AMD Radeon 680M vs NVIDIA GeForce GTX 770 Comparison

Head-to-Head Benchmarks

The benchmark data presents a clear generational divide between these two GPUs. In the two head-to-head tests available, the AMD Radeon 680M wins decisively on both occasions, though the margins vary significantly by workload type. The most striking result comes from the Geekbench OpenCL test, where the 680M scores 23,468 points against the GTX 770's 15,707 points. That represents a 49.4% advantage for the integrated part—a substantial lead that underscores the architectural efficiency gains AMD has achieved over the past decade.

The gap narrows considerably in the Vulkan API test, but the 680M still comes out ahead. Its score of 21,965 beats the GTX 770's 19,166 by 14.6%. This smaller delta suggests that the older Kepler architecture retains more competitive performance under modern low-level graphics APIs, likely due to its higher raw shading unit count and texture throughput. The GTX 770 fields 1,536 shading units and 128 texture mapping units, compared to the 680M's 768 shading units and 48 TMUs, yet the newer RDNA 2.0 design still finds a way to win.

Looking at the broader benchmark picture beyond the direct head-to-head, the 680M maintains its edge in average score. The AMD part posts an average benchmark score of 15,270 across all tested workloads, while the GTX 770 manages 14,747. Both GPUs sit at the 57th percentile among all GPUs in the database, which places them in the same performance tier overall. The 680M's nearest rivals include the NVIDIA GeForce GTX 580 (average score 15,283, deltaPct -0.1%), the GeForce RTX 2060 (15,290, -0.1%), the RTX 3050 OEM (15,199, +0.5%), and the AMD Radeon RX 7600 (15,171, +0.7%). The GTX 770's competitive set is notably different, with the Radeon RX 5500 XT (14,692, +0.4%), Quadro M2000 (14,532, +1.5%), GeForce GTX 660 Ti (15,063, -2.1%), and Radeon Pro 560X (15,082, -2.2%) all clustering within a few percentage points.

The 680M's OpenCL result is particularly strong when viewed against its rival set. It sits 0.7% above the RX 7600 and 0.5% above the RTX 3050 OEM, both of which are discrete desktop parts with dedicated memory. That the integrated 680M can trade blows with these cards in compute workloads speaks to the efficiency of its 6nm process node and RDNA 2.0 architecture. Meanwhile, the GTX 770's 1.5% lead over the Quadro M2000 and 0.4% edge over the RX 5500 XT shows it remains competitive with mid-range GPUs from several generations later, but it cannot match the modern integrated solution.

The Verdict

The data points to a clear conclusion: the AMD Radeon 680M outperforms the NVIDIA GeForce GTX 770 in every benchmark where they directly compete. The 49.4% OpenCL victory is the defining statistic, while the 14.6% Vulkan win reinforces the trend. For users choosing between these two GPUs today, the 680M is the stronger performer in compute-heavy tasks and holds a meaningful advantage in modern graphics APIs.

However, the choice is not purely about raw performance. The GTX 770 is a dual-slot, 230W discrete card requiring a 550W power supply and both a 6-pin and 8-pin power connector. The 680M is an IGP with a 50W TDP, no power connectors, and system-shared memory. The GTX 770 does offer 2GB of dedicated GDDR5 memory across a 256-bit bus with 224.4 GB/s bandwidth—the 680M's memory bandwidth is system dependent and shared with the CPU. For workloads that are sensitive to memory bandwidth, the GTX 770's dedicated VRAM could prove advantageous, even if the compute scores favor the 680M.

The GTX 770 also carries a launch MSRP of 399 USD, though it is now end-of-life. The 680M remains in active production. Users with existing GTX 770 boards may find them serviceable for legacy applications, but the 680M represents the more future-proof choice given its DirectX 12 Ultimate (12_2) support versus the GTX 770's DirectX 12 (11_0), and Vulkan 1.4 versus 1.2.175. The verdict from the data is straightforward: the 680M wins on performance, power efficiency, and API support.

Architecture Differences

The two GPUs come from entirely different eras of GPU design. The AMD Radeon 680M uses the Rembrandt+ chip built on RDNA 2.0 architecture, fabricated on TSMC's 6nm process. It packs 13,100 million transistors into a 208 mm² die, yielding a transistor density of 63.0 million per square millimeter. The GTX 770 uses the GK104 chip on Kepler architecture, built on TSMC's 28nm process with 3,540 million transistors across a 294 mm² die—a density of just 12.0 million per square millimeter.

The process node difference is stark: 6nm versus 28nm represents multiple generations of manufacturing improvements. This explains how the 680M achieves comparable or better performance while consuming only 50W against the GTX 770's 230W. The 680M's transistor density is more than five times greater, allowing far more logic per square millimeter.

Core configurations differ substantially. The 680M fields 768 shading units, 48 TMUs, and 32 ROPs, along with 12 ray tracing cores—a feature the GTX 770 lacks entirely. The GTX 770 counters with 1,536 shading units and 128 TMUs, but also 32 ROPs. This gives the Kepler part a theoretical texture rate of 138.9 GTexel/s versus the 680M's 105.6 GTexel/s. However, the 680M's pixel rate of 70.40 GPixel/s more than doubles the GTX 770's 34.72 GPixel/s, suggesting better fill-rate efficiency per ROP.

Compute throughput is nearly identical on paper. The 680M delivers 3.379 TFLOPS FP32, while the GTX 770 manages 3.333 TFLOPS. The 680M also supports FP16 at 6.758 TFLOPS (2:1 rate), whereas the GTX 770 has no FP16 capability listed. Clock speeds favor the AMD part: 2000 MHz base and 2200 MHz boost versus 1046 MHz base and 1085 MHz boost for the GTX 770. The GTX 770's memory runs at 1753 MHz (7 Gbps effective) across a 256-bit bus.

Memory architecture is fundamentally different. The 680M uses system-shared memory with system-dependent bandwidth; the GTX 770 uses 2GB of dedicated GDDR5. The bus interface also differs: PCIe 4.0 x8 for the 680M versus PCIe 3.0 x16 for the GTX 770. The 680M is an IGP with no slot width and portable-device-dependent display outputs, while the GTX 770 is a 267mm dual-slot card with 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2 outputs.

FAQ

Q: Which GPU is faster in OpenCL compute workloads?

A: The AMD Radeon 680M scores 23,468 in Geekbench OpenCL versus the GTX 770's 15,707, a 49.4% advantage for the 680M.

Q: Does the GTX 770 win any benchmark comparison?

A: No. In the two head-to-head tests recorded, the 680M wins both: OpenCL by 49.4% and Vulkan by 14.6%. The GTX 770 has zero wins in the head-to-head data.

Q: How do their power requirements compare?

A: The 680M is an IGP with a 50W TDP and no power connectors. The GTX 770 has a 230W TDP, requires one 6-pin and one 8-pin power connector, and needs a 550W suggested power supply.

Q: What are the API support differences?

A: The 680M supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The GTX 770 supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175.

Q: Does the GTX 770 have more shading units?

A: Yes, the GTX 770 has 1,536 shading units and 128 TMUs, while the 680M has 768 shading units and 48 TMUs. Despite this, the 680M wins in compute benchmarks.

Q: What is the GTX 770's launch MSRP?

A: The GTX 770 had a launch MSRP of 399 USD.

Where Each One Wins

The AMD Radeon 680M wins decisively in compute performance. Its 49.4% lead in OpenCL and 14.6% lead in Vulkan make it the clear choice for applications that leverage these APIs. The 680M also wins on power efficiency by a wide margin—50W versus 230W TDP—making it suitable for compact, low-power systems where the GTX 770's dual-slot, 267mm card format with external power connectors would be impractical. The 680M's ray tracing cores provide hardware acceleration for ray-traced workloads, a feature entirely absent from the GTX 770. Its newer API support (DirectX 12 Ultimate, Vulkan 1.4) also positions it better for modern games and applications.

The NVIDIA GeForce GTX 770 retains advantages in specific areas, though not in the direct benchmark comparisons. Its 2GB of dedicated GDDR5 memory on a 256-bit bus with 224.4 GB/s bandwidth provides predictable, consistent memory performance that does not depend on system DRAM configuration. The 680M's system-shared memory with system-dependent bandwidth means real-world performance can vary based on the host system's memory speed and capacity. The GTX 770's higher texture rate (138.9 GTexel/s versus 105.6 GTexel/s) could benefit texture-heavy workloads, and its 1,536 shading units provide raw parallel processing capacity that might advantage certain legacy workloads optimized for high shader counts.

The GTX 770 also wins on connectivity for desktop use with its 2x DVI, HDMI 1.4a, and DisplayPort 1.2 outputs, versus the 680M's portable-device-dependent display outputs. For users with existing desktop systems that can accommodate a 230W card and have the necessary power connectors, the GTX 770 remains a functional option for older DirectX 11-era applications. However, the benchmark data shows that in modern compute and graphics API workloads, the 680M's architectural advantages—6nm process, RDNA 2.0 design, ray tracing support, and higher clock speeds—deliver superior results despite the GTX 770's larger memory bus and higher shader count.