AMD Radeon R9 370X vs NVIDIA GeForce GTX 965M Comparison

Both the AMD Radeon R9 370X and the NVIDIA GeForce GTX 965M are end-of-life products from the 28 nm era, but they were built for fundamentally different environments. The R9 370X is a dual-slot desktop card with a 180 W TDP and two 6-pin power connectors, while the GTX 965M is an MXM module with no power connectors, designed for portable devices. The benchmark data shows a clear split: AMD dominates in OpenCL compute, while NVIDIA wins decisively in Vulkan graphics workloads. Their average benchmark scores sit close together—15,862 for the AMD part versus 14,404 for the NVIDIA part—placing them at the 58th and 56th percentiles of all GPUs respectively. This makes the choice between them less about overall capability and more about which API and workload profile matters most.

Where Each One Wins

The AMD Radeon R9 370X is the clear compute champion. In the Geekbench OpenCL test, it scores 25,893 against the GTX 965M’s 14,509, a commanding 78.5% advantage. This is not a marginal lead; it is a near-doubling of raw throughput in a widely-used compute API. The R9 370X also holds a higher average benchmark score overall (15,862 vs 14,404), and its FP32 throughput of 2.637 TFLOPS outpaces the GTX 965M’s 1.946 TFLOPS. For any workload that leverages OpenCL—scientific simulation, video encoding, or general-purpose GPU compute—the data strongly favors the AMD card.

The NVIDIA GeForce GTX 965M wins the Vulkan battle outright. It scores 14,299 in Geekbench Vulkan, versus the R9 370X’s 9,018, a 36.9% deficit for the AMD card. This is a substantial reversal of the OpenCL result. The GTX 965M’s Vulkan support is also newer (version 1.4 vs 1.2.170), and its DirectX 12 support is higher-tier (12_1 vs 11_1). For gaming or graphics applications that leverage Vulkan or modern DirectX 12 features, the NVIDIA module is the better performer. The head-to-head record is tied at one win each, which underscores that this is not a universally superior product—it is a workload-dependent split.

Architecture Differences

The two GPUs come from different architectural generations and design philosophies. The AMD Radeon R9 370X uses the Trinidad chip based on GCN 1.0 architecture, manufactured on a 28 nm process at TSMC. It contains 2,800 million transistors on a 212 mm² die, yielding a transistor density of 13.2M per mm². The GTX 965M, by contrast, uses the GM204 chip with Maxwell 2.0 architecture, also on 28 nm at TSMC, but packs 5,200 million transistors on a 398 mm² die—a density of 13.1M per mm². The NVIDIA chip is physically much larger and has nearly double the transistor count, yet it is configured with fewer active cores.

The R9 370X features 1,280 shading units, 80 texture mapping units, and 32 ROPs. Its memory subsystem is a 2 GB GDDR5 pool on a 256-bit bus, delivering 179.2 GB/s of bandwidth. Clock speeds are 980 MHz base and 1030 MHz boost, with memory at 1400 MHz (5.6 Gbps effective). The GTX 965M has 1,024 shading units, 64 TMUs, and 32 ROPs. Its memory is also 2 GB GDDR5, but on a narrower 128-bit bus, yielding only 80.19 GB/s—less than half the AMD card’s bandwidth. The NVIDIA module runs at 924 MHz base and 950 MHz boost, with memory at 1253 MHz (5 Gbps effective). The AMD part’s higher clock speeds and wider memory bus explain its lead in fill-rate metrics: 32.96 GPixel/s vs 30.40, and 82.40 GTexel/s vs 60.80.

The power and form factor differences are stark. The R9 370X is a dual-slot desktop card, 221 mm long, with a 180 W TDP and a suggested 450 W PSU. The GTX 965M is an MXM module with no power connectors and no suggested PSU, making it suitable only for laptops or compact systems. The AMD card requires external power and substantial cooling; the NVIDIA module is self-contained. Both support PCIe 3.0, but the R9 370X uses a standard x16 slot while the GTX 965M uses the MXM-B (3.0) interface.

The Verdict

The data supports a straightforward recommendation based on use case. For compute-heavy workloads that rely on OpenCL, the AMD Radeon R9 370X is the superior choice. Its 78.5% lead in Geekbench OpenCL is the single largest performance gap in the head-to-head results. Higher FP32 throughput, more shading units, and double the memory bandwidth all contribute to this result. The R9 370X also edges out the GTX 965M on average benchmark score by about 10.1% (15,862 vs 14,404). Users running OpenCL-based applications—whether for rendering, analysis, or other GPU-accelerated tasks—should favor the AMD card.

For Vulkan-based graphics workloads, the NVIDIA GeForce GTX 965M is the clear winner. It leads by 36.9% in Geekbench Vulkan, and its newer Vulkan 1.4 API support plus DirectX 12 (12_1) feature level give it a modern API advantage over the AMD card’s Vulkan 1.2.170 and DirectX 12 (11_1). The GTX 965M’s 14,299 Vulkan score is actually higher than its own OpenCL score, while the R9 370X’s Vulkan score is less than half its OpenCL result—suggesting the NVIDIA architecture is better optimized for graphics pipelines. Gamers or developers targeting Vulkan should pick the GTX 965M.

The choice also depends on the physical platform. The R9 370X is a desktop-only card requiring a dual-slot space and a 450 W PSU. The GTX 965M is a mobile module, meaning it is only relevant for systems that support MXM upgrades. If the platform is fixed, the decision is made. If both are possible, the tie in benchmark wins (one each) means the user must prioritize their API of choice. The average scores are close, but the performance profiles are not interchangeable.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon R9 370X has an average benchmark score of 15,862, which is 10.1% higher than the NVIDIA GeForce GTX 965M’s 14,404.

Q: How much faster is the R9 370X in OpenCL?

A: The R9 370X scores 25,893 in Geekbench OpenCL, which is 78.5% higher than the GTX 965M’s 14,509.

Q: Does the GTX 965M beat the R9 370X in any benchmark?

A: Yes, the GTX 965M wins the Geekbench Vulkan test by 36.9%, scoring 14,299 against the R9 370X’s 9,018.

Q: What is the memory bandwidth difference between the two?

A: The R9 370X has 179.2 GB/s of bandwidth on a 256-bit bus, while the GTX 965M has 80.19 GB/s on a 128-bit bus.

Q: Which GPU supports a newer version of Vulkan?

A: The NVIDIA GeForce GTX 965M supports Vulkan 1.4, while the AMD Radeon R9 370X supports Vulkan 1.2.170.

Q: Are both GPUs the same physical size?

A: No, the R9 370X is a dual-slot desktop card measuring 221 mm in length, while the GTX 965M is an MXM module with no listed dimensions.

Head-to-Head Benchmarks

The two GPUs split their head-to-head results exactly evenly, with one win each. The largest margin comes in Geekbench OpenCL, where the AMD Radeon R9 370X scores 25,893. This is a 78.5% improvement over the NVIDIA GeForce GTX 965M’s 14,509. The magnitude of this gap is significant; it is not a small edge but a dominant performance difference. The R9 370X’s higher clock speeds (980 MHz base vs 924 MHz), more shading units (1,280 vs 1,024), and wider memory bus (256-bit vs 128-bit) all contribute to this result. The OpenCL score also aligns with the FP32 throughput difference: 2.637 TFLOPS vs 1.946 TFLOPS.

The Vulkan result reverses the picture completely. The NVIDIA GeForce GTX 965M scores 14,299, which is 36.9% higher than the AMD Radeon R9 370X’s 9,018. This is a substantial loss for the AMD card, and it is notable that the GTX 965M’s Vulkan score is close to its OpenCL score, while the R9 370X’s Vulkan score is only about 35% of its OpenCL score. The NVIDIA card’s Maxwell 2.0 architecture appears better suited to Vulkan’s low-level draw call management and graphics pipeline. The GTX 965M’s newer Vulkan 1.4 support and higher DirectX 12 feature level (12_1 vs 11_1) further reinforce this advantage.

Looking at the broader context, both GPUs sit near the same performance tier. The R9 370X’s nearest rivals include the AMD Radeon RX 9060 (average score 16,014, 1% higher) and the NVIDIA GeForce RTX 3060 Ti (average score 16,129, 1.7% higher). The GTX 965M’s nearest rivals include the AMD Radeon RX Vega 11 (14,385, 0.1% higher) and the NVIDIA GeForce GTX TITAN (14,373, 0.2% higher). These deltas are small, indicating that both cards perform near the expected level for their respective classes. However, the within-pair differences are much larger—the 78.5% OpenCL lead and 36.9% Vulkan lead show that API choice matters more than the overall average score.

The transistor counts tell an interesting story. The GTX 965M’s GM204 chip has 5,200 million transistors on a 398 mm² die, nearly double the R9 370X’s 2,800 million on 212 mm². Yet the NVIDIA part uses fewer active shading units and delivers lower FP32 throughput. This suggests the Maxwell architecture is more efficient per transistor in graphics workloads, which is consistent with its Vulkan win. The AMD GCN 1.0 architecture, while older, achieves higher raw compute and memory bandwidth, which is why it wins so decisively in OpenCL. The 28 nm process node is identical for both, and transistor density is essentially the same (13.2M vs 13.1M per mm²), so the performance differences come down to architectural design choices rather than manufacturing advantages.