AMD Radeon 740M vs NVIDIA GeForce GTX 680 Comparison

# NVIDIA GeForce GTX 680 vs AMD Radeon 740M

The NVIDIA GeForce GTX 680 and AMD Radeon 740M represent two very different eras of GPU design. The GTX 680 is a 2012 discrete desktop card built on Kepler architecture, while the Radeon 740M is a 2024 integrated processor from the Phoenix2 chip using RDNA 3.0. In the database, the GTX 680 holds a 55th percentile ranking among all GPUs, while the Radeon 740M sits at the 53rd percentile. Their average benchmark scores are close at first glance: the GTX 680 averages 14,150 across its recorded tests, versus 12,870 for the Radeon 740M. However, the head-to-head results show a clear pattern: the GTX 680 wins both recorded comparisons, with a 66.2% lead in OpenCL and a 13.3% lead in Vulkan.

Where Each One Wins

The split between these two GPUs is about workload compatibility rather than raw capability. The GTX 680 wins both recorded head-to-head benchmarks, but the margin tells the story. In Geekbench OpenCL, the GTX 680 scores 16,906 against 10,172 for the Radeon 740M, a decisive 66.2% advantage. That is the kind of result that indicates a substantial compute throughput gap in general-purpose workloads. In Geekbench Vulkan, the GTX 680 scores 17,646 against 15,568, a narrower 13.3% lead. The Vulkan result is the closer contest, suggesting that the Radeon 740M's modern architecture narrows the gap in graphics-oriented API workloads.

The Radeon 740M's wins are not reflected in the head-to-head table, as it records zero victories in the two shared tests. However, its profile shows strengths that the GTX 680 cannot match in other contexts. The Radeon 740M is an integrated GPU with a 45 W TDP, no power connectors, and a slot width of IGP, making it suitable for compact, low-power systems. The GTX 680, by contrast, is a dual-slot card with a 195 W TDP, two 6-pin power connectors, and a suggested 450 W power supply. The Radeon 740M supports DirectX 12 Ultimate (12_2), while the GTX 680 only reaches DirectX 12 (11_0). For modern API features, the Radeon 740M is the more forward-looking option.

The GTX 680's advantage in OpenCL is substantial enough to recommend it for compute-heavy tasks that leverage that API. The Radeon 740M's Vulkan score, while lower, is competitive enough that it can handle Vulkan-based games without the same performance deficit. The data shows a clear hierarchy: the GTX 680 is the faster GPU in both shared tests, but the Radeon 740M offers a different set of tradeoffs in power, size, and API support.

Architecture Differences

The architectural gap between these two GPUs is generational. The GTX 680 uses the GK104 chip on a 28 nm TSMC process, with 3,540 million transistors on a 294 mm² die. The Radeon 740M uses the Phoenix2 chip on a 4 nm TSMC process, with 20,900 million transistors on a 137 mm² die. The transistor density difference is stark: 12.0M per mm² for the GTX 680 versus 152.6M per mm² for the Radeon 740M. The newer process allows AMD to pack nearly six times more transistors per square millimeter, which explains how an integrated GPU can approach the performance of a decade-old flagship.

The compute configurations differ significantly. The GTX 680 has 1,536 shading units, 128 texture mapping units, and 32 ROPs. The Radeon 740M has 256 shading units, 16 TMUs, and 8 ROPs. Despite having far fewer shading units, the Radeon 740M reaches 2.867 TFLOPS of FP32 performance, while the GTX 680 reaches 3.250 TFLOPS. The GTX 680 still leads in raw FP32 throughput, but the Radeon 740M's higher boost clock of 2800 MHz (versus 1058 MHz for the GTX 680) helps compensate for the smaller execution footprint. The Radeon 740M also includes 4 ray tracing cores, a feature the GTX 680 lacks entirely.

Memory is another major divergence. The GTX 680 has 2 GB of GDDR5 on a 256 bit bus, delivering 192.3 GB/s of bandwidth. The Radeon 740M uses system shared memory, with bandwidth described as system dependent. This gives the discrete card a significant advantage in dedicated memory bandwidth, but it also means the Radeon 740M's performance scales with the host system's memory configuration. The GTX 680's pixel rate is 33.86 GPixel/s versus 22.40 GPixel/s for the Radeon 740M, and its texture rate is 135.4 GTexel/s versus 44.80 GTexel/s.

API support is a point in AMD's favor. The Radeon 740M supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the GTX 680 supports DirectX 12 (11_0) and Vulkan 1.2.175. Both support OpenGL 4.6. The Radeon 740M also offers FP16 at 2.867 TFLOPS (1:1), while the GTX 680 has no recorded FP16 capability. The bus interface differs as well: PCIe 3.0 x16 for the GTX 680 versus PCIe 4.0 x8 for the Radeon 740M.

Head-to-Head Benchmarks

The OpenCL test is the largest gap in the recorded data. The GTX 680 scores 16,906, while the Radeon 740M scores 10,172. This is a 66.2% advantage for the NVIDIA card, and it reflects the GTX 680's higher FP32 throughput and dedicated memory bandwidth. In compute workloads that scale with raw shading power and memory bandwidth, the older discrete card maintains a commanding lead.

The Vulkan test is closer. The GTX 680 scores 17,646, and the Radeon 740M scores 15,568. The 13.3% margin is significant but far from the OpenCL gap. The Radeon 740M's modern RDNA 3.0 architecture likely contributes to its relatively stronger showing in a modern graphics API, even with fewer shading units and shared memory. The GTX 680's Vulkan score is also its highest recorded benchmark, while the Radeon 740M's Vulkan score is its best as well.

Looking at the nearest rivals in the database provides context. The GTX 680's average score of 14,150 places it 0.9% above the NVIDIA Tesla K10 (14,029), 0.9% below the GeForce GTX 1070 Ti (14,277), 1.2% below the Intel Iris Xe MAX Graphics (14,315), and 1.4% below the AMD Radeon Vega 11 (14,352). The Radeon 740M's average of 12,870 sits 0.2% above the AMD FirePro W5100 (12,847), 0.3% above the AMD Radeon Pro 455 (12,831), 0.3% above the NVIDIA GeForce GTX 590 (12,830), and 0.4% below the AMD Radeon RX 580 (12,928). These positions show that both GPUs sit in a crowded performance band, but the GTX 680 occupies a slightly higher tier.

FAQ

Q: Which GPU is faster overall?

A: The NVIDIA GeForce GTX 680 is faster in both recorded head-to-head tests. It leads by 66.2% in Geekbench OpenCL and by 13.3% in Geekbench Vulkan.

Q: How do their average benchmark scores compare?

A: The GTX 680 has an average benchmark score of 14,150, while the Radeon 740M averages 12,870. The GTX 680 also ranks at the 55th percentile among all GPUs, versus the 53rd percentile for the Radeon 740M.

Q: Can the Radeon 740M compete in modern APIs?

A: The Radeon 740M supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the GTX 680 supports DirectX 12 (11_0) and Vulkan 1.2.175. In the Vulkan benchmark, the Radeon 740M comes within 13.3% of the GTX 680.

Q: What are the power requirements for each?

A: The GTX 680 has a TDP of 195 W, requires two 6-pin power connectors, and needs a suggested 450 W power supply. The Radeon 740M has a TDP of 45 W, uses no power connectors, and is an integrated GPU.

Q: Does the Radeon 740M support ray tracing?

A: Yes, the Radeon 740M includes 4 ray tracing cores. The GTX 680 has no ray tracing cores.

Q: How does memory differ between the two?

A: The GTX 680 has 2 GB of GDDR5 on a 256 bit bus with 192.3 GB/s bandwidth. The Radeon 740M uses system shared memory with system dependent bandwidth.

The Verdict

The data supports a clear choice for users who prioritize raw performance: the NVIDIA GeForce GTX 680 wins both recorded benchmark comparisons, and its 66.2% OpenCL lead is the largest margin in the head-to-head results. For compute workloads that rely on OpenCL, the GTX 680 is the stronger option by a wide margin. Its higher FP32 throughput (3.250 TFLOPS versus 2.867 TFLOPS), dedicated 192.3 GB/s memory bandwidth, and higher pixel and texture rates all contribute to this result.

However, the Radeon 740M is not without a case. Its 45 W TDP, integrated form factor, and lack of power connectors make it suitable for systems where the GTX 680's dual-slot size and 195 W power draw would be impractical. The Radeon 740M also supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, bringing modern API features and ray tracing cores that the GTX 680 cannot offer. Its Vulkan score of 15,568 is within 13.3% of the GTX 680, showing that the architectural gap narrows substantially in newer graphics APIs.

Users with an existing desktop platform that can accommodate a discrete, dual-slot card should favor the GTX 680 for its benchmark dominance. Users building a compact, power-efficient system with an integrated GPU should consider the Radeon 740M, accepting lower benchmark scores in exchange for modern features and minimal power consumption. The database positions the GTX 680 at the 55th percentile and the Radeon 740M at the 53rd, a two-point difference that reflects the GTX 680's edge in average score. The GTX 680 is the performance winner, while the Radeon 740M wins on efficiency and feature currency.