AMD Radeon Pro W6800X vs NVIDIA GeForce RTX 5090 D Comparison

Head-to-Head Benchmarks

The only directly comparable benchmark in the database is Geekbench OpenCL, and the result is decisive. The NVIDIA GeForce RTX 5090 D scores 310,674, while the AMD Radeon Pro W6800X scores 124,498. That puts the NVIDIA card 59.9% ahead, a massive margin that reflects the generational gulf between the two designs. The AMD card does not win a single head-to-head test in the recorded data, while the RTX 5090 D wins the only shared workload.

Looking beyond that single comparison, the broader benchmark profiles tell different stories. The AMD Radeon Pro W6800X has an average benchmark score of 160,671 across its recorded tests, which places it at the 97th percentile of all GPUs in the database. Its nearest rivals are the NVIDIA A100 PCIe 40 GB at 162,504 (1.1% ahead), the AMD Radeon PRO W7800 at 164,894 (2.6% ahead), the NVIDIA RTX A5500 at 165,217 (2.8% ahead), and the NVIDIA RTX 4500 Ada Generation at 166,094 (3.3% ahead). The W6800X sits just below all four, but the gaps are narrow, within 3.3% at worst. That is a tightly clustered field, and the W6800X is clearly competitive with those workstation-class parts.

The RTX 5090 D, by contrast, has an average benchmark score of 77,712 across its full test suite, placing it at the 92nd percentile. Its nearest rivals are the AMD Radeon RX 6650M XT at 76,904 (1.1% behind the RTX card), the AMD Radeon RX 6850M XT at 78,940 (1.6% ahead), the NVIDIA Tesla P100 PCIe 12 GB at 79,396 (2.1% ahead), and the NVIDIA Tesla P100 PCIe 16 GB at 79,605 (2.4% ahead). The average is dragged down by the PassMark DirectX tests, which are older workloads and return low absolute numbers (231 for DirectX 10, 371 for DirectX 11, 219 for DirectX 12, 434 for DirectX 9). The high scores in Geekbench OpenCL (310,674), Geekbench Vulkan (376,915), and PassMark G3D (44,065) show where the card actually excels. The average is a misleading summary for this GPU because the test set includes both modern and legacy workloads.

The OpenCL head-to-head result is the cleanest signal available. It uses the same API on both cards, and the RTX 5090 D is nearly 2.5 times faster in raw score. The W6800X's OpenCL score of 124,498 is respectable for a 2021 workstation card, but it is not in the same class as the newer NVIDIA part. The delta of 59.9% is the kind of margin that shows up in render times, compute jobs, and any OpenCL-accelerated workflow.

FAQ

Q: Which GPU wins the only direct head-to-head benchmark?

A: The NVIDIA GeForce RTX 5090 D wins Geekbench OpenCL with a score of 310,674 against 124,498 for the AMD Radeon Pro W6800X, a 59.9% advantage.

Q: How do the two cards compare in average benchmark scores?

A: The AMD Radeon Pro W6800X has an average score of 160,671, while the NVIDIA GeForce RTX 5090 D averages 77,712. However, the RTX 5090 D's average is pulled down by legacy PassMark DirectX tests, while its modern API scores are much higher.

Q: What percentile do these GPUs occupy in the database?

A: The AMD Radeon Pro W6800X is at the 97th percentile of all GPUs, while the NVIDIA GeForce RTX 5090 D is at the 92nd percentile.

Q: Is the W6800X competitive with its nearest rivals?

A: Yes. The W6800X trails the NVIDIA A100 PCIe 40 GB by 1.1%, the AMD Radeon PRO W7800 by 2.6%, the NVIDIA RTX A5500 by 2.8%, and the NVIDIA RTX 4500 Ada Generation by 3.3%. All gaps are under 4%.

Q: What are the RTX 5090 D's strongest benchmark results?

A: Its Geekbench Vulkan score is 376,915, Geekbench OpenCL is 310,674, PassMark G3D is 44,065, and PassMark GPU Compute is 28,396. Its legacy DirectX scores are far lower.

Q: What is the W6800X's strongest benchmark result?

A: Its Geekbench Metal score is 196,844, which is its highest recorded result. Its OpenCL score is 124,498.

Architecture Differences

The two GPUs come from different architectural generations and design philosophies. The AMD Radeon Pro W6800X uses the Navi 21 chip built on RDNA 2.0 architecture, manufactured on a 7 nm process at TSMC. It packs 26,800 million transistors into a 520 mm² die, giving a transistor density of 51.5 million per square millimeter. The NVIDIA GeForce RTX 5090 D uses the GB202 chip on Blackwell 2.0 architecture, also fabricated by TSMC but on a 5 nm process. It contains 92,200 million transistors on a 750 mm² die, for a density of 122.9 million per square millimeter. The RTX 5090 D has more than three times the transistor count and more than double the density, which explains much of its performance advantage.

Ray tracing hardware differs significantly. The W6800X has 60 ray accelerators, while the RTX 5090 D has 170 RT cores. Tensor cores are absent from the AMD card entirely, while the NVIDIA card includes 680 tensor cores. That means the RTX 5090 D can accelerate AI and machine learning workloads that the W6800X cannot handle with dedicated hardware. Shading units also diverge sharply: the W6800X has 3,840 shading units, while the RTX 5090 D has 21,760. Texture mapping units number 240 on the AMD card versus 680 on the NVIDIA card, and render output units are 96 versus 176.

The memory subsystems reflect different generations as well. The W6800X uses 32 GB of GDDR6 on a 256-bit bus, delivering 512.0 GB/s of bandwidth. The RTX 5090 D uses 32 GB of GDDR7 on a 512-bit bus, delivering 1.79 TB/s. That is more than three times the bandwidth, which matters for large datasets, high-resolution textures, and compute workloads that stream data through memory. The NVIDIA card also uses a PCIe 5.0 x16 bus interface, while the AMD card uses Apple MPX, which ties it to Apple's Mac Pro ecosystem.

Compute throughput follows the same pattern. The W6800X delivers 16.03 TFLOPS of FP32 performance and 32.06 TFLOPS of FP16 (at a 2:1 ratio). The RTX 5090 D delivers 104.8 TFLOPS of FP32 and 104.8 TFLOPS of FP16 (at a 1:1 ratio). Those are enormous gaps. Pixel rate is 200.4 GPixel/s on the AMD card versus 423.6 GPixel/s on the NVIDIA card, and texture rate is 500.9 GTexel/s versus 1,636.8 GTexel/s. Every throughput metric favors the RTX 5090 D by a wide margin.

Specification Differences

The two cards differ on nearly every measurable specification. The W6800X has a base clock of 1800 MHz and a boost clock of 2087 MHz. The RTX 5090 D has a base clock of 2017 MHz and a boost clock of 2407 MHz. Memory clocks also differ: the AMD card runs at 2000 MHz with 16 Gbps effective, while the NVIDIA card runs at 1750 MHz with 28 Gbps effective. The AMD card has 32 GB of GDDR6, the NVIDIA card has 32 GB of GDDR7. Bus width is 256-bit on the AMD card and 512-bit on the NVIDIA card, and bandwidth is 512.0 GB/s versus 1.79 TB/s.

Power requirements diverge substantially. The W6800X has a TDP of 200 W and suggests a 550 W power supply. The RTX 5090 D has a TDP of 575 W and suggests a 950 W power supply. Physical dimensions differ too: the AMD card is 267 mm long and 120 mm tall, while the NVIDIA card is 304 mm long, 137 mm tall, and 48 mm wide. The AMD card is quad-slot with an Apple MPX power connector, while the NVIDIA card is dual-slot with a single 16-pin connector. Display outputs are 1x HDMI 2.1 and 4x Thunderbolt on the AMD card, versus 1x HDMI 2.1b and 3x DisplayPort 2.1b on the NVIDIA card.

Both cards support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The W6800X was released on 2021-08-02 and is end-of-life, with a launch MSRP of 2,799 USD. The RTX 5090 D was released on 2025-01-29 and is active, with a launch MSRP of 2,299 USD. The NVIDIA card lists the GeForce 40 as its predecessor and GeForce 60 as its successor. The AMD card lists no predecessor or successor in the database.

Where Each One Wins

The AMD Radeon Pro W6800X wins in the context of its own benchmark pool. Its average score of 160,671 is higher than that of the RTX 5090 D, and it sits at the 97th percentile, above the RTX 5090 D's 92nd. It also has a much lower power draw at 200 W versus 575 W, and it requires a smaller power supply (550 W versus 950 W). For systems where power delivery or cooling is constrained, the W6800X is the more practical choice. Its Apple MPX interface and Thunderbolt outputs make it the only option of the two for Apple Mac Pro systems. Its Geekbench Metal score of 196,844 shows strong performance in Apple's graphics API, and its nearest rivals are all within 3.3%, meaning it sits comfortably in workstation-class territory.

The NVIDIA GeForce RTX 5090 D wins in raw compute across the board. The head-to-head OpenCL result is a 59.9% margin, and its modern API scores are far above the AMD card's recorded results. Its Geekbench Vulkan score of 376,915 and PassMark G3D score of 44,065 indicate strong performance in current graphics workloads. The 680 tensor cores give it a capability the AMD card simply does not have for AI and machine learning tasks. The 1.79 TB/s memory bandwidth, 170 RT cores, and 104.8 TFLOPS of FP32 throughput all point to a card built for heavy compute, rendering, and high-end gaming. The dual-slot form factor is also easier to fit into standard PC cases compared to the quad-slot AMD card.

The Verdict

The data points to a clear split. For Apple Mac Pro users, the AMD Radeon Pro W6800X is the only compatible option due to its Apple MPX bus interface and Thunderbolt outputs. Its 200 W power draw and 550 W suggested PSU make it far easier to integrate into existing systems, and its 97th percentile standing with narrow gaps to its workstation rivals means it remains a capable card. The 32 GB of GDDR6 memory and 512.0 GB/s bandwidth are adequate for many professional workloads, and the 196,844 Metal score shows it performs well in Apple's ecosystem.

For anyone building a PC and prioritizing raw performance, the NVIDIA GeForce RTX 5090 D is the stronger choice. It wins the only direct head-to-head benchmark by 59.9%, and its architectural advantages are overwhelming: more than triple the shading units, nearly triple the TMUs, more than double the ROPs, 680 tensor cores, and 1.79 TB/s of memory bandwidth. The 104.8 TFLOPS of FP32 performance is more than six times the AMD card's 16.03 TFLOPS. The higher power draw of 575 W is a real cost, but the performance data justifies it for compute-heavy workloads. The RTX 5090 D's average score is misleading due to legacy DirectX tests, but its modern API results are the relevant metric for current software.

The choice comes down to platform and workload. The W6800X is the pick for Apple systems and power-sensitive environments. The RTX 5090 D is the pick for maximum compute performance in a PC, especially where tensor cores, ray tracing, or massive memory bandwidth matter. The benchmark data does not support the AMD card in a straight performance comparison, but its niche advantages in power, platform compatibility, and percentile standing keep it relevant for specific use cases.