AMD Radeon HD 8970M vs AMD Radeon RX 5700 Comparison

The AMD Radeon RX 5700 and AMD Radeon HD 8970M represent two distinct eras of GPU design, and the benchmark data reflects a massive gulf in performance. The RX 5700, a desktop part from the RDNA 1.0 generation, utterly dominates the older mobile HD 8970M in every measurable way. The only shared benchmark, Geekbench OpenCL, shows the RX 5700 scoring 69,603 against the HD 8970M’s 21,237, a delta of 227.7%. This is not a close contest; it is a generational wipeout.

Head-to-Head Benchmarks

The single direct comparison available in the data is Geekbench OpenCL, and the result is decisive. The RX 5700 achieves a score of 69,603, while the HD 8970M manages only 21,237. This translates to a 227.7% advantage for the RX 5700, meaning it delivers more than three times the compute performance in this particular workload. For context, the RX 5700’s average benchmark score across all tests is 22,170, while the HD 8970M’s average is 21,237—the RX 5700’s single OpenCL result alone is over three times the HD 8970M’s entire average.

Looking at the broader benchmark suite for the RX 5700, its strengths are evident across multiple APIs. In Passmark DirectX 9, it scores 209, which is more than double its DirectX 11 score of 101 and roughly four times its DirectX 12 score of 54. This suggests the card handles legacy titles exceptionally well. Its Passmark G3D score of 14,314 and GPU Compute score of 6,517 further cement its position as a capable all-rounder. The HD 8970M, by contrast, has no other benchmark results in the pack, leaving its OpenCL score as the sole data point. That score places it in the 66th percentile of all GPUs, while the RX 5700 sits in the 67th percentile—a surprisingly close overall ranking, though the raw numbers tell a different story.

The RX 5700’s nearest rivals in the database include the RX 6700S (average score 22,154, delta 0.1%), the RTX 5060 Mobile (22,435, delta -1.2%), and the GTX 1060 6 GB (21,856, delta 1.4%). It essentially trades blows with these cards, sitting within a 1.7% margin of the RX 7700 XT (22,549, delta -1.7%). The HD 8970M’s nearest rivals are the RX Vega M GL (21,153, delta 0.4%), RTX A4000 Mobile (21,379, delta -0.7%), and RTX 5050 (21,035, delta 1%). Both cards are statistically tied with their immediate competitors, but the performance tier is entirely different.

Where Each One Wins

The RX 5700 wins in every category where data exists. Its architecture is built for modern workloads, with support for DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, whereas the HD 8970M only reaches DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The newer card also has a substantial edge in raw throughput: 7.949 TFLOPS FP32 versus 2.304 TFLOPS, and a pixel rate of 110.4 GPixel/s versus 28.80 GPixel/s. Texture rate is similarly lopsided at 248.4 GTexel/s versus 72.00 GTexel/s.

For gaming, the RX 5700 is the clear choice. Its Passmark scores show strong performance across DirectX 9, 10, and 11, with the DirectX 9 score of 209 being particularly notable. The HD 8970M has no such data, but given its 2013 release date and GCN 1.0 architecture, it is reasonable to infer it would struggle with modern titles. The RX 5700’s 8 GB of GDDR6 memory with 448.0 GB/s bandwidth dwarfs the HD 8970M’s 4 GB of GDDR5 at 153.6 GB/s, meaning the newer card can handle higher resolutions and larger textures without bottlenecking.

The HD 8970M’s only potential advantage is its form factor. As an MXM module with a 100 W TDP, it is designed for laptops, whereas the RX 5700 is a dual-slot desktop card with a 180 W TDP. For a mobile user, the HD 8970M is the only option in this comparison, but that is a niche use case. In pure performance, the RX 5700 wins every benchmark it appears in, and the HD 8970M has zero wins in the head-to-head data.

Architecture Differences

The two GPUs are separated by six years of architectural evolution. The RX 5700 uses the Navi 10 chip built on TSMC’s 7 nm process, packing 10,300 million transistors into a 251 mm² die. The HD 8970M uses the Neptune chip on a 28 nm process, with 2,800 million transistors on a 212 mm² die. The transistor density tells the story: 41.0M / mm² for the RX 5700 versus 13.2M / mm² for the HD 8970M. This is a 3x density improvement, enabling far more compute units in the same physical space.

The RX 5700’s RDNA 1.0 architecture is a major departure from the HD 8970M’s GCN 1.0. RDNA is designed for higher IPC and better power efficiency, which is reflected in the clock speeds: the RX 5700 boosts to 1725 MHz, while the HD 8970M peaks at 900 MHz. The RX 5700 also has significantly more execution resources—2,304 shading units, 144 texture mapping units, and 64 ROPs—versus the HD 8970M’s 1,280 shading units, 80 TMUs, and 32 ROPs.

Memory technology is another differentiator. The RX 5700 uses 8 GB of GDDR6 at 14 Gbps effective, yielding 448.0 GB/s of bandwidth. The HD 8970M uses 4 GB of GDDR5 at 4.8 Gbps effective, for 153.6 GB/s. Both use a 256-bit bus, but the newer memory standard gives the RX 5700 nearly triple the bandwidth. The RX 5700 also supports PCIe 4.0 x16, while the HD 8970M is limited to PCIe 3.0 x16. Neither card has ray tracing or tensor cores, as both predate those features in AMD’s lineup.

Power consumption scales with performance: the RX 5700 draws 180 W and requires a 450 W PSU, while the HD 8970M draws 100 W and has no specified PSU requirement. The RX 5700’s power connectors are 1x 6-pin plus 1x 8-pin, whereas the HD 8970M uses an MXM module interface. The RX 5700 also has a defined physical footprint—268 mm length, 111 mm height, 36 mm width—while the HD 8970M’s dimensions are listed as null, being dependent on the host laptop.

The Verdict

The data is unambiguous: the AMD Radeon RX 5700 is the superior GPU by every metric available. It delivers 227.7% higher OpenCL performance, has 3.4x the shading units, 4.5x the ROPs, and 3.4x the FP32 throughput. Its memory bandwidth is 2.9x higher, and it supports newer API versions. The RX 5700 is a desktop card intended for high-refresh 1080p or 1440p gaming, and its Passmark scores confirm it handles both old and new titles well. The HD 8970M, by contrast, is a 2013 mobile part that is now thoroughly obsolete for anything beyond basic tasks or very old games.

If you are building a desktop system, the RX 5700 is the only sensible choice, provided your case can accommodate a 268 mm dual-slot card and your PSU can handle 180 W. Its launch MSRP was 349 USD. The HD 8970M should only be considered if you have a specific laptop that requires an MXM module and you cannot replace the entire machine. In that scenario, it remains a functional GPU, but its 66th percentile ranking means it sits at the bottom of modern performance charts. For anyone choosing between these two, the RX 5700 wins on performance, features, and longevity. The HD 8970M wins only on portability, and even that is a stretch given its age.

FAQ

Q: How much faster is the RX 5700 than the HD 8970M in OpenCL?

A: The RX 5700 scores 69,603 in Geekbench OpenCL, while the HD 8970M scores 21,237. This is a 227.7% advantage for the RX 5700.

Q: Which GPU has more shading units?

A: The RX 5700 has 2,304 shading units, compared to the HD 8970M’s 1,280. This is a 1.8x difference in raw shader count.

Q: Does the HD 8970M support ray tracing?

A: No. Neither GPU has ray tracing cores; both list null for RT cores in their specifications.

Q: What is the memory bandwidth of each card?

A: The RX 5700 has 448.0 GB/s of bandwidth from 8 GB of GDDR6, while the HD 8970M has 153.6 GB/s from 4 GB of GDDR5. The RX 5700’s bandwidth is 2.9x higher.

Q: Which card is more power-efficient?

A: The HD 8970M draws 100 W, while the RX 5700 draws 180 W. However, the RX 5700 delivers 227.7% more OpenCL performance, making it far more efficient per watt.

Q: Can the RX 5700 run older DirectX titles?

A: Yes. Its Passmark DirectX 9 score is 209, which is higher than its DirectX 11 score of 101 and DirectX 12 score of 54, indicating strong legacy support.

Specification Differences

| Specification | AMD Radeon RX 5700 | AMD Radeon HD 8970M |

|---|---|---|

| Architecture | RDNA 1.0 | GCN 1.0 |

| Process Node | 7 nm | 28 nm |

| Transistors | 10,300 million | 2,800 million |

| Die Size | 251 mm² | 212 mm² |

| Base Clock | 1465 MHz | 850 MHz |

| Boost Clock | 1725 MHz | 900 MHz |

| Memory Size | 8 GB | 4 GB |

| Memory Type | GDDR6 | GDDR5 |

| Memory Bus | 256 bit | 256 bit |

| Bandwidth | 448.0 GB/s | 153.6 GB/s |

| Shading Units | 2304 | 1280 |

| TMUs | 144 | 80 |

| ROPs | 64 | 32 |

| Pixel Rate | 110.4 GPixel/s | 28.80 GPixel/s |

| Texture Rate | 248.4 GTexel/s | 72.00 GTexel/s |

| FP32 | 7.949 TFLOPS | 2.304 TFLOPS |

| TDP | 180 W | 100 W |

| Bus Interface | PCIe 4.0 x16 | PCIe 3.0 x16 |

| DirectX | 12 (12_1) | 12 (11_1) |

| Vulkan | 1.4 | 1.2.170 |

| Slot Width | Dual-slot | MXM Module |

| Power Connectors | 1x 6-pin + 1x 8-pin | None listed |

| Suggested PSU | 450 W | None listed |

| Release Date | 2019-07-06 | 2013-05-13 |