AMD Radeon RX 6800 vs NVIDIA P104-100 Comparison

Head-to-Head Benchmarks

The recorded data presents a stark contrast between these two GPUs, with each card claiming decisive victories in different workloads. In the modern DirectX 12-based 3DMark Steel Nomad test, the AMD Radeon RX 6800 delivers a commanding performance, scoring 3188 points. The NVIDIA P104-100 manages only 1413 points, a deficit of 55.7%. This is not a marginal gap; the AMD card more than doubles the NVIDIA card's output in this particular benchmark, indicating a massive generational and architectural advantage in contemporary gaming workloads.

However, the narrative flips dramatically in compute-oriented tests. In the Geekbench OpenCL benchmark, the NVIDIA P104-100 posts an impressive score of 52368, dwarfing the AMD Radeon RX 6800's score of 24508. This represents a 113.7% advantage for the NVIDIA card, meaning it is more than twice as fast as its rival in this specific test. This result is particularly notable given that the P104-100 is a mining-specific product with no display outputs, yet its raw compute throughput in this API is exceptional.

The Geekbench Vulkan test shows another decisive win for the AMD Radeon RX 6800, which scores 115107 against the NVIDIA P104-100's 45165. This is a 60.8% lead for the AMD card, reinforcing its strength in modern graphics APIs. While the OpenCL result suggests the NVIDIA card has raw compute power that can outperform in certain scenarios, the Vulkan and DirectX 12 results tell a consistent story: the AMD Radeon RX 6800 is the superior graphics processor for contemporary rendering tasks.

Looking at the broader benchmark landscape, the two cards occupy similar percentile positions among all GPUs. The NVIDIA P104-100 sits at the 77th percentile, while the AMD Radeon RX 6800 is at the 75th percentile. Their average benchmark scores are close, with the NVIDIA card averaging 32982 and the AMD card averaging 30095. This proximity in overall standing is somewhat misleading given the head-to-head results, but it reflects the different workloads each card excels at. The NVIDIA card's nearest rivals include the NVIDIA T600 Mobile, T550 Mobile, GeForce RTX 3050 Mobile, and AMD Radeon Pro 570, all within a narrow band of 0.7% or less. The AMD card's nearest rivals are the NVIDIA GeForce RTX 3070 Ti, RTX 5070 Mobile, RTX 2080 Ti, and AMD Radeon RX 6700, with deltas ranging from 1.1% to 1%. The AMD card is essentially performance-equivalent to an RTX 3070 Ti based on average scores, while the NVIDIA P104-100 trades blows with mobile-class GPUs.

FAQ

Q: Which GPU is faster in 3DMark Steel Nomad?

A: The AMD Radeon RX 6800 is significantly faster, scoring 3188 compared to the NVIDIA P104-100's 1413, a 55.7% advantage for the AMD card.

Q: Does the NVIDIA P104-100 win any benchmark?

A: Yes, the NVIDIA P104-100 wins the Geekbench OpenCL test with a score of 52368, which is 113.7% higher than the AMD Radeon RX 6800's 24508.

Q: How do these GPUs compare in average benchmark scores?

A: The NVIDIA P104-100 has a slightly higher average benchmark score of 32982, while the AMD Radeon RX 6800 averages 30095. The NVIDIA card also sits at the 77th percentile versus the AMD card's 75th percentile.

Q: What is the memory configuration difference?

A: The AMD Radeon RX 6800 has 16 GB of GDDR6 memory on a 256-bit bus with 512.0 GB/s bandwidth. The NVIDIA P104-100 has 4 GB of GDDR5X memory on a 256-bit bus with 320.3 GB/s bandwidth.

Q: Which card supports hardware ray tracing?

A: Only the AMD Radeon RX 6800, which includes 60 ray accelerators (RT cores) and supports DirectX 12 Ultimate. The NVIDIA P104-100 has no RT cores listed and only supports DirectX 12 (12_1).

Q: What are the display output capabilities?

A: The AMD Radeon RX 6800 has 1x HDMI 2.1, 2x DisplayPort 1.4a, and 1x USB Type-C. The NVIDIA P104-100 has no display outputs at all, as it is a mining-specific product.

Architecture Differences

The architectural divide between these two GPUs is substantial, reflecting their different origins and purposes. The NVIDIA P104-100 is built on the Pascal architecture, using the GP104 chip fabricated on a 16 nm process at TSMC. It contains 7,200 million transistors on a 314 mm² die, resulting in a transistor density of 22.9 million per square millimeter. The AMD Radeon RX 6800, by contrast, uses the RDNA 2.0 architecture with the Navi 21 chip, produced on TSMC's 7 nm process. This newer node allows for 26,800 million transistors on a 520 mm² die, yielding a much higher transistor density of 51.5 million per square millimeter. The process node advantage alone explains a significant portion of the performance differential, as the AMD card fits nearly four times as many transistors into a die that is only about 65% larger.

The compute resources differ dramatically. The NVIDIA P104-100 has 1920 shading units, 120 texture mapping units, and 64 raster output units. The AMD Radeon RX 6800 more than doubles these counts with 3840 shading units, 240 TMUs, and 96 ROPs. The AMD card also features 60 ray accelerators, while the NVIDIA card has none, reflecting the P104-100's design as a mining product that predates mainstream ray tracing adoption. The NVIDIA card's lack of ray tracing hardware and its DirectX 12 (12_1) support, versus the AMD card's DirectX 12 Ultimate (12_2) support, places them in different capability tiers for modern gaming features.

The memory subsystems also reflect their divergent design goals. The NVIDIA P104-100 uses 4 GB of GDDR5X memory, while the AMD Radeon RX 6800 uses 16 GB of GDDR6. Both use a 256-bit bus, but the AMD card achieves 512.0 GB/s of bandwidth versus 320.3 GB/s for the NVIDIA card, a 60% increase. Clock speeds favor the AMD card as well, with a base clock of 1700 MHz and boost clock of 2105 MHz, compared to the NVIDIA card's 1607 MHz base and 1733 MHz boost. The AMD card even lists a game clock of 1815 MHz, a figure the NVIDIA card does not provide.

Specification Differences

The recorded specifications show clear divergence in nearly every category. The NVIDIA P104-100 is a mining GPU from the Pascal generation, while the AMD Radeon RX 6800 belongs to the Radeon RX 6000 series under the RDNA 2.0 architecture. Manufacturing processes differ, with the NVIDIA card at 16 nm and the AMD card at 7 nm. Transistor counts are 7,200 million versus 26,800 million, and die sizes are 314 mm² versus 520 mm². Transistor density is 22.9 million per square millimeter for the NVIDIA card and 51.5 million for the AMD card.

Clock specifications show the AMD card running faster across the board: 1700 MHz base versus 1607 MHz, 2105 MHz boost versus 1733 MHz, and memory at 2000 MHz (16 Gbps effective) versus 1251 MHz (10 Gbps effective). Memory capacity is 4 GB versus 16 GB, with GDDR5X versus GDDR6 type. Bandwidth is 320.3 GB/s versus 512.0 GB/s. The shading units, TMUs, and ROPs are all higher on the AMD card: 3840 versus 1920, 240 versus 120, and 96 versus 64 respectively. The AMD card has 60 ray accelerators, while the NVIDIA card has none.

Compute throughput measures strongly favor the AMD card, with FP32 performance at 16.17 TFLOPS versus 6.655 TFLOPS. The FP16 comparison is even more lopsided: 32.33 TFLOPS for the AMD card versus 104.0 GFLOPS for the NVIDIA card, a ratio of over 300 to 1, due to the AMD card's 2:1 FP16 rate versus the NVIDIA card's 1:64 rate. Pixel rate is 202.1 GPixel/s versus 110.9 GPixel/s, and texture rate is 505.2 GTexel/s versus 208.0 GTexel/s.

Power and interface details also differ. The AMD Radeon RX 6800 has a TDP of 250 W and requires a 600 W suggested PSU with two 8-pin power connectors. The NVIDIA P104-100 has no listed TDP but suggests a 200 W PSU with a single 8-pin connector. The bus interface is PCIe 4.0 x16 for the AMD card versus PCIe 1.0 x4 for the NVIDIA card, an enormous difference in host connectivity. Display outputs are present only on the AMD card, which offers HDMI 2.1, DisplayPort 1.4a, and USB Type-C, while the NVIDIA card has no outputs. Physical dimensions are identical in length at 267 mm, but the AMD card adds height of 120 mm and width of 40 mm, while the NVIDIA card's height and width are not recorded. Both are dual-slot cards.

Where Each One Wins

The AMD Radeon RX 6800 is the clear choice for modern gaming and graphics workloads. Its 55.7% lead in 3DMark Steel Nomad and 60.8% lead in Geekbench Vulkan demonstrate overwhelming superiority in DirectX 12 and Vulkan rendering. The 16 GB of GDDR6 memory provides ample capacity for high-resolution textures and large datasets, while the 512.0 GB/s bandwidth ensures fast data movement. The inclusion of 60 ray accelerators and DirectX 12 Ultimate support makes it future-proof for ray-traced titles and advanced graphics features. Its higher pixel rate of 202.1 GPixel/s and texture rate of 505.2 GTexel/s further cement its position as a gaming and content creation powerhouse.

The NVIDIA P104-100, despite being a mining product with no display outputs, wins decisively in Geekbench OpenCL with a 113.7% advantage. This suggests that in certain compute workloads that leverage OpenCL, the Pascal architecture's design can outperform the newer RDNA 2.0 architecture. The card's higher average benchmark score of 32982 versus 30095 and its 77th percentile ranking versus 75th indicate that in aggregate, it holds its own across a wide range of tests. For scenarios where OpenCL compute performance is the primary metric and modern graphics features are unnecessary, the P104-100 demonstrates surprising capability. Its lower power requirements, with a suggested PSU of 200 W versus 600 W, and single 8-pin connector versus dual, make it a less demanding card from a system integration perspective.

The use-case split is therefore quite clear. For gamers, content creators using DirectX 12 or Vulkan, and anyone needing display outputs or ray tracing, the AMD Radeon RX 6800 is the overwhelmingly better option. For specialized compute tasks that rely on OpenCL and where the lack of display outputs is not a limitation, the NVIDIA P104-100 offers a compelling alternative, particularly given its higher average benchmark score and percentile ranking. The two cards target different niches, and the benchmark data reflects their distinct design philosophies.