AMD Radeon R9 M290X vs NVIDIA RTX PRO 2000 Blackwell Comparison
AMD Radeon R9 M290X
RTX PRO 2000 Blackwell
PERFORMANCE BENCHMARKS
Analysis: AMD Radeon R9 M290X vs NVIDIA RTX PRO 2000 Blackwell
Where Each One Wins
The benchmark data paints a starkly lopsided picture. The NVIDIA RTX PRO 2000 Blackwell wins the only head-to-head benchmark available, and it does so decisively. In Geekbench OpenCL, the NVIDIA card scores 106,087 against the AMD Radeon R9 M290X's 22,028 — a 381.6% delta. That is not a marginal victory; it is a generational gap expressed in raw compute throughput.
Looking at the broader benchmark suite, the NVIDIA card's strengths are spread across modern workloads. Its 3DMark Steel Nomad DX12 score of 2,374.5 indicates strong DirectX 12 Ultimate performance, which aligns with its API support for feature level 12_2. The Geekbench Vulkan score of 113,865 further reinforces that the card excels in contemporary graphics APIs. In Passmark, the NVIDIA card shows a G3D score of 20,049 and a GPU compute score of 8,396, suggesting balanced performance between rasterization and compute tasks.
The AMD Radeon R9 M290X, by contrast, only has two benchmarks in the record: Geekbench Metal at 24,524 and Geekbench OpenCL at 22,028. The Metal score is noteworthy because the NVIDIA card has no corresponding Metal benchmark listed — this is the only arena where the AMD card has a data point the NVIDIA card lacks. However, even that score is modest when placed next to the NVIDIA card's OpenCL and Vulkan results. The AMD card's 68th percentile ranking versus the NVIDIA card's 70th percentile shows they sit in a similar overall tier, but the average benchmark scores tell a different story: 25,269 for NVIDIA versus 23,276 for AMD.
The data implies a clear use-case split. The NVIDIA RTX PRO 2000 Blackwell is built for modern, compute-heavy professional workloads — ray tracing, tensor operations, and high-throughput parallel processing. The AMD Radeon R9 M290X, with its older architecture and limited benchmark footprint, appears suited to legacy applications or scenarios where Metal API support is mandatory. For anyone running current DX12 or Vulkan titles or GPU-accelerated compute, the NVIDIA card is the only rational choice based on this evidence.
Architecture Differences
The architectural gap between these two GPUs is vast, and the numbers quantify just how far GPU design has progressed. The NVIDIA RTX PRO 2000 Blackwell uses the GB206 chip built on TSMC's 5 nm process, packing 21,900 million transistors into a 181 mm² die. That yields a transistor density of 121.0 million per square millimeter. The AMD Radeon R9 M290X uses the Neptune chip on a 28 nm process, with 2,800 million transistors spread across a larger 212 mm² die, resulting in just 13.2 million transistors per square millimeter.
The compute resources differ by an order of magnitude. The NVIDIA card has 4,352 shading units, 136 texture mapping units, and 48 ROPs, alongside 34 RT cores and 136 tensor cores. The AMD card has 1,280 shading units, 80 TMUs, and 32 ROPs — no RT cores, no tensor cores. This means the NVIDIA card supports hardware-accelerated ray tracing and AI tensor operations, while the AMD card relies entirely on traditional shader-based rendering.
Memory architecture also diverges sharply. The NVIDIA card features 16 GB of GDDR7 memory on a 128-bit bus, delivering 288.0 GB/s of bandwidth. The AMD card has 4 GB of GDDR5 on a 256-bit bus, providing 153.6 GB/s. Despite having half the bus width, the NVIDIA card nearly doubles the memory bandwidth thanks to faster GDDR7 memory. The AMD card's memory clock runs at 1200 MHz with 4.8 Gbps effective transfer rate, while the NVIDIA card's memory runs at 1125 MHz but achieves 18 Gbps effective — a 3.75x improvement in per-pin transfer speed.
Clock speeds tell a similar story. The NVIDIA card has a base clock of 982 MHz and a boost of 1957 MHz. The AMD card's base is 850 MHz with a boost of 900 MHz. The NVIDIA card's boost clock is more than double the AMD card's boost clock, which partially explains the massive FP32 throughput difference: 17.03 TFLOPS versus 2.304 TFLOPS. The NVIDIA card also lists FP16 performance at 17.03 TFLOPS (1:1 ratio), a feature the AMD card lacks entirely.
Power and physical design further differentiate them. The NVIDIA card has a 70 W TDP, runs dual-slot, and draws power from the PCIe slot with no external connectors. The AMD card has a 100 W TDP and uses an MXM Module form factor. The NVIDIA card is a standard PCIe 5.0 x8 card measuring 167 mm by 69 mm by 20 mm, while the AMD card's dimensions are not listed, suggesting a proprietary mobile implementation.
FAQ
Q: Which GPU has higher raw compute throughput?
A: The NVIDIA RTX PRO 2000 Blackwell delivers 17.03 TFLOPS FP32 performance, while the AMD Radeon R9 M290X manages 2.304 TFLOPS. This represents a roughly 7.4x advantage for the NVIDIA card in single-precision compute.
Q: Does the AMD card support ray tracing or tensor operations?
A: No. The AMD Radeon R9 M290X has no RT cores or tensor cores listed in its specification. The NVIDIA RTX PRO 2000 Blackwell includes 34 RT cores and 136 tensor cores, enabling hardware-accelerated ray tracing and AI workloads.
Q: How do their memory subsystems compare?
A: The NVIDIA card has 16 GB of GDDR7 memory with 288.0 GB/s bandwidth on a 128-bit bus. The AMD card has 4 GB of GDDR5 with 153.6 GB/s bandwidth on a 256-bit bus. The NVIDIA card provides 4x the capacity and nearly double the bandwidth despite a narrower bus.
Q: Which GPU supports more modern graphics APIs?
A: The NVIDIA RTX PRO 2000 Blackwell supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The AMD Radeon R9 M290X supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The NVIDIA card's DX12 feature level is higher, and its Vulkan version is newer.
Q: What is the performance difference in the only shared benchmark?
A: In Geekbench OpenCL, the NVIDIA card scores 106,087 versus the AMD card's 22,028, a 381.6% advantage for the NVIDIA RTX PRO 2000 Blackwell. This is the sole head-to-head benchmark recorded between the two.
Q: How do their production statuses differ?
A: The NVIDIA RTX PRO 2000 Blackwell is listed as Active production status, released in August 2025. The AMD Radeon R9 M290X is End-of-life, released in January 2014, with its successor being Polaris Mobile.
Specification Differences
The two cards differ in nearly every measurable specification. The process node is 5 nm for NVIDIA versus 28 nm for AMD. Transistor count is 21,900 million versus 2,800 million. Die size is 181 mm² versus 212 mm², meaning NVIDIA fits nearly 8x more transistors into a smaller die.
Clock speeds: NVIDIA base 982 MHz / boost 1957 MHz; AMD base 850 MHz / boost 900 MHz. Memory: NVIDIA 16 GB GDDR7 at 1125 MHz with 18 Gbps effective; AMD 4 GB GDDR5 at 1200 MHz with 4.8 Gbps effective. Bus width: 128-bit versus 256-bit. Bandwidth: 288.0 GB/s versus 153.6 GB/s.
Compute units: 4,352 shading units, 136 TMUs, 48 ROPs, 34 RT cores, 136 tensor cores for NVIDIA; 1,280 shading units, 80 TMUs, 32 ROPs, no RT or tensor cores for AMD. Pixel rate: 93.94 GPixel/s versus 28.80 GPixel/s. Texture rate: 266.2 GTexel/s versus 72.00 GTexel/s. FP32: 17.03 TFLOPS versus 2.304 TFLOPS.
Power: 70 W TDP versus 100 W TDP. Form factor: dual-slot PCIe 5.0 x8 card versus MXM Module with PCIe 3.0 x16. Display outputs: 4x mini-DisplayPort 2.1b versus portable-device-dependent outputs. The NVIDIA card has a suggested PSU of 250 W; the AMD card has no suggested PSU listed. The NVIDIA card supports DirectX 12 Ultimate and Vulkan 1.4; the AMD card supports DirectX 12 (11_1) and Vulkan 1.2.170.
Head-to-Head Benchmarks
The single head-to-head benchmark is Geekbench OpenCL, and it is not close. The NVIDIA RTX PRO 2000 Blackwell scores 106,087, while the AMD Radeon R9 M290X scores 22,028. The delta is 381.6% in favor of NVIDIA. To put that in perspective, the NVIDIA card is roughly 4.8 times faster than the AMD card in this OpenCL compute test.
This result is consistent with the architectural disparity. The NVIDIA card has 3.4x more shading units, 1.7x more TMUs, 1.5x more ROPs, and a 7.4x higher FP32 throughput. When you combine the higher clock speeds, the massive shader advantage, and the faster GDDR7 memory, the OpenCL result makes perfect sense.
The AMD card's only other benchmark, Geekbench Metal, scores 24,524. There is no direct NVIDIA comparison available, but this score is comparable to the AMD card's OpenCL result, suggesting its compute performance is consistently in the low-20,000s range regardless of API. The NVIDIA card's nearest rivals in average score include the AMD Radeon RX 6700M at 25,633 and the NVIDIA GeForce RTX 3080 Ti Mobile at 25,740, both within 1.8% of the RTX PRO 2000 Blackwell's average of 25,269. Meanwhile, the AMD R9 M290X's nearest rival is the AMD Radeon RX 6600M at 23,273, a 0% delta, showing it sits in a much lower performance class.
The wins tally is 1-0 in favor of NVIDIA. The data does not support any scenario where the AMD card wins a compute benchmark. The only question is whether the AMD card's Metal support gives it relevance in Apple-centric workflows, but even that score of 24,524 is far below the NVIDIA card's Vulkan and OpenCL results.
The Verdict
The data is unambiguous. The NVIDIA RTX PRO 2000 Blackwell is the superior GPU in every measured category. It wins the only head-to-head benchmark by 381.6%, has a higher percentile ranking (70th versus 68th), a higher average benchmark score (25,269 versus 23,276), and offers architectural features — RT cores, tensor cores, GDDR7 memory, 16 GB capacity — that the AMD Radeon R9 M290X simply does not have.
Who should pick the NVIDIA card? Anyone running modern DirectX 12 or Vulkan applications, GPU compute workloads, ray-traced rendering, or AI inference. The card's 17.03 TFLOPS FP32 and FP16 (1:1) performance, combined with 16 GB of GDDR7 memory and 288.0 GB/s bandwidth, make it suitable for professional workstation tasks. Its 70 W TDP and lack of external power connectors also suggest it can fit into power-constrained systems.
Who should pick the AMD Radeon R9 M290X? The case is narrow. The card supports Metal, which the NVIDIA card does not have a listed benchmark for, and it is an MXM Module form factor — relevant for certain mobile or embedded systems. Its 68th percentile ranking and average score of 23,276 place it near the AMD Radeon RX 6600M (23,273) and NVIDIA P106-100 (23,249). But with no RT cores, no tensor cores, 4 GB of older GDDR5 memory, and an End-of-life production status, the AMD card is a legacy product. For any new deployment, the NVIDIA RTX PRO 2000 Blackwell is the data-backed choice. For those maintaining legacy systems with MXM slots and Metal-based software, the AMD card remains a functional, if dated, option.