GPU Comparison

NVIDIA RTX A4000 and AMD Radeon RX 590 represent two very different eras and purposes in GPU design. The RTX A4000 is a workstation-focused Ampere part built on an 8 nm Samsung process, while the RX 590 is a consumer Polaris card from AMD on GlobalFoundries' 12 nm node. The data shows a clear performance hierarchy, but the pairing is not without nuance, as each card has its own strengths in specific workloads. This analysis walks through the benchmark results, architectural differences, and specification gaps to determine where each card fits best.

Where Each One Wins

The benchmark data overwhelmingly favors the NVIDIA RTX A4000 in compute and modern API workloads. In the sole head-to-head benchmark available, 3DMark Steel Nomad DX12, the RTX A4000 scores 2604 against the RX 590’s 1081, a 140.9% delta. This is a decisive win for the Ampere card, indicating that in DirectX 12 gaming or workstation tasks, the RTX A4000 delivers more than double the performance of the RX 590. The RTX A4000 also holds a higher average benchmark score of 26683 compared to 24744 for the RX 590, placing it in the 72nd percentile of all GPUs versus the RX 590’s 70th percentile.

Where the RX 590 can claim a foothold is in legacy API performance. The data shows the RX 590 has a PassMark DirectX 10 score of 126, which is higher than the RTX A4000’s score of 72 in that same test. Similarly, in PassMark DirectX 9, the RX 590 scores 240 against the RTX A4000’s 240, showing parity in that older API. This suggests that for older games or applications built on DirectX 9 and 10, the RX 590 can hold its own or even outperform the newer card, which appears optimized for more recent API versions.

For compute workloads, the RTX A4000 dominates with a PassMark GPU Compute score of 9760, while the RX 590 does not have a comparable score listed in the data. The RTX A4000 also leads in Geekbench OpenCL with 105739 and Vulkan with 127645, whereas the RX 590 only has a Geekbench Metal score of 48407. This indicates the RTX A4000 is the superior choice for OpenCL and Vulkan compute, while the RX 590’s only listed compute result is in Apple’s Metal API, where it scores lower than the RTX A4000’s Vulkan score but is not directly comparable.

Architecture Differences

The architectural gap between these two GPUs is substantial. The NVIDIA RTX A4000 is built on the Ampere architecture, using the GA104 chip, and is fabricated on Samsung’s 8 nm process. It contains 17,400 million transistors on a 392 mm² die, yielding a transistor density of 44.4M per mm². In contrast, the AMD Radeon RX 590 uses the older GCN 4.0 architecture with the Polaris 30 chip, made on GlobalFoundries’ 12 nm process. It has just 5,700 million transistors on a 232 mm² die, with a density of 24.6M per mm². This means the RTX A4000 has over three times the transistor count in a larger but more densely packed die.

The RTX A4000 features 6144 shading units, 192 texture mapping units, and 96 render output units. It also includes 48 ray tracing cores and 192 tensor cores, which are absent on the RX 590. The RX 590 has 2304 shading units, 144 TMUs, and only 32 ROPs, with no dedicated ray tracing or tensor hardware. This fundamental difference in compute resources explains why the RTX A4000 achieves 19.17 TFLOPS of FP32 performance, while the RX 590 manages just 7.119 TFLOPS. Both cards offer 1:1 FP16 to FP32 ratios, but the RTX A4000’s raw throughput is far higher.

Memory architecture also diverges. The RTX A4000 uses 16 GB of GDDR6 memory on a 256-bit bus, delivering 448.0 GB/s of bandwidth. The RX 590 has 8 GB of GDDR5 on the same 256-bit bus, but only achieves 256.0 GB/s. The RTX A4000 also supports PCIe 4.0 x16, while the RX 590 is limited to PCIe 3.0 x16. In terms of API support, the RTX A4000 offers DirectX 12 Ultimate (12_2) and Vulkan 1.4, whereas the RX 590 is capped at DirectX 12 (12_0) and Vulkan 1.3. This makes the RTX A4000 more future-proof for modern software that leverages the latest API features.

Head-to-Head Benchmarks

The only direct head-to-head benchmark in the data is 3DMark Steel Nomad DX12, and it is a landslide. The NVIDIA RTX A4000 scores 2604 points, while the AMD Radeon RX 590 scores 1081 points. The delta is 140.9%, meaning the RTX A4000 is nearly two and a half times faster in this DirectX 12 test. This is the most definitive metric for comparing modern gaming or compute performance, and it clearly shows the architectural advantage of Ampere over Polaris.

Outside of this single test, the comparison relies on individual benchmark scores. The RTX A4000’s PassMark G3D score is 19459, which is not directly matched by the RX 590 in the data, but the average benchmark scores provide context. The RTX A4000’s average of 26683 is 7.8% higher than the RX 590’s 24744. Breaking down the PassMark results, the RTX A4000 wins in DirectX 11 with a score of 158 versus the RX 590’s 126 in DirectX 10, though these are different tests. The RTX A4000 also scores 1024 in PassMark G2D, which measures 2D graphics performance, a metric not listed for the RX 590.

The RTX A4000’s nearest rivals by average score include the AMD Radeon RX 5700 XT 50th Anniversary at 26553 (0.5% slower) and the NVIDIA GeForce RTX 5060 at 26331 (1.3% slower). The RX 590’s nearest rivals include the NVIDIA RTX A5000 Mobile at 24763 (0.1% faster) and the AMD Radeon RX 6600 XT at 24442 (1.2% slower). This places the RX 590 in a lower performance tier, closer to entry-level mobile GPUs, while the RTX A4000 sits alongside mid-range desktop cards.

FAQ

Q: Which GPU has more VRAM, and does it matter?

A: The NVIDIA RTX A4000 has 16 GB of GDDR6 memory, while the AMD Radeon RX 590 has 8 GB of GDDR5. The RTX A4000 also offers 448.0 GB/s bandwidth versus 256.0 GB/s. This means the RTX A4000 can handle larger textures and datasets, making it more suitable for modern workloads.

Q: Can the RX 590 run ray tracing?

A: No. The AMD Radeon RX 590 has no ray tracing cores listed in the data. The NVIDIA RTX A4000 includes 48 ray tracing cores and 192 tensor cores, enabling hardware-accelerated ray tracing and AI features.

Q: How do the power requirements compare?

A: The RTX A4000 has a TDP of 140 W and requires a 300 W suggested PSU with a single 6-pin connector. The RX 590 has a higher TDP of 175 W and needs a 450 W PSU with a single 8-pin connector. The RTX A4000 is more power-efficient despite its higher performance.

Q: Which card is better for older DirectX games?

A: The RX 590 shows higher scores in legacy APIs. It scores 126 in PassMark DirectX 10 and 240 in PassMark DirectX 9, while the RTX A4000 scores 72 and 240 in the same tests. For DirectX 9, they are tied, but the RX 590 is ahead in DirectX 10.

Q: What is the form factor difference?

A: The RTX A4000 is a single-slot card measuring 241 mm in length and 112 mm in height. The RX 590 is a dual-slot card with the same 241 mm length. This makes the RTX A4000 more suitable for compact or multi-GPU workstation builds.

Q: What is the release timing for each card?

A: The RX 590 was released on 2018-11-14, while the RTX A4000 came later on 2021-04-11. Both are end-of-life products, but the RTX A4000 is from a newer generation with a successor in Workstation Ada.

Specification Differences

The two cards differ in nearly every core specification. The NVIDIA RTX A4000 uses the GA104 chip on an 8 nm Samsung process, while the AMD Radeon RX 590 uses Polaris 30 on a 12 nm GlobalFoundries process. Transistor count is 17,400 million versus 5,700 million, and die size is 392 mm² versus 232 mm². The RTX A4000 has 6144 shading units, 192 TMUs, and 96 ROPs, while the RX 590 has 2304 shading units, 144 TMUs, and 32 ROPs. The RTX A4000 also adds 48 RT cores and 192 tensor cores, which the RX 590 lacks entirely.

Clock speeds differ, with the RTX A4000 running at 735 MHz base and 1560 MHz boost, while the RX 590 runs higher at 1469 MHz base and 1545 MHz boost. Memory is 16 GB GDDR6 at 1750 MHz (14 Gbps effective) for the RTX A4000 versus 8 GB GDDR5 at 2000 MHz (8 Gbps effective) for the RX 590. Bandwidth is 448.0 GB/s versus 256.0 GB/s. Pixel rate is 149.8 GPixel/s for the RTX A4000 versus 49.44 GPixel/s for the RX 590, and texture rate is 299.5 GTexel/s versus 222.5 GTexel/s. FP32 performance is 19.17 TFLOPS versus 7.119 TFLOPS.

Power and physical specs also vary. The RTX A4000 has a 140 W TDP, single-slot design, and one 6-pin connector, with a suggested 300 W PSU. The RX 590 has a 175 W TDP, dual-slot design, and one 8-pin connector, with a suggested 450 W PSU. The RTX A4000 supports PCIe 4.0 x16, while the RX 590 is PCIe 3.0 x16. Display outputs are 4x DisplayPort 1.4a on the RTX A4000 versus 1x HDMI 2.0b and 3x DisplayPort 1.4a on the RX 590. The RTX A4000 also supports DirectX 12 Ultimate and Vulkan 1.4, while the RX 590 is limited to DirectX 12 and Vulkan 1.3.

The Verdict

The data is unambiguous: the NVIDIA RTX A4000 is the superior GPU in almost every measurable aspect. In 3DMark Steel Nomad DX12, it beats the RX 590 by 140.9%. It offers more than double the memory, nearly triple the bandwidth, and over 2.7 times the FP32 compute throughput. The RTX A4000 also brings ray tracing and tensor cores, which the RX 590 does not have. For any modern workstation task, 3D rendering, or compute workload, the RTX A4000 is the clear choice.

The AMD Radeon RX 590 holds a narrow advantage only in legacy DirectX 10 performance, scoring 126 versus the RTX A4000’s 72 in PassMark DirectX 10. Both cards tie in DirectX 9 with a score of 240. This makes the RX 590 a viable option only for users running very old software that relies on these APIs and cannot benefit from the RTX A4000’s architectural advancements. The RX 590 also has a lower percentile ranking at 70 versus 72, and its nearest rivals include much weaker hardware.

For the vast majority of users, the RTX A4000 is the better purchase, provided the higher power efficiency and single-slot form factor are priorities. The RX 590 is a legacy card from 2018 with a 175 W TDP and no modern features. The verdict from the benchmark data is straightforward: the RTX A4000 wins on performance, memory, bandwidth, and feature set. The RX 590 is only relevant for those specifically targeting DirectX 9 and 10 workloads where it matches or slightly edges out the newer card.