AMD Radeon 740M vs NVIDIA GeForce GTX 960A Comparison

The AMD Radeon 740M and NVIDIA GeForce GTX 960A represent two very different approaches to mobile graphics, separated by nearly a decade of architectural evolution. The data shows a narrow but decisive victory for the older NVIDIA part in the available benchmark, yet the AMD integrated processor brings substantial modern feature support and a significantly higher peak compute rating. This comparison is less about raw speed and more about what each chip prioritizes: legacy bandwidth and shader count versus modern process technology and API compliance.

Head-to-Head Benchmarks

The only directly comparable benchmark in the dataset is Geekbench OpenCL, and the results are surprisingly close. The NVIDIA GeForce GTX 960A scores 11,998 points, while the AMD Radeon 740M scores 10,172 points. This gives the GTX 960A a 15.2% lead in this specific test, marking the sole win for NVIDIA in the head-to-head comparison.

This result is notable because the Radeon 740M carries a much higher theoretical FP32 throughput of 2.867 TFLOPS compared to the GTX 960A's 1.505 TFLOPS. The 960A is effectively 47.5% lower in raw compute, yet it still outperforms the newer part in OpenCL. This suggests that the GTX 960A's dedicated memory subsystem — a 128-bit bus with 80.19 GB/s of bandwidth and 2 GB of GDDR5 — provides a practical advantage that raw shader count cannot overcome in this workload.

Contextualizing the scores against their respective rivals highlights how close these parts are in the broader market. The 740M's average benchmark score of 12,870 places it just 0.2% ahead of the AMD FirePro W5100 and 0.3% ahead of both the AMD Radeon Pro 455 and the NVIDIA GeForce GTX 590. Meanwhile, the GTX 960A's average score of 11,998 puts it 0.3% ahead of the NVIDIA GeForce GTX 1080, a much higher-tier part, and 1.3% ahead of the AMD Radeon RX 6500 XT. Both chips occupy a similar performance percentile — the 740M sits at the 53rd percentile of all GPUs, while the 960A sits at the 51st percentile — indicating they are within the same tier of overall capability.

The single benchmark win is a narrow one, but it is consistent. The GTX 960A's advantage in OpenCL may stem from its dedicated VRAM and mature driver optimizations for that API. The Radeon 740M, which relies on system memory with bandwidth described as "System Dependent," is more susceptible to memory performance variability.

FAQ

Q: Which GPU has the higher raw compute throughput?

A: The AMD Radeon 740M, with an FP32 rating of 2.867 TFLOPS. The NVIDIA GeForce GTX 960A is rated at 1.505 TFLOPS, which is roughly half the compute capability on paper.

Q: Does the NVIDIA card support modern graphics APIs?

A: The GTX 960A supports DirectX 12 (11_0) and Vulkan 1.4. It does not support DirectX 12 Ultimate features, and its OpenGL support is 4.6.

Q: What is the memory situation for each chip?

A: The GTX 960A has a fixed 2 GB of GDDR5 memory on a 128-bit bus, providing 80.19 GB/s of bandwidth. The Radeon 740M uses system-shared memory, with its bus width and bandwidth listed as "System Shared" and "System Dependent," respectively.

Q: How do their transistor counts compare?

A: The Radeon 740M is built with 20,900 million transistors, while the GTX 960A has 1,870 million. This difference reflects the vastly different die sizes and process nodes.

Q: Which GPU has a higher pixel fill rate?

A: The Radeon 740M, at 22.40 GPixel/s, edges out the GTX 960A's 18.82 GPixel/s. However, the GTX 960A has a higher texture fill rate at 47.04 GTexel/s compared to the 740M's 44.80 GTexel/s.

Q: What is the form factor of each product?

A: The Radeon 740M is an integrated graphics processor (IGP) with no power connectors, designed to be part of a CPU package. The GTX 960A is an MXM module, a removable mobile graphics card standard.

Architecture Differences

The two GPUs are built on fundamentally different architectures from different eras. The AMD Radeon 740M is based on the RDNA 3.0 architecture, using the Phoenix2 chip, and is part of the Navi III IGP generation. It is fabricated on a 4 nm process at TSMC, a leading-edge node that allows for an extremely dense design.

In contrast, the NVIDIA GeForce GTX 960A uses the Maxwell architecture with the GM107 chip, belonging to the GeForce 900A generation. It is manufactured on a much older 28 nm process, also at TSMC. The process node difference is stark: 4 nm versus 28 nm. This allows the AMD chip to pack 20,900 million transistors into a die size of 137 mm², yielding a transistor density of 152.6M per mm². The NVIDIA chip has only 1,870 million transistors on a larger 148 mm² die, resulting in a density of just 12.6M per mm².

The core configurations also differ significantly. The Radeon 740M features 256 shading units, 16 texture mapping units (TMUs), and 8 render output units (ROPs). Critically, it also includes 4 dedicated ray tracing cores, a feature entirely absent from the GTX 960A. The NVIDIA part, by contrast, has a much higher count of 640 shading units, 40 TMUs, and 16 ROPs. This means the GTX 960A relies on a wider, more traditional shader array, while the Radeon 740M uses a narrower but more modern design with specialized hardware.

The memory architecture is another major divergence. The Radeon 740M uses system-shared memory, where the GPU accesses the main system RAM. This makes its bandwidth and capacity dependent on the host system's configuration. The GTX 960A has dedicated 2 GB of GDDR5 memory on a 128-bit interface, providing a consistent 80.19 GB/s of bandwidth. This dedicated memory is a key reason for its competitive performance despite its older architecture.

Specification Differences

The most obvious difference is the process node and resulting physical characteristics. The Radeon 740M is built on a 4 nm process with a die size of 137 mm², while the GTX 960A uses a 28 nm process with a 148 mm² die. Transistor counts are 20,900 million and 1,870 million, respectively.

Clock speeds tell a story of efficiency versus brute force. The Radeon 740M has a base clock of 800 MHz and a boost clock of 2800 MHz. The GTX 960A has a lower base clock of 1097 MHz but also a much lower boost clock of 1176 MHz. The AMD chip's boost clock is more than double that of the NVIDIA chip, a direct result of the advanced process node.

Memory specifications are entirely different. The Radeon 740M has system-shared memory with a type of "System Shared" and a bus width of "System Shared." Its bandwidth is "System Dependent." The GTX 960A has 2 GB of GDDR5, a 128-bit bus, and 80.19 GB/s of bandwidth. Its memory clock is listed as 1253 MHz with 5 Gbps effective data rate.

Compute and fill rates are mixed. The Radeon 740M leads in FP32 performance (2.867 TFLOPS vs. 1.505 TFLOPS) and pixel rate (22.40 GPixel/s vs. 18.82 GPixel/s). The GTX 960A leads in texture rate (47.04 GTexel/s vs. 44.80 GTexel/s) and has more TMUs and ROPs.

Power and form factor also differ. The Radeon 740M has a TDP of 45 W and is an IGP with no power connectors. The GTX 960A has a TDP of 75 W and is an MXM Module with no power connectors. The Radeon 740M uses a PCIe 4.0 x8 bus interface, while the GTX 960A uses an MXM-B (3.0) interface.

Finally, the API support differs. The Radeon 740M supports DirectX 12 Ultimate (12_2), while the GTX 960A supports DirectX 12 (11_0). Both support OpenGL 4.6 and Vulkan 1.4. The Radeon 740M also has 4 ray tracing cores, a feature the GTX 960A lacks entirely.

Where Each One Wins

The NVIDIA GeForce GTX 960A wins in the only head-to-head benchmark available, the Geekbench OpenCL test. It scores 11,998 versus the Radeon's 10,172, a 15.2% advantage. This makes it the better choice for compute workloads that rely heavily on OpenCL, particularly those that benefit from dedicated memory bandwidth. Its 2 GB of GDDR5 and 80.19 GB/s of bandwidth provide a stable foundation that does not depend on system memory performance.

The AMD Radeon 740M wins on raw compute capability and modern features. Its FP32 throughput of 2.867 TFLOPS is nearly double that of the GTX 960A. It also has a higher pixel rate of 22.40 GPixel/s. The inclusion of 4 ray tracing cores and support for DirectX 12 Ultimate makes it the more future-proof option for gaming and applications that leverage these newer APIs. Its 45 W TDP is significantly lower than the GTX 960A's 75 W, making it more power-efficient for integrated use.

The Radeon 740M also wins on technological longevity. It is an active production part with a release date of 2024, while the GTX 960A is end-of-life with a 2015 release date. The Radeon's 4 nm process and higher transistor density suggest it can achieve its performance with far less power draw per unit of compute.

The Verdict

The data presents a clear trade-off. For users who need maximum performance in legacy OpenCL applications and have a system that supports MXM modules, the NVIDIA GeForce GTX 960A is the definitive winner. Its 15.2% lead in the head-to-head benchmark and its dedicated 80.19 GB/s memory bandwidth make it a more reliable performer in that specific context. Its 640 shading units and 40 TMUs provide substantial traditional rasterization muscle, as evidenced by its higher texture rate of 47.04 GTexel/s.

However, for anyone building a modern system or upgrading an integrated graphics setup, the AMD Radeon 740M is the more sensible choice. Despite losing the OpenCL test, it offers double the FP32 compute, a higher pixel rate, and a much lower TDP. Its support for DirectX 12 Ultimate and ray tracing cores makes it capable of handling modern visual effects that the GTX 960A simply cannot process. The 740M's performance is also tied to the system's memory, meaning a fast dual-channel configuration could potentially close the gap seen in the benchmark.

Ultimately, the GTX 960A is a specialized legacy part that wins on a narrow metric. The Radeon 740M is a general-purpose modern IGP that loses on that metric but wins on nearly every other technical specification. The choice depends on whether the user prioritizes a single benchmark score from 2015-era software or the broad feature set and efficiency of a 2024-era design. The benchmark data shows the GTX 960A as the faster part in OpenCL, but the specification data overwhelmingly favors the Radeon 740M for contemporary use.