AMD Radeon R9 M390X vs Intel Arc B570 Comparison

The AMD Radeon R9 M390X and Intel Arc B570 represent two very different eras of GPU design, separated by nearly a decade of architectural evolution. The data shows that while the R9 M390X is a legacy 28nm mobile part from the GCN 3.0 generation, the Arc B570 is a modern 5nm desktop GPU from Intel’s Battlemage lineup. Their average benchmark scores are remarkably close—the R9 M390X posts 20,662 against the Arc B570’s 20,556—but their underlying capabilities and feature sets diverge sharply, making the choice between them a matter of workload rather than raw aggregate performance.

FAQ

Q: How do the average benchmark scores of the two GPUs compare?

A: The AMD Radeon R9 M390X has an average benchmark score of 20,662, while the Intel Arc B570 scores 20,556. This puts the R9 M390X just 0.5% ahead of the Arc B570.

Q: Which GPU wins in the Vulkan benchmark?

A: The Intel Arc B570 wins decisively in Geekbench Vulkan, scoring 96,844 compared to the R9 M390X’s 16,951. The deltaPct is -82.5%, meaning the R9 M390X trails by over 82% in this specific test.

Q: What is the difference in memory capacity and type?

A: The R9 M390X comes with 4 GB of GDDR5 memory on a 256-bit bus, delivering 160.0 GB/s bandwidth. The Arc B570 offers 10 GB of GDDR6 on a 160-bit bus, with 380.0 GB/s bandwidth—more than double the bandwidth.

Q: How do their process nodes and transistor counts differ?

A: The R9 M390X uses a 28 nm process from TSMC with 5,000 million transistors on a 366 mm² die. The Arc B570 is built on a 5 nm TSMC process, packing 19,600 million transistors into a smaller 272 mm² die.

Q: Which GPU has higher pixel and texture fill rates?

A: The Arc B570 is far ahead, with a pixel rate of 200.0 GPixel/s versus 23.14 GPixel/s for the R9 M390X. Its texture rate is 360.0 GTexel/s compared to 92.54 GTexel/s.

Q: What API feature levels does each support?

A: The R9 M390X supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.2.170. The Arc B570 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, along with 18 ray tracing cores.

Where Each One Wins

The benchmark data shows a single head-to-head result, and it belongs entirely to the Intel Arc B570. In the Geekbench Vulkan test, the Arc B570 scores 96,844 against 16,951 for the R9 M390X, a margin of 82.5%. This is not a close contest; it reflects a fundamental gap in modern API efficiency and driver maturity. The Arc B570’s Vulkan support (version 1.4) pairs with its Xe2-HPG architecture to deliver massive performance in compute-heavy, low-level graphics workloads. The R9 M390X, with its older GCN 3.0 design and Vulkan 1.2.170, simply cannot keep pace in this scenario.

However, the aggregate averages tell a different story. The R9 M390X edges ahead in average score (20,662 vs 20,556), suggesting that in older or less demanding workloads—where GCN’s raw shader throughput and 256-bit memory bus still matter—it remains competitive. The R9 M390X’s nearest rival list includes the AMD Radeon RX 5600 XT (deltaPct -0.2) and Intel Arc A750 (deltaPct 0.4), while the Arc B570 sits just 0.5% behind. This implies the R9 M390X holds its own in legacy titles or DirectX 11-era applications, where its 2048 shading units and 128 TMUs are still relevant. The Arc B570, by contrast, is built for the future: DirectX 12 Ultimate, ray tracing, and Vulkan 1.4 are its home turf.

For users prioritizing modern features and forward-looking API support, the Arc B570 is the clear winner. For those locked into older software stacks or seeking a drop-in upgrade for a legacy system, the R9 M390X’s aggregate score shows it can still deliver comparable average performance, though without the same headroom.

Architecture Differences

The R9 M390X is built on AMD’s GCN 3.0 architecture, codenamed Amethyst, and belongs to the Gem System (R9 M300) generation. It is fabricated on a 28 nm process at TSMC, with 5,000 million transistors on a 366 mm² die. This results in a transistor density of 13.7M per mm². The chip is a mobile-focused MXM Module, designed for laptops and portable devices, with display outputs that are listed as "Portable Device Dependent." Its compute capabilities are straightforward: 2048 shading units, 128 TMUs, and 32 ROPs, with FP32 performance rated at 2.961 TFLOPS and FP16 at the same 2.961 TFLOPS (1:1 ratio).

The Intel Arc B570, in contrast, uses the Xe2-HPG architecture on the BMG-G21 chip, part of the Battlemage (Arc 5) generation. It is built on a 5 nm TSMC process with 19,600 million transistors on a 272 mm² die, yielding a density of 72.1M per mm²—over five times denser than the R9 M390X. The Arc B570 packs 2304 shading units, 144 TMUs, 80 ROPs, and 18 dedicated ray tracing cores. Its FP32 throughput is 11.52 TFLOPS, while FP16 reaches 23.04 TFLOPS (2:1 ratio), reflecting a modern architecture optimized for parallel compute and accelerated graphics features.

The generational gap is stark. The R9 M390X supports DirectX 12 (12_0) and Vulkan 1.2.170, while the Arc B570 supports DirectX 12 Ultimate (12_2) and Vulkan 1.4. The Arc B570 also includes hardware ray tracing cores, a feature entirely absent from the R9 M390X. Additionally, the Arc B570’s memory subsystem uses GDDR6 with 380.0 GB/s bandwidth, while the R9 M390X relies on GDDR5 with 160.0 GB/s—a 2.4x difference in memory throughput.

Specification Differences

The two GPUs differ across nearly every major specification, reflecting their different target markets and eras. The R9 M390X has a TDP of 75 W and uses no power connectors, typical for a mobile MXM module. The Arc B570 has a TDP of 150 W, requires a 1x 8-pin power connector, and suggests a 450 W PSU. The bus interface also differs: the R9 M390X uses PCIe 3.0 x16, while the Arc B570 uses PCIe 4.0 x8.

Memory configurations are a major divergence. The R9 M390X offers 4 GB of GDDR5 on a 256-bit bus, with 160.0 GB/s bandwidth. The Arc B570 offers 10 GB of GDDR6 on a 160-bit bus, with 380.0 GB/s bandwidth—more than double the bandwidth despite a narrower bus, thanks to faster memory clocks (2375 MHz / 19 Gbps effective vs 1250 MHz / 5 Gbps effective).

Clock speeds also differ: the R9 M390X lists no base or boost clock, only a memory clock, while the Arc B570 has a base and boost clock of 2500 MHz. The Arc B570’s pixel rate (200.0 GPixel/s) and texture rate (360.0 GTexel/s) dwarf the R9 M390X’s (23.14 GPixel/s and 92.54 GTexel/s). The Arc B570 is a dual-slot card measuring 272 mm in length and 115 mm in height, with outputs of 1x HDMI 2.1a and 3x DisplayPort 2.1. The R9 M390X has no listed dimensions and its display outputs are dependent on the portable device.

Finally, production status and release dates separate them: the R9 M390X was released on 2015-05-04 and is end-of-life, while the Arc B570 was released on 2025-01-15 and remains active. The Arc B570 has a launch MSRP of 219 USD.

Head-to-Head Benchmarks

The only direct head-to-head benchmark in the data is the Geekbench Vulkan test, and it is a landslide. The Intel Arc B570 scores 96,844, while the AMD Radeon R9 M390X scores 16,951. The deltaPct of -82.5% indicates the R9 M390X loses by 82.5%, meaning the Arc B570 delivers roughly 5.7 times the Vulkan performance. This is the single largest gap between the two, and it highlights the Arc B570’s modern driver stack and Vulkan 1.4 support, which enable far better utilization of its 2304 shading units and 18 ray tracing cores.

While the R9 M390X has no individual benchmark wins in the head-to-head data, its aggregate performance tells a more nuanced story. Its average score of 20,662 is 0.5% higher than the Arc B570’s 20,556, and its nearest rivals include the RX 5600 XT (-0.2%) and Arc A750 (+0.4%). This suggests that in Geekbench Metal or other unspecified workloads, the R9 M390X can match or exceed the Arc B570. The R9 M390X’s benchmarks include a Geekbench Metal score of 24,373 and a Geekbench Vulkan score of 16,951, the latter being the same test where it loses to the Arc B570.

The Arc B570’s broader benchmark suite—including 3DMark Steel Nomad DX12 (2,649), Geekbench OpenCL (83,514), and Passmark G3D (14,195)—shows consistent strength across modern APIs. Its Passmark DirectX 9 score of 164 and DirectX 11 score of 118 indicate solid legacy performance, though not stellar. The R9 M390X, with only two benchmarks listed, appears to excel in Metal-based workloads, which are common in macOS environments, while the Arc B570 dominates Vulkan and OpenCL.

The Verdict

The data points to a clear split. The Intel Arc B570 is the superior GPU for modern workloads, leading by 82.5% in the only direct head-to-head test (Geekbench Vulkan). Its 5 nm process, 19,600 million transistors, 18 ray tracing cores, and DirectX 12 Ultimate support make it the obvious choice for users running current games, using Vulkan APIs, or needing high-bandwidth GDDR6 memory (380.0 GB/s). Its dual-slot design, 150 W TDP, and PCIe 4.0 x8 interface are standard for a desktop card, and its active production status ensures ongoing support.

The AMD Radeon R9 M390X, while end-of-life and limited to a 28 nm mobile design, still posts a marginally higher average benchmark score (20,662 vs 20,556). Its 256-bit memory bus and 75 W TDP make it a low-power, legacy-friendly option for portable devices, and its Geekbench Metal score of 24,373 suggests it remains viable in Apple-centric environments. However, its 4 GB GDDR5 memory and lack of ray tracing cores are severe limitations for anything beyond older or lighter workloads.

Choose the Arc B570 if you need modern API support, high memory bandwidth, and ray tracing, or if you are building a new desktop system. Choose the R9 M390X only if you are constrained to a legacy mobile platform, require minimal power draw, or have workloads that specifically benefit from Metal performance. The aggregate scores are close, but the architecture gap makes the Arc B570 the forward-looking pick.