AMD Radeon RX 6900 XT vs NVIDIA Quadro P6000 Comparison

Where Each One Wins

The benchmark data splits cleanly along a single axis: the AMD Radeon RX 6900 XT wins every recorded head-to-head test, while the NVIDIA Quadro P6000 does not win a single one. In the two shared workloads, Geekbench OpenCL and Geekbench Vulkan, AMD leads by margins of 64.6% and 50.5% respectively. That is not a close contest; it is a decisive sweep.

For compute-heavy tasks that rely on OpenCL, the RX 6900 XT is in another class entirely. Its OpenCL score of 187,673 dwarfs the Quadro’s 66,382. The gap is so wide that the Quadro P6000 would need to nearly triple its output to match. In Vulkan, the same story repeats: 148,525 versus 73,590, with AMD ahead by more than half. If your workload is built around either API, the choice is unambiguous.

The Quadro P6000’s only consolation is context. Its average benchmark score of 69,986 places it in the 90th percentile of all GPUs in the database, and it sits within 1.2% of the AMD Radeon RX 6600 LE, a modern midrange card. The RX 6900 XT, by contrast, averages 50,951 and sits in the 86th percentile. That lower percentile is misleading, because the RX 6900 XT’s average is dragged down by a much broader benchmark suite that includes older DirectX 9 and DirectX 10 tests, where it scores a meager 268 and 167 respectively. In the modern workloads that matter for current software, AMD wins outright.

So the use-case split is not about different strengths; it is about the RX 6900 XT being the stronger card in every measured scenario. The Quadro P6000’s advantage, if any, lies outside these benchmarks: its 24 GB memory capacity, its professional display outputs, and its end-of-life status as a workstation part. But in raw compute performance, the data gives AMD the win everywhere.

Architecture Differences

The two cards come from different eras and different design philosophies. The NVIDIA Quadro P6000 uses the GP102 chip on the Pascal architecture, built on a 16 nm process at TSMC. It packs 11,800 million transistors into a 471 mm² die, yielding a transistor density of 25.1 million per square millimeter. The AMD Radeon RX 6900 XT uses the Navi 21 chip on RDNA 2.0, built on a 7 nm process, also at TSMC. It holds 26,800 million transistors on a 520 mm² die, for a density of 51.5 million per square millimeter. That is more than double the density, a direct result of the smaller process node.

Memory is another clear divider. The Quadro P6000 carries 24 GB of GDDR5X on a 384-bit bus, delivering 432.8 GB/s of bandwidth. The RX 6900 XT has 16 GB of GDDR6 on a 256-bit bus, but delivers 512.0 GB/s. Fewer gigabytes, wider effective throughput. The Quadro’s memory runs at 9 Gbps effective, while AMD’s runs at 16 Gbps effective. For bandwidth-bound tasks, the RX 6900 XT has a 18% advantage despite having a narrower bus.

Compute resources differ substantially. The Quadro P6000 has 3,840 shading units, 240 texture mapping units, and 96 ROPs. The RX 6900 XT has 5,120 shading units, 320 TMUs, and 128 ROPs. AMD also includes 80 ray tracing cores, a feature the Pascal-based Quadro lacks entirely. The Quadro has no tensor cores either, while the RX 6900 XT’s architecture natively supports ray tracing hardware. In raw throughput, AMD leads in every category: pixel rate is 288.0 GPixel/s versus 157.9 GPixel/s, texture rate is 720.0 GTexel/s versus 394.8 GTexel/s, and FP32 compute is 23.04 TFLOPS versus 12.63 TFLOPS.

The FP16 comparison is stark. The Quadro P6000 delivers 197.4 GFLOPS at a 1:64 ratio, meaning FP16 is heavily de-emphasized. The RX 6900 XT delivers 46.08 TFLOPS at a 2:1 ratio, making FP16 a first-class citizen. That is a 233x difference in raw FP16 throughput, which matters for machine learning inference and certain scientific workloads.

Power and physical design also diverge. The Quadro P6000 is rated at 250 W TDP, uses a single 8-pin connector, and is dual-slot. The RX 6900 XT is rated at 300 W TDP, requires two 8-pin connectors, and occupies a triple-slot design. The Quadro recommends a 600 W PSU; AMD recommends 700 W. Both are 267 mm long, but the RX 6900 XT is slightly taller at 120 mm versus 111 mm, and it has a defined width of 50 mm.

Interface and output differences round out the picture. The Quadro uses PCIe 3.0 x16, while the RX 6900 XT uses PCIe 4.0 x16, doubling the theoretical bandwidth to the host. Display outputs differ too: the Quadro offers 1x DVI and 4x DisplayPort 1.4a, while the RX 6900 XT offers 1x HDMI 2.1, 2x DisplayPort 1.4a, and 1x USB Type-C. API support favors AMD with DirectX 12 Ultimate (12_2), while NVIDIA only reaches DirectX 12 (12_1). Both support OpenGL 4.6 and Vulkan 1.4.

Head-to-Head Benchmarks

Only two benchmarks appear in both cards’ recorded results, but they are decisive. In Geekbench OpenCL, the RX 6900 XT scores 187,673 against the Quadro P6000’s 66,382. The delta is -64.6% from AMD’s perspective, meaning the Quadro is roughly one third of AMD’s performance. That is not a marginal gap; it is a generational chasm. The Quadro’s FP32 output of 12.63 TFLOPS versus AMD’s 23.04 TFLOPS explains part of the difference, but the OpenCL score suggests AMD’s driver and architecture extract more efficiency from the hardware.

In Geekbench Vulkan, the RX 6900 XT scores 148,525 against the Quadro’s 73,590. The delta is -50.5%, so AMD is twice as fast. Vulkan is a lower-level API that rewards raw hardware throughput, and AMD’s RDNA 2 design with its 5,120 shading units and 80 ray tracing cores simply out-muscles the older Pascal chip. The Quadro’s Vulkan score is actually higher than its OpenCL score (73,590 versus 66,382), which is unusual and suggests the Pascal driver handles Vulkan relatively well. But relative improvement does not close a 50% gap.

Beyond the shared tests, the RX 6900 XT has a broader benchmark footprint. Its PassMark G3D score is 26,732, its GPU compute score is 14,547, and its Geekbench Metal score is 177,021. The Quadro P6000 has no recorded scores in those tests, so direct comparison is impossible. The RX 6900 XT’s lower average score (50,951) comes from including legacy DirectX tests: DirectX 9 at 268, DirectX 10 at 167, DirectX 11 at 279, and DirectX 12 at 113. These are low scores that pull the average down, but they reflect old API paths, not real-world modern performance.

The nearest rival data puts the Quadro in a tight cluster. Its average score of 69,986 is 0.2% below the AMD Radeon Pro WX 8200, 0.2% above the NVIDIA RTX A3000 Mobile, 1.2% below the AMD Radeon RX 6600 LE, and 1.4% above the NVIDIA CMP 90HX. The RX 6900 XT’s average of 50,951 sits 1.6% below the NVIDIA CMP 50HX, 1.9% above the AMD Radeon RX Vega 64, 2% above the GeForce RTX 5070 Ti, and 2.4% above the Intel Arc A550M. In both cases, the cards are within a few percentage points of their closest competitors, but the head-to-head results show a massive gap between the two cards themselves.

FAQ

Q: Which card has more memory?

A: The NVIDIA Quadro P6000 has 24 GB of GDDR5X, while the AMD Radeon RX 6900 XT has 16 GB of GDDR6. The Quadro also uses a wider 384-bit bus, but AMD’s faster 16 Gbps effective memory gives it higher bandwidth at 512.0 GB/s versus 432.8 GB/s.

Q: Does the RX 6900 XT support ray tracing?

A: Yes. The AMD Radeon RX 6900 XT includes 80 ray tracing cores on its RDNA 2.0 architecture. The NVIDIA Quadro P6000 has no ray tracing cores, as its Pascal architecture predates hardware ray tracing support.

Q: Which card is faster in OpenCL?

A: The AMD Radeon RX 6900 XT is significantly faster. In Geekbench OpenCL, it scores 187,673 versus the Quadro P6000’s 66,382, a 64.6% advantage for AMD.

Q: What is the transistor density difference?

A: The RX 6900 XT has a density of 51.5 million transistors per square millimeter, while the Quadro P6000 has 25.1 million per square millimeter. This is due to AMD’s 7 nm process versus NVIDIA’s 16 nm process.

Q: Are these cards still in production?

A: Both are end-of-life. The Quadro P6000 was released on 2016-09-30, and the RX 6900 XT on 2020-10-27. The Quadro’s predecessor is Quadro Maxwell, its successor is Quadro Volta. The RX 6900 XT’s predecessor is Navi, its successor is Navi III.

Q: Which card has better API support?

A: The RX 6900 XT supports DirectX 12 Ultimate (12_2), while the Quadro P6000 only supports DirectX 12 (12_1). Both support OpenGL 4.6 and Vulkan 1.4.

The Verdict

The data is unambiguous: the AMD Radeon RX 6900 XT outperforms the NVIDIA Quadro P6000 in every recorded head-to-head benchmark. In OpenCL, AMD leads by 64.6%; in Vulkan, by 50.5%. The RX 6900 XT also has higher raw compute specs across the board: 23.04 TFLOPS FP32 versus 12.63 TFLOPS, 5,120 shading units versus 3,840, and 128 ROPs versus 96. It even has 80 ray tracing cores, which the Quadro lacks entirely.

Who should pick the Quadro P6000? Only someone who specifically needs 24 GB of VRAM for large datasets that exceed 16 GB, or who requires its professional display output configuration: 1x DVI and 4x DisplayPort 1.4a. The Quadro’s 90th percentile ranking among all GPUs is respectable, and its average score of 69,986 is close to modern midrange cards. But in raw compute, it is outclassed by a card that was released four years later.

Who should pick the RX 6900 XT? Anyone running modern compute workloads, especially those using OpenCL, Vulkan, or Metal. Its Geekbench Metal score of 177,021 and OpenCL score of 187,673 indicate strong performance across heterogeneous compute APIs. Its higher TDP of 300 W and triple-slot design are trade-offs, but the performance advantage is decisive. The RX 6900 XT’s lower percentile (86th) is an artifact of its broader test suite, not a sign of weakness. In the tests both cards share, AMD wins by a landslide.

Specification Differences

| Field | NVIDIA Quadro P6000 | AMD Radeon RX 6900 XT |

|---|---|---|

| Architecture | Pascal | RDNA 2.0 |

| Process node | 16 nm | 7 nm |

| Transistors | 11,800 million | 26,800 million |

| Die size | 471 mm² | 520 mm² |

| Transistor density | 25.1M / mm² | 51.5M / mm² |

| Base clock | 1506 MHz | 1825 MHz |

| Boost clock | 1645 MHz | 2250 MHz |

| Memory size | 24 GB | 16 GB |

| Memory type | GDDR5X | GDDR6 |

| Memory bus | 384 bit | 256 bit |

| Memory bandwidth | 432.8 GB/s | 512.0 GB/s |

| Shading units | 3840 | 5120 |

| TMUs | 240 | 320 |

| ROPs | 96 | 128 |

| Ray tracing cores | None | 80 |

| Pixel rate | 157.9 GPixel/s | 288.0 GPixel/s |

| Texture rate | 394.8 GTexel/s | 720.0 GTexel/s |

| FP32 compute | 12.63 TFLOPS | 23.04 TFLOPS |

| FP16 compute | 197.4 GFLOPS (1:64) | 46.08 TFLOPS (2:1) |

| TDP | 250 W | 300 W |

| Slot width | Dual-slot | Triple-slot |

| Power connectors | 1x 8-pin | 2x 8-pin |

| Suggested PSU | 600 W | 700 W |

| Bus interface | PCIe 3.0 x16 | PCIe 4.0 x16 |

| Display outputs | 1x DVI, 4x DisplayPort 1.4a | 1x HDMI 2.1, 2x DisplayPort 1.4a, 1x USB Type-C |

| DirectX support | 12 (12_1) | 12 Ultimate (12_2) |

| Height | 111 mm | 120 mm |

| Width | Not specified | 50 mm |

| Release date | 2016-09-30 | 2020-10-27 |

| Launch MSRP | 5,999 USD | 999 USD |