 8 raster operation pipelines. The NVIDIA GeForce GTX 670MX has 960 shading units, 80 texture mapping units, and 24 ROPs. The NVIDIA chip has 2.5 times the shader count, over 3 times the TMUs, and 3 times the ROPs. This explains why the NVIDIA part still posts competitive fill rates despite its lower clock speeds: its pixel rate is 12.02 GPixel/s versus 7.680 GPixel/s for AMD, and its texture rate is 48.08 GTexel/s versus 23.04 GTexel/s.

Clock speeds flip the picture. The AMD Radeon R7 M370 runs at 875 MHz base and boosts to 960 MHz, while the NVIDIA GeForce GTX 670MX is locked at 601 MHz with no boost. The AMD card's higher clocks partially compensate for its smaller compute core. Still, in raw FP32 throughput, the NVIDIA part leads with 1,153.9 GFLOPS against 737.3 GFLOPS for AMD.

Memory configurations also differ. The AMD card uses 2 GB of GDDR5 on a 128-bit bus with a memory clock of 900 MHz, or 3.6 Gbps effective, yielding 57.60 GB/s of bandwidth. The NVIDIA card uses 3 GB of GDDR5 on a 192-bit bus with a memory clock of 700 MHz, or 2.8 Gbps effective, yielding 67.20 GB/s. The NVIDIA part has more capacity and more bandwidth despite slower memory clocks, thanks to the wider bus.

API support shows a split. The AMD card supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The NVIDIA card supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. Both support the same OpenGL version, but NVIDIA's Vulkan version is slightly newer. The AMD card's DX12 feature level is one notch higher. Power draw is only listed for the NVIDIA part at 75 W, while the AMD card has no TDP figure in the database. The NVIDIA card also lists no power connectors, indicating it draws entirely from the motherboard slot.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon R7 M370 has an average benchmark score of 6764, while the NVIDIA GeForce GTX 670MX scores 5721. The AMD card sits at the 38th percentile among all GPUs, the NVIDIA card at the 33rd.

Q: How large is the performance gap in Vulkan compute?

A: In geekbench_vulkan, the AMD Radeon R7 M370 scores 6465 against 5316 for the NVIDIA GeForce GTX 670MX, a 21.6% advantage for AMD.

Q: Does the NVIDIA card have any architectural advantage?

A: Yes, the NVIDIA GeForce GTX 670MX has significantly more compute resources. It packs 960 shading units, 80 TMUs, and 24 ROPs, versus 384 shading units, 24 TMUs, and 8 ROPs for the AMD card. It also leads in pixel rate at 12.02 GPixel/s and texture rate at 48.08 GTexel/s.

Q: Which GPU has more memory bandwidth?

A: The NVIDIA GeForce GTX 670MX has 67.20 GB/s of bandwidth from its 192-bit bus and 3 GB of GDDR5. The AMD Radeon R7 M370 has 57.60 GB/s from a 128-bit bus and 2 GB of GDDR5.

Q: Are both GPUs still in production?

A: No, both are listed as end-of-life. The AMD Radeon R7 M370 was released in 2015 and succeeded by Polaris Mobile. The NVIDIA GeForce GTX 670MX was released in 2012 and succeeded by GeForce 700M.

Q: How do the closest rivals compare for each card?

A: For the AMD Radeon R7 M370, the closest rival is the AMD FirePro M5100 at 6830, which is 1% slower. For the NVIDIA GeForce GTX 670MX, the closest rival is the Intel Iris Pro Graphics P6300 at 5712, which is 0.2% slower.

Specification Differences

The two GPUs differ across nearly every major specification category. The AMD Radeon R7 M370 uses the GCN 1.0 architecture with a Litho chip, while the NVIDIA GeForce GTX 670MX uses Kepler with a GK104 chip. The AMD chip contains 950 million transistors on a 77 mm² die, while the NVIDIA chip contains 3,540 million transistors on a 294 mm² die. Transistor density is nearly identical: 12.3M per mm² for AMD versus 12.0M per mm² for NVIDIA.

Clock speeds differ substantially. The AMD card runs at 875 MHz base and 960 MHz boost. The NVIDIA card runs at 601 MHz with no boost capability. Memory clocks also differ, with AMD at 900 MHz (3.6 Gbps effective) and NVIDIA at 700 MHz (2.8 Gbps effective).

Memory capacity and bus width favor NVIDIA: 3 GB on a 192-bit bus versus 2 GB on a 128-bit bus. Bandwidth follows the bus width, with NVIDIA at 67.20 GB/s and AMD at 57.60 GB/s.

Compute resources heavily favor NVIDIA. The NVIDIA card has 960 shading units, 80 TMUs, and 24 ROPs. The AMD card has 384 shading units, 24 TMUs, and 8 ROPs. Pixel rate is 12.02 GPixel/s for NVIDIA versus 7.680 GPixel/s for AMD. Texture rate is 48.08 GTexel/s versus 23.04 GTexel/s. FP32 throughput is 1,153.9 GFLOPS versus 737.3 GFLOPS.

The bus interface differs: AMD uses PCIe 3.0 x8, NVIDIA uses PCIe 3.0 x16. Display outputs are listed as portable device dependent for NVIDIA, while AMD has no listing. Power draw is listed only for NVIDIA at 75 W, with no power connectors required. The AMD card has no TDP figure recorded.

API support shows minor differences. AMD supports DirectX 12 (11_1) and Vulkan 1.2.170. NVIDIA supports DirectX 12 (11_0) and Vulkan 1.2.175. Both support OpenGL 4.6.

Release timing differs by roughly three years. The AMD Radeon R7 M370 launched in 2015 as part of the Gem System R7 M300 generation. The NVIDIA GeForce GTX 670MX launched in 2012 as part of the GeForce 600M generation.

Head-to-Head Benchmarks

The database records two head-to-head benchmark tests between these GPUs, and the AMD Radeon R7 M370 wins both. In geekbench_opencl, the AMD card scores 7063 against 6125 for the NVIDIA GeForce GTX 670MX. That is a 15.3% delta in favor of AMD. This is a meaningful gap, roughly the same margin that separates the AMD card from its nearest rival, the AMD FirePro M5100, which sits only 1% behind at 6830.

The Vulkan test shows an even larger divide. The AMD Radeon R7 M370 scores 6465 in geekbench_vulkan, while the NVIDIA GeForce GTX 670MX manages 5316. The 21.6% delta is the biggest margin in either direction across the head-to-head set. In the context of the NVIDIA card's nearest rivals, this score of 5316 would place it below every one of its closest competitors: the Intel Iris Pro Graphics P6300 at 5712, the NVIDIA GeForce GTX 550 Ti at 5731, the AMD Radeon HD 8790M at 5691, and the NVIDIA Quadro M500M at 5604.

The AMD card's OpenCL score of 7063 also holds up well against its own competitive set. It beats the NVIDIA GeForce GTX 675M at 6946, the AMD FirePro M5100 at 6830, and the NVIDIA GeForce GT 1010 at 6698. Only the AMD Radeon R7 M460 at 6612 sits further behind, at 2.3% slower.

These benchmark results are somewhat surprising given the raw specification sheet. The NVIDIA card has more than double the shader count, three times the ROPs, and 56% higher FP32 throughput. Yet in both recorded tests, the AMD card wins by double-digit margins. The data suggests that the AMD GCN architecture maps more efficiently onto OpenCL and Vulkan workloads, or that the NVIDIA card's low 601 MHz clock and lack of boost severely limit its real-world compute performance. The AMD card's 960 MHz boost clock gives it a 59% clock advantage, which appears to outweigh the NVIDIA card's wider compute core in these specific tests.

The Verdict

The recorded data supports a clear split. The AMD Radeon R7 M370 is the better choice for compute-oriented tasks. It wins both head-to-head benchmarks, holds a higher average score of 6764 against 5721, and ranks in the 38th percentile versus the 33rd for NVIDIA. Its 15.3% OpenCL lead and 21.6% Vulkan lead are substantial enough that users running OpenCL or Vulkan workloads should prefer the AMD card without hesitation.

The NVIDIA GeForce GTX 670MX still has arguments in its favor, but they come from the specification sheet rather than the benchmark data. It offers 3 GB of memory versus 2 GB, a wider 192-bit bus, higher memory bandwidth at 67.20 GB/s, and significantly more shading units, TMUs, and ROPs. Its 1,153.9 GFLOPS of FP32 throughput and 12.02 GPixel/s pixel rate are class-leading figures for its era. For applications that rely on raw fill rate or memory capacity rather than compute API performance, the NVIDIA card may hold an edge that the recorded benchmarks do not capture.

Users who prioritize compute benchmarks should choose the AMD Radeon R7 M370. Its higher clock speeds, more modern GCN architecture, and superior OpenCL and Vulkan scores make it the stronger all-around performer in the database's measurements. Users who need more video memory or higher fill rates in a legacy mobile workstation may still find the NVIDIA GeForce GTX 670MX serviceable, but the data shows it is the slower GPU in every recorded head-to-head test.