AMD Radeon RX 470 vs AMD Radeon RX 6800M Comparison

AMD Radeon RX 470 and AMD Radeon RX 6800M occupy the same 74th percentile among all GPUs, yet their aggregate benchmark scores are nearly identical — 28,996 versus 28,874, a 0.4% difference in favor of the RX 470. That statistical tie, however, masks a decisive generational split: the RX 6800M wins all four head-to-head tests by margins exceeding 57%, while the RX 470’s overall average is buoyed by a narrower benchmark suite. The data presents a rare case where aggregate standings suggest equivalence, but individual workload results tell a story of complete dominance by the newer mobile part.

Head-to-Head Benchmarks

The RX 6800M outperforms the RX 470 in every directly comparable test, with the smallest margin still representing a landslide. In the 3DMark Steel Nomad DX12 test, the RX 6800M scores 2,238 against the RX 470’s 842, a delta of -62.4% from the RX 470’s perspective — meaning the RX 6800M delivers roughly 2.7 times the performance. This is the least lopsided of the four matchups, yet it still places the RX 470 firmly in a lower performance tier.

Geekbench results widen the gap further. In the Metal API test, the RX 6800M posts 113,721 versus 41,690, a -63.3% delta. OpenCL shows a similar pattern: 87,621 for the RX 6800M against 33,568 for the RX 470, a -61.7% difference. The Vulkan test is the closest contest, with the RX 6800M scoring 94,766 compared to the RX 470’s 39,884 — still a -57.9% delta, meaning the RX 6800M is more than 2.3 times faster. Across all four tests, the RX 6800M’s advantage ranges from roughly 2.4x to 2.7x, with no single benchmark where the RX 470 closes the gap to under double.

The absence of any RX 470 win is notable. With zero wins out of four head-to-head tests, the RX 470’s aggregate parity with the RX 6800M is entirely an artifact of its benchmark selection. The RX 470’s suite includes only 3DMark Steel Nomad and three Geekbench compute tests, while the RX 6800M’s 11-test suite adds PassMark DirectX 9/10/11/12, G2D, G3D, and GPU Compute results. Those additional PassMark scores — ranging from 65 in DirectX 12 to 147 in DirectX 9, with a G3D score of 13,261 — drag the RX 6800M’s average down toward the RX 470’s level, even though the RX 6800M wins every shared test by a wide margin.

The Verdict

From the data alone, the RX 6800M is the superior GPU for any workload represented in the head-to-head tests. It wins all four shared benchmarks with deltas between -57.9% and -63.3%, and its raw compute throughput — 12.24 TFLOPS FP32 versus 4.940 TFLOPS — is roughly 2.5 times higher. The RX 470’s aggregate score parity is misleading; it stems from the RX 6800M’s inclusion of legacy DirectX 9/10/11 tests where it scores low (147, 101, and 127 respectively) relative to its modern API results. A buyer choosing between these two should select the RX 6800M if any of the four compared workloads matter.

Still, the RX 470 is not without a case. Its average benchmark score of 28,996 is higher than the RX 6800M’s 28,874, and it sits ahead of the RX 6800M in the nearestRivals comparison by 0.4%. The RX 470 also matches the RX 6800M in the 74th percentile ranking. For users whose applications resemble the RX 470’s test suite — primarily Geekbench compute and 3DMark Steel Nomad — the performance gap is enormous, but the aggregate metrics suggest that in a broader, unmeasured context, the two GPUs land in the same overall performance bracket. The RX 470’s dual-slot, 120 W desktop design with a 1x 6-pin connector also contrasts sharply with the RX 6800M’s integrated (IGP) mobile form factor, which draws 145 W and requires no power connectors.

The verdict is workload-dependent. For modern DX12 and Vulkan compute, the RX 6800M is the only rational choice. For legacy DirectX 9-11 titles, the RX 6800M’s PassMark scores (101-147) are low in absolute terms, but no RX 470 PassMark data exists for comparison — so the RX 470 cannot be declared superior there. The data simply shows the RX 6800M wins every shared test, and the RX 470’s aggregate edge is a statistical artifact.

Architecture Differences

The two GPUs represent distinct architectural generations. The RX 470 uses the Ellesmere chip built on GCN 4.0, fabricated on GlobalFoundries’ 14 nm process with 5,700 million transistors on a 232 mm² die, yielding a transistor density of 24.6M per mm². The RX 6800M uses Navi 22 based on RDNA 2.0, manufactured by TSMC on 7 nm, packing 17,200 million transistors into a 335 mm² die — a density of 51.3M per mm², more than double the RX 470’s.

Compute resources differ substantially. The RX 470 has 2,048 shading units, 128 TMUs, and 32 ROPs. The RX 6800M scales up to 2,560 shading units, 160 TMUs, and 64 ROPs — 25% more shaders, 25% more texture units, and double the render output units. The RX 6800M also adds 40 ray tracing cores, a feature entirely absent from the RX 470. Clock speeds reflect the newer process: the RX 6800M runs at 2,116 MHz base and 2,390 MHz boost (2,300 MHz game clock), versus the RX 470’s 926 MHz base and 1,206 MHz boost. This combination of higher clocks and more units yields 12.24 TFLOPS FP32 for the RX 6800M against 4.940 TFLOPS for the RX 470. The RX 6800M also offers 24.47 TFLOPS FP16 (2:1 ratio) while the RX 470 is capped at 4.940 TFLOPS FP16 (1:1).

Memory systems are equally divergent. The RX 470 uses 4 GB of GDDR5 on a 256-bit bus, delivering 211.2 GB/s bandwidth at 1,650 MHz (6.6 Gbps effective). The RX 6800M moves to 12 GB of GDDR6 on a 192-bit bus, with 384.0 GB/s bandwidth at 2,000 MHz (16 Gbps effective). The RX 6800M’s bandwidth is 82% higher despite a narrower bus, thanks to faster memory. Pixel and texture rates follow: the RX 6800M achieves 153.0 GPixel/s and 382.4 GTexel/s, versus 38.59 GPixel/s and 154.4 GTexel/s for the RX 470.

API support also differs. The RX 6800M supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the RX 470 is limited to DirectX 12 (12_0) and Vulkan 1.3. Both support OpenGL 4.6. The RX 6800M uses PCIe 4.0 x16 versus the RX 470’s PCIe 3.0 x16. Power profiles are reversed in form factor: the RX 470 is a dual-slot desktop card with a 120 W TDP and 1x 6-pin connector, while the RX 6800M is an integrated mobile GPU (IGP) with a 145 W TDP and no power connectors. The RX 470 measures 240 mm by 95 mm by 35 mm; the RX 6800M has no listed dimensions.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon RX 470, with an average benchmark score of 28,996 versus the RX 6800M’s 28,874 — a 0.4% difference in the RX 470’s favor.

Q: Does the RX 6800M win every directly compared test?

A: Yes. In the four head-to-head benchmarks (3DMark Steel Nomad DX12, Geekbench Metal, OpenCL, and Vulkan), the RX 6800M wins all four, with deltas ranging from -57.9% to -63.3% relative to the RX 470.

Q: What is the largest performance gap between the two?

A: The Geekbench Metal test shows the biggest delta at -63.3%, where the RX 6800M scores 113,721 versus the RX 470’s 41,690.

Q: How do their memory configurations differ?

A: The RX 470 has 4 GB of GDDR5 on a 256-bit bus with 211.2 GB/s bandwidth. The RX 6800M has 12 GB of GDDR6 on a 192-bit bus with 384.0 GB/s bandwidth.

Q: Do both GPUs support the same DirectX version?

A: No. The RX 6800M supports DirectX 12 Ultimate (12_2), while the RX 470 supports DirectX 12 (12_0). The RX 6800M also supports Vulkan 1.4 versus the RX 470’s Vulkan 1.3.

Q: What is the transistor count difference?

A: The RX 6800M contains 17,200 million transistors on a 335 mm² die, while the RX 470 contains 5,700 million transistors on a 232 mm² die — a threefold difference in transistor count.

Where Each One Wins

The RX 6800M wins every benchmark where both GPUs were tested. In 3DMark Steel Nomad DX12, it scores 2,238 versus 842, a 62.4% advantage. In Geekbench Metal, it scores 113,721 versus 41,690, a 63.3% advantage. In OpenCL, it scores 87,621 versus 33,568, a 61.7% advantage. In Vulkan, it scores 94,766 versus 39,884, a 57.9% advantage. These results make the RX 6800M the clear choice for any modern DX12 workload, compute-heavy applications using Metal, OpenCL, or Vulkan, and any scenario where ray tracing is required — the RX 6800M’s 40 ray tracing cores have no counterpart in the RX 470.

The RX 470’s wins are limited to aggregate statistics. It holds a higher average benchmark score (28,996 versus 28,874) and a 0.4% deltaPct advantage over the RX 6800M in the nearestRivals comparison. Its benchmark suite — which excludes PassMark tests — does not include any test where it beats the RX 6800M. For legacy DirectX 9/10/11 workloads, the RX 6800M’s PassMark scores (147, 101, and 127 respectively) are available, but no equivalent RX 470 scores exist, so no direct comparison can be made. The RX 470’s only clear advantage is its form factor: a dual-slot desktop card with a 120 W TDP and 1x 6-pin power connector, versus the RX 6800M’s integrated mobile design at 145 W with no power connectors — though this is a physical difference, not a performance win.

In practical terms, the RX 6800M is the winner for all four tested metrics, with margins that range from 2.4x to 2.7x. The RX 470’s aggregate parity is a function of its smaller, more modern-focused test set. Users prioritizing the exact benchmarks in this comparison should choose the RX 6800M without hesitation. Users weighing the aggregate average score might consider the RX 470, but that decision would ignore the RX 6800M’s complete dominance in every shared test.