AMD Radeon 680M vs NVIDIA GeForce GTX 480 Comparison

Where Each One Wins

The benchmark data splits cleanly between these two very different designs. The AMD Radeon 680M wins the only directly comparable workload, Geekbench OpenCL, with a score of 23,468 against the GTX 480's 13,300. That is a 76.5% advantage, a decisive margin that places the integrated part firmly ahead in compute-oriented tasks.

The Radeon 680M also holds advantages in several other measurable categories. Its average benchmark score sits at 15,270, while the GTX 480 averages 13,300. The 680M's percentile ranking among all GPUs is 57, versus 53 for the GTX 480. The 680M records additional results in 3DMark Steel Nomad DX12 (378) and Geekbench Vulkan (21,965), while the GTX 480 has no recorded result in either test. The GTX 480 has no benchmark wins in the head-to-head comparison, and no separate wins in any other recorded test.

The GTX 480's strengths are not visible in the recorded benchmark suite. Its only data point is the Geekbench OpenCL score, where it loses outright. What remains are architectural and specification differences that favor it in legacy scenarios: a dedicated 1536 MB GDDR5 frame buffer with a 384-bit bus and 177.4 GB/s bandwidth, versus the 680M's system-shared memory with dependent bandwidth. The GTX 480 also has more texture mapping units (60 versus 48) and more raster output pipelines (48 versus 32), which historically benefits fill-rate-bound workloads. None of those advantages translate into a winning benchmark score in the database.

The practical split is therefore clear: the 680M wins every measured comparison, while the GTX 480 retains theoretical advantages in raw memory throughput and fill rate that do not appear in the recorded data. For any workload represented in the database, the 680M is the stronger part.

Architecture Differences

The two GPUs come from opposite ends of GPU design history. The AMD Radeon 680M uses the Rembrandt+ chip built on RDNA 2.0 architecture, fabricated on a 6 nm process at TSMC with 13,100 million transistors on a 208 mm² die. That yields a transistor density of 63.0 million per mm². The NVIDIA GeForce GTX 480 uses the GF100 chip on Fermi architecture, fabricated on a 40 nm process, also at TSMC, with 3,100 million transistors on a 529 mm² die. Its transistor density is 5.9 million per mm², a dramatic 10.7x difference in density that reflects over a decade of process advancement.

The 680M is an integrated graphics processor (IGP) with no standalone slot width, no power connectors, and a 50 W TDP. It interfaces via PCIe 4.0 x8 and relies entirely on system-shared memory, meaning its bandwidth is system dependent. The GTX 480 is a dual-slot discrete card, 267 mm long, with 1x 6-pin and 1x 8-pin power connectors, a 250 W TDP, and a suggested 600 W power supply. It uses PCIe 2.0 x16 and has 1536 MB of dedicated GDDR5 memory on a 384-bit bus with 177.4 GB/s of bandwidth.

Compute resources differ substantially. The 680M has 768 shading units, 48 TMUs, 32 ROPs, and 12 ray tracing cores. The GTX 480 has 480 shading units, 60 TMUs, and 48 ROPs, with no ray tracing cores at all. Clock behavior also differs: the 680M runs at a 2000 MHz base and 2200 MHz boost, while the GTX 480's core clocks are not recorded in the database. Memory clock for the GTX 480 is 924 MHz, or 3.7 Gbps effective.

The 680M supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The GTX 480 supports DirectX 12 (11_0) and OpenGL 4.6, with no Vulkan support recorded. The 680M posts 70.40 GPixel/s pixel rate, 105.6 GTexel/s texture rate, 3.379 TFLOPS FP32, and 6.758 TFLOPS FP16 (2:1). The GTX 480 posts 21.03 GPixel/s, 42.06 GTexel/s, and 1,345.0 GFLOPS FP32, with no FP16 figure recorded.

Release timing further separates them: the 680M launched in January 2023, while the GTX 480 launched in March 2010. The 680M's predecessor is the Vega II IGP and its successor is the Navi III IGP. The GTX 480's predecessor is GeForce 200 and its successor is GeForce 500. The GTX 480 remains in an end-of-life production status, while the 680M is active. The GTX 480 carries a launch MSRP of 499 USD.

Head-to-Head Benchmarks

The only directly comparable benchmark between the two is Geekbench OpenCL. The AMD Radeon 680M scores 23,468, while the NVIDIA GeForce GTX 480 scores 13,300. The delta is 76.5% in favor of the 680M. That is not a narrow margin; it represents a generational leap in compute throughput per the recorded data.

The 680M's FP32 output of 3.379 TFLOPS is roughly 2.5x the GTX 480's 1,345.0 GFLOPS. Its texture rate of 105.6 GTexel/s is 2.5x the GTX 480's 42.06 GTexel/s. Its pixel rate of 70.40 GPixel/s is 3.3x the GTX 480's 21.03 GPixel/s. These specification deltas align with the benchmark result: the 680M is faster across every measured compute and fill-rate metric.

The GTX 480 does hold a memory bandwidth advantage in absolute terms. Its 177.4 GB/s dedicated bandwidth exceeds whatever the 680M can draw from system-shared memory, which is described only as system dependent. However, this advantage does not produce a win in the recorded OpenCL test. The 680M's 12 ray tracing cores and 768 shading units more than compensate in the compute workload captured by the database.

The 680M also records a 3DMark Steel Nomad DX12 score of 378 and a Geekbench Vulkan score of 21,965. The GTX 480 has no entries for either test, likely due to its lack of Vulkan support and its older DirectX 12 (11_0) feature level. The 680M's Vulkan score of 21,965 is close to its OpenCL score of 23,468, indicating consistent performance across different compute APIs. The GTX 480 cannot be measured in those APIs because the data does not include any result for them.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon 680M averages 15,270, while the NVIDIA GeForce GTX 480 averages 13,300. The 680M also ranks at the 57th percentile among all GPUs, versus the 53rd percentile for the GTX 480.

Q: How much faster is the Radeon 680M in Geekbench OpenCL?

A: The 680M scores 23,468 against the GTX 480's 13,300, a 76.5% advantage. This is the only head-to-head benchmark with recorded scores for both parts.

Q: Does the GTX 480 have any architectural advantages?

A: Yes. The GTX 480 has 60 TMUs versus 48, 48 ROPs versus 32, and a dedicated 1536 MB GDDR5 frame buffer with 177.4 GB/s bandwidth over a 384-bit bus. The 680M relies on system-shared memory with system-dependent bandwidth.

Q: What process nodes do the two GPUs use?

A: The Radeon 680M uses a 6 nm process at TSMC with 13,100 million transistors on a 208 mm² die. The GTX 480 uses a 40 nm process at TSMC with 3,100 million transistors on a 529 mm² die.

Q: Does the GTX 480 support Vulkan or ray tracing?

A: No. The GTX 480 has no recorded Vulkan support and no ray tracing cores. The Radeon 680M supports Vulkan 1.4 and includes 12 ray tracing cores.

Q: What is the power consumption difference?

A: The Radeon 680M has a 50 W TDP and is an integrated part with no power connectors. The GTX 480 has a 250 W TDP, requires 1x 6-pin and 1x 8-pin connectors, and carries a suggested 600 W power supply.

The Verdict

The data supports only one conclusion for performance: the AMD Radeon 680M is the faster GPU in every recorded benchmark. Its OpenCL score is 76.5% higher, its average benchmark score is 14.8% higher, and its percentile rank is four points higher. It also adds modern feature support in DirectX 12 Ultimate, Vulkan 1.4, and hardware ray tracing, none of which the GTX 480 can offer.

The GTX 480's case rests entirely on specifications that do not materialize as wins in the database. Its dedicated 177.4 GB/s memory bandwidth, higher TMU and ROP counts, and larger physical frame buffer are real differences, but the only test that compares both parts shows a decisive loss. The GTX 480 also carries a 250 W TDP, a dual-slot cooler, and a 600 W suggested power supply, while the 680M operates at 50 W with no connectors.

For anyone choosing between these two based on measured data, the Radeon 680M is the clear pick. It wins the compute benchmark, offers substantially higher pixel and texture rates, supports modern APIs, and does so at a fraction of the power draw. The GTX 480 is an end-of-life product from 2010 with a 499 USD launch MSRP and no recorded benchmark victory over the 680M.

The only scenario where the GTX 480 might be preferable is one that depends on its dedicated memory subsystem or its higher ROP count, but the database contains no test where that translates into a higher score. Until such a workload appears in the recorded data, the 680M stands as the superior part across every measurable dimension.