AMD Radeon 680M vs AMD Radeon 820M Comparison

FAQ

Q: What are the core architecture differences between the AMD Radeon 680M and the AMD Radeon 820M?

A: The Radeon 680M uses the RDNA 2.0 architecture on a 6 nm process, while the Radeon 820M uses RDNA 3.5 on a 4 nm process. The 680M has 768 shading units, 48 TMUs, and 32 ROPs, whereas the 820M has 128 shading units, 8 TMUs, and 4 ROPs.

Q: How do the clock speeds compare between the two GPUs?

A: The Radeon 680M has a base clock of 2000 MHz and a boost clock of 2200 MHz. The Radeon 820M has a much lower base clock of 400 MHz but a higher boost clock of 2800 MHz.

Q: What is the thermal design power (TDP) of each GPU?

A: The Radeon 680M has a TDP of 50 W, while the Radeon 820M has a TDP of 15 W. This makes the 820M considerably more power-efficient in terms of rated power draw.

Q: Which GPU has higher raw compute throughput?

A: The Radeon 680M delivers 3.379 TFLOPS of FP32 compute, compared to the Radeon 820M's 716.8 GFLOPS. The 680M also achieves higher pixel rate at 70.40 GPixel/s versus 11.20 GPixel/s, and higher texture rate at 105.6 GTexel/s versus 22.40 GTexel/s.

Q: What benchmark scores are available for each GPU?

A: The Radeon 680M has recorded scores of 378 in 3DMark Steel Nomad DX12, 23468 in Geekbench OpenCL, and 21965 in Geekbench Vulkan, with an average benchmark score of 15270. The Radeon 820M has no recorded benchmark scores in the database.

Q: How do the two GPUs compare in terms of ray tracing capabilities?

A: The Radeon 680M includes 12 ray tracing cores, while the Radeon 820M includes 2. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Architecture Differences

The two integrated GPUs come from different architectural generations. The Radeon 680M is built on RDNA 2.0, using the Rembrandt+ chip, and is part of the Navi II IGP generation. The Radeon 820M uses the newer RDNA 3.5 architecture, based on the Krackan Point 2 chip, and belongs to the Navi III IGP generation.

Manufacturing processes differ significantly. The 680M uses a 6 nm process at TSMC, while the 820M is fabricated on a 4 nm process at the same foundry. The 680M has a transistor count of 13,100 million on a 208 mm² die, giving a transistor density of 63.0M per mm². The 820M's transistor count and die size are listed as unknown in the database.

The execution unit configurations diverge sharply. The 680M deploys 768 shading units, 48 texture mapping units, and 32 raster output units. The 820M has only 128 shading units, 8 TMUs, and 4 ROPs. Ray tracing hardware also differs, with 12 RT cores on the 680M versus 2 on the 820M.

Clock behavior shows an interesting contrast. The 680M has a higher base clock at 2000 MHz versus 400 MHz on the 820M. However, the 820M has a higher boost clock at 2800 MHz versus 2200 MHz on the 680M. This suggests the 820M is designed to ramp up aggressively under load but idle at a very low frequency to conserve power.

The FP16 compute capability also differs. The 680M achieves 6.758 TFLOPS at a 2:1 ratio relative to FP32, while the 820M achieves 716.8 GFLOPS at a 1:1 ratio. This indicates the 680M has dedicated fast FP16 throughput, whereas the 820M does not double its FP16 rate.

Both GPUs share several features. They use system shared memory with system dependent bandwidth, have a PCIe 4.0 x8 bus interface, are IGP slot width, have no power connectors, and list display outputs as portable device dependent. Both support the same API set: DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

The release timeline shows a generational gap. The 680M was released in January 2023 and lists Vega II IGP as its predecessor and Navi III IGP as its successor. The 820M was released in February 2025 and lists Navi II IGP as its predecessor, matching that successor relationship.

Head-to-Head Benchmarks

The database contains no direct head-to-head benchmark comparisons between the Radeon 680M and the Radeon 820M. The head-to-head benchmark array is empty, and neither wins nor losses are recorded for either GPU in direct competition.

However, the Radeon 680M has three individual benchmark scores recorded. The 3DMark Steel Nomad DX12 test produced a score of 378. Geekbench OpenCL delivered 23468, and Geekbench Vulkan produced 21965. These combine to an average benchmark score of 15270.

The Radeon 820M has no benchmark scores recorded at all, with an average benchmark score of 0. This makes direct numerical comparison impossible from the recorded data. The database shows the 680M at the 57th percentile among all GPUs, while the 820M sits at the 50th percentile.

The 680M's average score places it among several discrete NVIDIA and AMD cards. The database lists the NVIDIA GeForce GTX 580 with an average score of 15283, a delta of -0.1 percent relative to the 680M. The NVIDIA GeForce RTX 2060 scores 15290, also -0.1 percent. The NVIDIA GeForce RTX 3050 OEM scores 15199, a delta of 0.5 percent. The AMD Radeon RX 7600 scores 15171, a delta of 0.7 percent.

These rival comparisons show the 680M performing essentially on par with the GTX 580 and RTX 2060, slightly behind the RTX 3050 OEM, and slightly ahead of the RX 7600. The deltas are all within one percentage point, indicating near-identical average performance among this group.

The 820M has no nearest rivals listed, which aligns with its lack of recorded benchmark data. Its percentile ranking of 50 is based on its position among all GPUs but without measured scores to substantiate a precise comparison.

The Verdict

The recorded data presents a clear hierarchy. The Radeon 680M delivers substantially higher compute throughput in every measured category. Its FP32 performance of 3.379 TFLOPS is roughly 4.7 times the 820M's 716.8 GFLOPS. Its texture rate of 105.6 GTexel/s is about 4.7 times the 820M's 22.40 GTexel/s. Its pixel rate of 70.40 GPixel/s is approximately 6.3 times the 820M's 11.20 GPixel/s.

The 680M also carries a much larger execution resource pool with 768 shading units, 48 TMUs, and 32 ROPs versus 128, 8, and 4 respectively. It has six times the ray tracing cores at 12 versus 2.

The 820M counters with a more modern architecture, RDNA 3.5 versus RDNA 2.0, and a more efficient process node at 4 nm versus 6 nm. Its TDP of 15 W is one-third of the 680M's 50 W. Its boost clock of 2800 MHz exceeds the 680M's 2200 MHz, though its base clock of 400 MHz is far lower.

For workloads that demand raw graphics throughput, the data points squarely to the Radeon 680M. Its higher shading unit count, TMU count, ROP count, and compute rates give it a decisive advantage in fill-rate-limited and shader-heavy scenarios. The 680M also has recorded benchmark evidence of competitive performance against discrete GPUs from the GTX 580 and RTX 2060 class.

For workloads that prioritize power efficiency and the newer architectural feature set, the Radeon 820M holds appeal. Its 15 W TDP makes it suited to the most power-constrained portable designs. The RDNA 3.5 architecture brings the newer instruction set and feature support, though the database does not list specific architectural feature differences beyond the API set, which is identical.

The absence of benchmark scores for the 820M means its real-world performance cannot be quantified from the database. The verdict must therefore rest on the architectural specifications and the 680M's measured results.

Specification Differences

The following fields differ between the two GPUs:

• Chip: Rembrandt+ (680M) versus Krackan Point 2 (820M)
• Architecture: RDNA 2.0 (680M) versus RDNA 3.5 (820M)
• Generation: Navi II IGP (Rembrandt Mobile) (680M) versus Navi III IGP (Strix Point Mobile) (820M)
• Process node: 6 nm (680M) versus 4 nm (820M)
• Transistors: 13,100 million (680M) versus unknown (820M)
• Die size: 208 mm² (680M) versus unknown (820M)
• Transistor density: 63.0M / mm² (680M) versus null (820M)
• Base clock: 2000 MHz (680M) versus 400 MHz (820M)
• Boost clock: 2200 MHz (680M) versus 2800 MHz (820M)
• Shading units: 768 (680M) versus 128 (820M)
• TMUs: 48 (680M) versus 8 (820M)
• ROPs: 32 (680M) versus 4 (820M)
• RT cores: 12 (680M) versus 2 (820M)
• Pixel rate: 70.40 GPixel/s (680M) versus 11.20 GPixel/s (820M)
• Texture rate: 105.6 GTexel/s (680M) versus 22.40 GTexel/s (820M)
• FP32: 3.379 TFLOPS (680M) versus 716.8 GFLOPS (820M)
• FP16: 6.758 TFLOPS (2:1) (680M) versus 716.8 GFLOPS (1:1) (820M)
• TDP: 50 W (680M) versus 15 W (820M)
• Release date: January 2023 (680M) versus February 2025 (820M)
• Predecessor: Vega II IGP (680M) versus Navi II IGP (820M)
• Successor: Navi III IGP (680M) versus null (820M)
• Average benchmark score: 15270 (680M) versus 0 (820M)
• Percentile vs all GPUs: 57 (680M) versus 50 (820M)

Fields that are identical include the manufacturer (AMD), foundry (TSMC), memory configuration (system shared), bus interface (PCIe 4.0 x8), slot width (IGP), power connectors (none), display outputs (portable device dependent), API set (DirectX 12 Ultimate, OpenGL 4.6, Vulkan 1.4), production status (active), and launch MSRP (not listed for either).

Where Each One Wins

The Radeon 680M wins on raw graphics throughput. The recorded data shows it at 4.7 times the FP32 compute, 4.7 times the texture fill rate, and 6.3 times the pixel fill rate of the 820M. Its 768 shading units provide six times the parallel shader capacity. For any workload that is compute-bound or fill-rate-bound, the 680M dominates by specification.

The Radeon 680M wins on ray tracing resources. With 12 RT cores versus 2, it has six times the dedicated ray tracing hardware. The API support is identical, so the difference comes down to hardware capacity.

The Radeon 680M wins on measured evidence. Its average benchmark score of 15270 places it in the 57th percentile of all GPUs, with direct comparisons to discrete cards like the GTX 580 and RTX 2060 within one percentage point. The 820M has no recorded scores.

The Radeon 820M wins on power efficiency. Its 15 W TDP is 30 percent of the 680M's 50 W. For thermally constrained laptops and fanless designs, this is a substantial advantage.

The Radeon 820M wins on process technology. The 4 nm node is newer than the 6 nm node, which typically allows higher transistor density and lower leakage per transistor, though the 820M's transistor count is unknown.

The Radeon 820M wins on peak boost clock. Its 2800 MHz boost exceeds the 680M's 2200 MHz. When the 820M does ramp up to its maximum frequency, its per-clock efficiency from RDNA 3.5 can partially offset its smaller execution resource pool.

The workload split is therefore clear. Graphics-intensive applications, gaming, 3D rendering, and any GPU-accelerated compute that leverages high shader counts or high fill rates favor the 680M. Lightweight desktop compositing, video playback, office productivity, and battery-conscious mobile use favor the 820M due to its dramatically lower power draw. The 680M targets performance-oriented thin-and-light laptops, while the 820M targets ultra-portable and entry-level devices where the 15 W TDP is the primary constraint.