AMD Radeon HD 8970M vs AMD Radeon RX 590 Comparison

Head-to-Head Benchmarks

The recorded data places the AMD Radeon RX 590 and the AMD Radeon HD 8970M in very different performance tiers, despite both carrying the Radeon name. The RX 590 posts an average benchmark score of 24744, while the HD 8970M trails far behind at 21237. That is a raw gap of roughly 16.5% in the database’s aggregate metric, a margin that shows up clearly across the few directly comparable tests.

The RX 590’s strongest recorded result is in Geekbench Metal, where it scores 48407. The HD 8970M has no Metal result in the database, so the comparison must lean on the OpenCL test, where the older card scores 21237. Looking at the relative standings, the RX 590 sits at the 70th percentile of all GPUs, while the HD 8970M lands at the 66th percentile. That four-point percentile gap might sound small, but the underlying scores tell a different story: the RX 590 is 16.5% ahead in average performance, which is substantial for two products from the same vendor.

The RX 590 also records a 3DMark Steel Nomad DX12 score of 1081, a test the HD 8970M does not appear in. Given the HD 8970M’s much older GCN 1.0 architecture and its lower raw compute figures, the database suggests the RX 590 would dominate any DX12 workload. The HD 8970M’s only listed API advantage is Vulkan 1.2.170 support versus the RX 590’s Vulkan 1.3, but the newer card also supports DirectX 12_0 features, while the HD 8970M only reaches DirectX 12 (11_1). That means the RX 590 is better equipped for modern titles that rely on DX12 feature levels.

Looking at the nearest rivals in the database, the RX 590’s average score of 24744 is nearly identical to the NVIDIA RTX A5000 Mobile at 24763, a delta of just -0.1%. It also edges out the Intel Arc A350M by 0.4%, the AMD Radeon RX 6600 XT by 1.2%, and the NVIDIA GeForce GTX 1630 by 1.9%. These are tight margins, meaning the RX 590 sits right at the boundary of a cluster of mid-range GPUs from multiple generations. For the HD 8970M, its average score of 21237 puts it 0.4% ahead of the AMD Radeon RX Vega M GL, 1% ahead of the NVIDIA GeForce RTX 5050, but 0.7% behind the NVIDIA RTX A4000 Mobile and 1.8% behind the NVIDIA Quadro RTX 5000. So the HD 8970M is competitive with a few mobile and desktop parts from later eras, but it is nowhere near the RX 590’s performance class.

The biggest win for the RX 590 is in raw compute throughput. The database lists the RX 590 at 7.119 TFLOPS FP32, versus 2.304 TFLOPS for the HD 8970M. That is a 3.09x advantage in single-precision floating-point math. Pixel fill rate tells a similar story: 49.44 GPixel/s for the RX 590 against 28.80 GPixel/s for the HD 8970M, a 1.72x lead. Texture fill rate is even more lopsided: 222.5 GTexel/s versus 72.00 GTexel/s, a 3.09x advantage. These are not marginal improvements; they represent a full generational leap in architecture and execution resources.

Where Each One Wins

The RX 590 wins in every category where the database records comparable data. It has more shading units (2304 versus 1280), more texture mapping units (144 versus 80), and the same 32 ROPs. The memory subsystem also favors the newer card: 8 GB of GDDR5 on a 256-bit bus with 256.0 GB/s bandwidth, versus 4 GB of GDDR5 on the same bus width but only 153.6 GB/s. That means the RX 590 can feed its larger compute pipeline much more effectively, which matters for texture-heavy scenes and high-resolution rendering.

For gaming, the RX 590 is the clear choice. Its DX12 support at feature level 12_0, combined with Vulkan 1.3, makes it compatible with the current generation of game engines. The HD 8970M’s DX12 support is limited to feature level 11_1, which excludes certain advanced rendering techniques used in newer titles. The RX 590 also has a higher boost clock (1545 MHz versus 900 MHz) and a higher base clock (1469 MHz versus 850 MHz), so even clock-for-clock it is faster, and it has far more compute units behind those clocks.

Where the HD 8970M has a claim is in power efficiency relative to its era, though the database does not directly compare efficiency. The HD 8970M has a TDP of 100 W versus the RX 590’s 175 W, so it draws less power. But that lower power budget comes with a 3x reduction in FP32 throughput, so the efficiency ratio is not favorable to the older part. The HD 8970M also uses an MXM module form factor, which is designed for laptops, while the RX 590 is a dual-slot desktop card. So the HD 8970M’s niche is older mobile workstations or gaming laptops, not desktop performance.

The HD 8970M does have a small edge in the database’s rival comparisons: its nearest rival list includes the NVIDIA RTX A4000 Mobile and NVIDIA Quadro RTX 5000, and it is only 0.7% and 1.8% behind those two parts, respectively. That suggests the HD 8970M, despite its age, can still hold its own against some professional mobile GPUs from a few years later. But those are not the same performance tier as the RX 590’s rivals, which include the RX 6600 XT and GTX 1630.

The Verdict

The data is unambiguous: the AMD Radeon RX 590 is the superior product. It is 16.5% ahead in average benchmark score, has 3.09x the FP32 throughput, 3.09x the texture fill rate, 1.72x the pixel rate, double the memory capacity, and 1.67x the memory bandwidth. Every recorded metric that can be compared favors the RX 590. The HD 8970M’s only advantages are lower power draw (100 W versus 175 W) and a mobile-friendly MXM form factor, but those do not translate into better performance in any benchmark.

For a desktop user looking for a capable 1080p or 1440p gaming card, the RX 590 is the obvious pick. It sits at the 70th percentile of all GPUs, meaning it outperforms the majority of the database’s entries, and it is within 1.9% of the GTX 1630 and 1.2% of the RX 6600 XT, both of which are much newer products. The HD 8970M, at the 66th percentile, is not far behind in percentile terms, but its raw scores place it in a lower performance bracket.

For someone constrained to a laptop or an MXM-based system, the HD 8970M is the only option of the two, since the RX 590 is a desktop card. But the data shows that option is significantly slower. The HD 8970M’s 2.304 TFLOPS and 153.6 GB/s bandwidth will struggle with modern game settings, while the RX 590’s 7.119 TFLOPS and 256.0 GB/s are still competitive with entry-level and mid-range cards released years later.

The verdict from the recorded data: choose the RX 590 for any desktop build. Choose the HD 8970M only if the form factor forces you into a mobile platform, and even then, expect substantially lower performance in every workload.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon RX 590 scores 24744 on average, while the AMD Radeon HD 8970M scores 21237. The RX 590 is 16.5% ahead.

Q: How does the RX 590 compare to its nearest rivals?

A: The RX 590 is 0.1% behind the NVIDIA RTX A5000 Mobile, 0.4% ahead of the Intel Arc A350M, 1.2% ahead of the AMD Radeon RX 6600 XT, and 1.9% ahead of the NVIDIA GeForce GTX 1630.

Q: What is the biggest performance gap between the two cards?

A: The largest recorded gap is in FP32 compute: the RX 590 delivers 7.119 TFLOPS, while the HD 8970M delivers 2.304 TFLOPS, a 3.09x difference.

Q: Does the HD 8970M outperform any of its rivals?

A: Yes, the HD 8970M is 0.4% ahead of the AMD Radeon RX Vega M GL and 1% ahead of the NVIDIA GeForce RTX 5050, but it trails the NVIDIA RTX A4000 Mobile by 0.7% and the NVIDIA Quadro RTX 5000 by 1.8%.

Q: Which card has better memory bandwidth?

A: The RX 590 has 256.0 GB/s bandwidth, compared to 153.6 GB/s for the HD 8970M. The RX 590 also has 8 GB of memory versus 4 GB.

Q: Are both cards still in production?

A: No, both are end-of-life products. The RX 590 was released in 2018, and the HD 8970M was released in 2013.

Architecture Differences

The two GPUs come from different generations of AMD’s Graphics Core Next architecture. The RX 590 uses Polaris 30, built on GCN 4.0, while the HD 8970M uses Neptune, built on GCN 1.0. That architectural gap explains most of the performance difference.

The process node is a major factor. The RX 590 is fabricated on a 12 nm process at GlobalFoundries, with 5,700 million transistors on a 232 mm² die. The HD 8970M uses a 28 nm process at TSMC, with 2,800 million transistors on a 212 mm² die. The newer process allows the RX 590 to nearly double the transistor count while keeping the die size similar, resulting in a transistor density of 24.6M per mm² versus 13.2M per mm² for the HD 8970M.

Compute resources differ substantially. The RX 590 has 2304 shading units, 144 texture mapping units, and 32 ROPs. The HD 8970M has 1280 shading units, 80 texture mapping units, and 32 ROPs. The RX 590 also runs at higher clocks: 1469 MHz base and 1545 MHz boost, versus 850 MHz base and 900 MHz boost for the HD 8970M. Combined, those give the RX 590 a pixel rate of 49.44 GPixel/s and a texture rate of 222.5 GTexel/s, while the HD 8970M manages 28.80 GPixel/s and 72.00 GTexel/s.

Memory configuration is another divider. Both use a 256-bit bus, but the RX 590 has 8 GB of GDDR5 at 2000 MHz (8 Gbps effective) for 256.0 GB/s bandwidth. The HD 8970M has 4 GB of GDDR5 at 1200 MHz (4.8 Gbps effective) for 153.6 GB/s. The RX 590’s memory also runs faster and has more capacity, which helps with texture streaming and high-resolution assets.

API support reflects the architectural age. The RX 590 supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.3. The HD 8970M supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The RX 590 also has a higher FP32 throughput of 7.119 TFLOPS, and it supports FP16 at the same rate (7.119 TFLOPS 1:1), while the HD 8970M has no recorded FP16 capability.

Power and form factor are the final differentiators. The RX 590 has a 175 W TDP, requires a single 8-pin power connector, and is a dual-slot PCIe 3.0 x16 card with 1x HDMI 2.0b and 3x DisplayPort 1.4a outputs. The HD 8970M has a 100 W TDP, uses an MXM module form factor, and has display outputs that are portable-device dependent. The RX 590 is 241 mm long, while the HD 8970M has no recorded dimensions. The RX 590’s launch MSRP was 279 USD. Both cards are end-of-life, with the RX 590 succeeding the Arctic Islands generation and preceding Vega, while the HD 8970M succeeds London and precedes Gem System.