AMD Radeon RX 570X vs NVIDIA P106-090 Comparison

AMD Radeon RX 570X vs NVIDIA P106-090: the RX 570X is the clear winner in the only shared benchmark, posting an 82.8% higher Geekbench OpenCL score, but the P106-090 counters with a lower power draw and a smaller die. The RX 570X also holds a slight edge in overall average benchmark score, 13871 versus 13470, placing both cards within a percentage point of each other in global GPU percentile rankings (55th vs 54th).

FAQ

Q: Which card has the higher average benchmark score?

A: The AMD Radeon RX 570X leads with an average benchmark score of 13871, compared to the NVIDIA P106-090’s 13470. This places the RX 570X just 0.4% above its nearest rival (NVIDIA RTX A2000 Mobile), while the P106-090 sits 0.3% below the NVIDIA GeForce GTX 570.

Q: How do the cards compare in the Geekbench OpenCL test?

A: The RX 570X scores 38939, which is 82.8% higher than the P106-090’s 21304. This is the only head-to-head benchmark available, and the RX 570X wins it decisively.

Q: What is the difference in memory capacity and bandwidth?

A: The RX 570X has 8 GB of GDDR5 on a 256-bit bus, delivering 224.0 GB/s bandwidth. The P106-090 has 3 GB of GDDR5 on a 192-bit bus, delivering 192.2 GB/s. The RX 570X offers more capacity and higher bandwidth.

Q: Which card has a lower power requirement?

A: The NVIDIA P106-090 has a TDP of 75 W and a suggested PSU of 250 W, while the AMD Radeon RX 570X has a TDP of 150 W and a suggested PSU of 450 W. The P106-090 is far less power-hungry.

Q: Are there differences in display output capabilities?

A: Yes. The RX 570X features 1x DVI, 1x HDMI 2.0b, and 3x DisplayPort 1.4a outputs. The P106-090 has no display outputs at all, as it is designed for mining.

Q: What is the transistor density difference between the two?

A: The RX 570X packs 5,700 million transistors into a 232 mm² die, yielding 24.6M transistors per mm². The P106-090 has 4,400 million transistors on a 200 mm² die, yielding 22.0M per mm². The RX 570X has a higher density.

Architecture Differences

The AMD Radeon RX 570X is built on GCN 4.0 architecture using the Polaris 20 chip, fabricated on a 14 nm process at GlobalFoundries. The NVIDIA P106-090 uses the Pascal architecture with the GP106 chip, manufactured on a 16 nm process at TSMC. This process difference gives the RX 570X a slightly denser transistor layout, 24.6M per mm² versus 22.0M per mm², despite the RX 570X having a larger physical die at 232 mm² compared to 200 mm².

The RX 570X carries 2,048 shading units, 128 texture mapping units, and 32 ROPs. The P106-090 has only 768 shading units and 48 TMUs, but it has 48 ROPs, which is 50% more than the RX 570X. This ROP advantage explains the P106-090’s higher pixel rate of 73.49 GPixel/s versus 39.81 GPixel/s for the RX 570X. However, the RX 570X dominates in texture rate, achieving 159.2 GTexel/s versus 73.49 GTexel/s.

Compute throughput is heavily skewed toward AMD. The RX 570X delivers 5.095 TFLOPS of FP32 performance, more than double the P106-090’s 2.352 TFLOPS. In FP16, the gap is even more extreme: the RX 570X outputs 5.095 TFLOPS (1:1), while the P106-090 manages just 36.74 GFLOPS (1:64). Neither card has dedicated ray tracing or tensor cores.

Memory architecture differs significantly. The RX 570X uses 8 GB of GDDR5 on a 256-bit bus at 1750 MHz (7 Gbps effective), yielding 224.0 GB/s. The P106-090 has 3 GB of GDDR5 on a 192-bit bus at 2002 MHz (8 Gbps effective), yielding 192.2 GB/s. The RX 570X offers both more capacity and higher bandwidth, though the P106-090 runs its memory at a higher effective clock speed.

API support is comparable, with both supporting DirectX 12 (RX 570X: 12_0; P106-090: 12_1) and OpenGL 4.6. The P106-090 supports Vulkan 1.4, while the RX 570X supports Vulkan 1.3. The P106-090 also has a higher base clock (1354 MHz) and boost clock (1531 MHz) compared to the RX 570X’s 1168 MHz base and 1244 MHz boost, though this does not translate into higher compute performance.

Head-to-Head Benchmarks

The only direct benchmark comparison available is the Geekbench OpenCL test, where the AMD Radeon RX 570X scores 38939 against the NVIDIA P106-090’s 21304. This is an 82.8% advantage for the RX 570X, a massive gap that reflects the underlying compute difference. The RX 570X’s FP32 throughput of 5.095 TFLOPS versus 2.352 TFLOPS for the P106-090 supports this result, as OpenCL workloads typically scale with raw shading power.

For the P106-090, its strengths appear in other metrics not captured by this single test. Its pixel rate of 73.49 GPixel/s is nearly double the RX 570X’s 39.81 GPixel/s, which could benefit certain rasterization-heavy tasks, though no benchmark confirms this. Its higher ROP count (48 vs 32) and higher boost clock (1531 MHz vs 1244 MHz) suggest fill-rate advantages, but the lack of a head-to-head test means these remain theoretical.

The overall average benchmark scores tell a closer story. The RX 570X averages 13871, while the P106-090 averages 13470, a difference of roughly 3%. Both cards sit at nearly the same percentile among all GPUs: the RX 570X at the 55th percentile and the P106-090 at the 54th. This suggests that despite the OpenCL blowout, the two cards are broadly comparable in real-world aggregate performance, likely because the P106-090’s strengths in other areas offset its compute deficit.

The RX 570X’s nearest rival is the NVIDIA RTX A2000 Mobile with an average score of 13821 (0.4% delta), while the P106-090’s nearest rival is the NVIDIA GeForce GTX 570 at 13515 (-0.3% delta). These margins are tiny, indicating both cards are well-positioned against their immediate competition, but the RX 570X has a slight edge in absolute terms.

Specification Differences

| Specification | AMD Radeon RX 570X | NVIDIA P106-090 |

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

| Architecture | GCN 4.0 | Pascal |

| Process Node | 14 nm (GlobalFoundries) | 16 nm (TSMC) |

| Transistors | 5,700 million | 4,400 million |

| Die Size | 232 mm² | 200 mm² |

| Transistor Density | 24.6M / mm² | 22.0M / mm² |

| Base Clock | 1168 MHz | 1354 MHz |

| Boost Clock | 1244 MHz | 1531 MHz |

| Memory Size | 8 GB GDDR5 | 3 GB GDDR5 |

| Memory Bus | 256 bit | 192 bit |

| Memory Clock | 1750 MHz (7 Gbps) | 2002 MHz (8 Gbps) |

| Memory Bandwidth | 224.0 GB/s | 192.2 GB/s |

| Shading Units | 2048 | 768 |

| TMUs | 128 | 48 |

| ROPs | 32 | 48 |

| Pixel Rate | 39.81 GPixel/s | 73.49 GPixel/s |

| Texture Rate | 159.2 GTexel/s | 73.49 GTexel/s |

| FP32 | 5.095 TFLOPS | 2.352 TFLOPS |

| FP16 | 5.095 TFLOPS (1:1) | 36.74 GFLOPS (1:64) |

| TDP | 150 W | 75 W |

| Suggested PSU | 450 W | 250 W |

| Bus Interface | PCIe 3.0 x16 | PCIe 1.0 x1 |

| Display Outputs | 1x DVI, 1x HDMI 2.0b, 3x DisplayPort 1.4a | No outputs |

| Vulkan Support | 1.3 | 1.4 |

| Release Date | 2018-04-10 | 2017-07-30 |

The Verdict

The data clearly favors the AMD Radeon RX 570X for compute-heavy workloads. Its 82.8% lead in Geekbench OpenCL, combined with 5.095 TFLOPS FP32 performance and 8 GB of memory, makes it the superior choice for any task that leverages shading units or large datasets. The RX 570X also edges out the P106-090 in average benchmark score (13871 vs 13470) and holds a higher global percentile ranking (55th vs 54th).

The NVIDIA P106-090 is not without merit. Its 75 W TDP and 250 W suggested PSU make it far more power-efficient, and its 73.49 GPixel/s pixel rate suggests fill-rate advantages. However, with only 3 GB of memory and no display outputs, it is fundamentally a specialized mining part. Its PCIe 1.0 x1 interface also limits bandwidth in general-purpose systems, whereas the RX 570X uses PCIe 3.0 x16.

For gamers or general users, the RX 570X is the obvious pick due to its display outputs and higher compute capacity. For a mining or low-power compute scenario where display output is irrelevant, the P106-090’s lower power draw might be attractive, but its performance deficit in the only shared benchmark is a significant drawback.

Where Each One Wins

AMD Radeon RX 570X wins in:

• OpenCL compute: 82.8% higher score (38939 vs 21304).
• Texture throughput: 159.2 GTexel/s vs 73.49 GTexel/s.
• FP32 and FP16 compute: 5.095 TFLOPS in both vs 2.352 TFLOPS and 36.74 GFLOPS.
• Memory capacity and bandwidth: 8 GB / 224.0 GB/s vs 3 GB / 192.2 GB/s.
• Overall average benchmark score: 13871 vs 13470.
• Display connectivity: full set of outputs vs none.

NVIDIA P106-090 wins in:

• Pixel fill rate: 73.49 GPixel/s vs 39.81 GPixel/s.
• ROP count: 48 vs 32.
• Power efficiency: 75 W TDP and 250 W PSU vs 150 W and 450 W.
• Clock speeds: 1354 MHz base / 1531 MHz boost vs 1168 MHz / 1244 MHz.
• Vulkan API version: 1.4 vs 1.3.
• Smaller die size: 200 mm² vs 232 mm².