AMD Radeon RX 570 vs NVIDIA P104-100 Comparison

Head-to-Head Benchmarks

The benchmark data is decisively in favor of the NVIDIA P104-100 across every recorded test. The largest gap appears in Geekbench OpenCL, where the P104-100 scores 52,368 against the Radeon RX 570’s 32,084, a 63.2% advantage. This is a substantial margin that reflects the P104-100’s raw compute throughput, and it places the card comfortably ahead in any workload that leverages OpenCL acceleration.

In 3DMark Steel Nomad DX12, the P104-100 again dominates, posting 1,413 points versus 880 for the Radeon RX 570. That is a 60.6% delta in favor of the NVIDIA card. This particular benchmark stresses modern DX12 rendering pipelines, and the result indicates that the P104-100’s Pascal architecture handles the workload far more efficiently than GCN 4.0. The RX 570 trails by more than half a thousand points, which is a significant chasm in a synthetic test that typically shows smaller relative differences.

The closest contest comes in Geekbench Vulkan, where the P104-100 scores 45,165 and the RX 570 scores 42,752. The delta here is just 5.6%, still a win for NVIDIA, but much narrower. This suggests that when both cards are pushed through Vulkan’s lower-level API, the architectural gap narrows considerably. The RX 570’s GCN design has historically been competent in Vulkan titles, and the data supports that reputation, but the P104-100 still holds the edge.

Across all three head-to-head benchmarks, the P104-100 wins 3, and the RX 570 wins 0. The average benchmark score from the database reinforces this: the P104-100 sits at 32,982, while the RX 570 averages 28,766. That is a 14.7% overall lead for the NVIDIA card. The P104-100 also lands in the 77th percentile among all GPUs in the database, while the RX 570 sits at the 74th percentile. Both are respectable positions, but the P104-100 is clearly the stronger performer in aggregate.

Looking at the nearest rivals in the database provides context. The P104-100’s closest competitor is the NVIDIA T600 Mobile, with an average score of 32,849 and a delta of just 0.4%. That means the P104-100 is virtually tied with that mobile workstation card. Meanwhile, the RX 570’s nearest rival is the AMD Radeon RX 6800M, which scores 28,874 and trails by only 0.4%. The RX 570 also sits near the Radeon R9 M295X and the Radeon RX 470, with deltas of 0.7% and -0.8% respectively. These comparisons show that both cards are clustered tightly with their peers, but the P104-100’s cluster sits several thousand points higher.

Where Each One Wins

The NVIDIA P104-100 wins every benchmark category recorded in the database. There is no test in which the Radeon RX 570 comes out ahead. That means for users prioritizing pure compute performance in OpenCL, the P104-100 is the unambiguous choice. The 63.2% lead in Geekbench OpenCL is the largest margin of any test, making this card particularly well-suited for compute-heavy tasks such as rendering, simulation, or any workload that offloads parallel math to the GPU.

In DX12 gaming workloads, represented by 3DMark Steel Nomad, the P104-100 also holds a commanding 60.6% advantage. This is a modern, demanding test, and the data suggests that the P104-100 will deliver significantly higher frame rates in current DX12 titles. The RX 570, while capable, is simply outclassed here. Gamers running DX12 games should expect the P104-100 to pull far ahead.

The Vulkan result is the only area where the RX 570 shows relative strength. Its score of 42,752 is 5.6% behind the P104-100’s 45,165, which is a modest gap. In Vulkan-based games or applications, the RX 570 would be much closer in performance, and in some titles the difference might be imperceptible. However, the data still gives the win to NVIDIA, so even in Vulkan the P104-100 is the faster card.

For users who rely on Metal, the RX 570 has an additional benchmark that the P104-100 lacks: Geekbench Metal, where it scores 39,349. This is not a head-to-head comparison since the NVIDIA card has no Metal result, but it indicates that the RX 570 can serve macOS or Metal-based environments. The P104-100 has no display outputs at all, which makes it unsuitable for any interactive use, while the RX 570 offers 1x DVI, 1x HDMI 2.0b, and 3x DisplayPort 1.4a outputs. So while the RX 570 loses every performance comparison, it wins on connectivity and platform flexibility.

The Verdict

The data is clear: the NVIDIA P104-100 is the faster GPU in every direct benchmark comparison. It wins 3 out of 3 head-to-head tests, with margins ranging from 5.6% to 63.2%. Its average benchmark score of 32,982 is 14.7% higher than the RX 570’s 28,766. Anyone selecting purely on performance should choose the P104-100 without hesitation.

However, the P104-100 is a mining-specific card with no display outputs. It cannot drive a monitor, so it is useless for gaming or any visual output task. The RX 570, by contrast, has full display connectivity and can be used as a daily driver for gaming and general computing. If the workload requires a visible output, the RX 570 is the only viable option despite its lower scores.

The P104-100 also carries architectural advantages. It uses the GP104 chip on a 16 nm TSMC process, while the RX 570 uses Polaris 20 on a 14 nm GlobalFoundries process. The P104-100 has a higher boost clock, faster memory, and more bandwidth. These factors explain its benchmark dominance. But the RX 570 has a lower TDP at 150 W and uses a single 6-pin power connector, while the P104-100 requires an 8-pin connector and lists a 200 W suggested PSU, which is a higher system requirement.

For compute-focused users who do not need display output, the P104-100 is the superior choice. For gamers or general users who need a functional graphics card, the RX 570 is the practical pick, even though it loses every recorded benchmark. The verdict depends entirely on the use case, but on raw numbers alone, the P104-100 wins.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA P104-100 has an average benchmark score of 32,982, while the AMD Radeon RX 570 averages 28,766, a 14.7% difference.

Q: What is the largest performance gap between the two cards?

A: The largest gap is in Geekbench OpenCL, where the P104-100 scores 52,368 and the RX 570 scores 32,084, a 63.2% delta in favor of NVIDIA.

Q: Is the RX 570 closer to the P104-100 in any benchmark?

A: Yes, in Geekbench Vulkan the RX 570 scores 42,752 versus 45,165 for the P104-100, a margin of only 5.6%.

Q: Can the P104-100 connect to a display?

A: No, the P104-100 has no display outputs. The RX 570 offers 1x DVI, 1x HDMI 2.0b, and 3x DisplayPort 1.4a outputs.

Q: What are the memory specifications for each card?

A: The P104-100 has 4 GB of GDDR5X on a 256-bit bus with 320.3 GB/s bandwidth. The RX 570 has 4 GB of GDDR5 on a 256-bit bus with 224.0 GB/s bandwidth.

Q: How do the cards rank among all GPUs in the database?

A: The P104-100 is in the 77th percentile, while the RX 570 is in the 74th percentile.

Architecture Differences

The two cards come from fundamentally different architectural lineages. The NVIDIA P104-100 is built on the Pascal architecture, using the GP104 chip fabricated on a 16 nm process at TSMC. The AMD Radeon RX 570 uses the GCN 4.0 architecture with the Polaris 20 chip, produced on a 14 nm process at GlobalFoundries. The transistor counts differ significantly: the P104-100 packs 7,200 million transistors on a 314 mm² die, while the RX 570 has 5,700 million transistors on a 232 mm² die. This gives the NVIDIA chip a lower transistor density of 22.9M per mm² compared to the AMD chip’s 24.6M per mm².

Clock speeds also favor the P104-100. Its base clock is 1,607 MHz with a boost of 1,733 MHz, while the RX 570 runs at 1,168 MHz base and 1,244 MHz boost. The NVIDIA card’s memory is rated at 1,251 MHz with 10 Gbps effective speed, while the RX 570’s memory runs at 1,750 MHz with 7 Gbps effective. The memory type differs as well: GDDR5X on the P104-100 versus GDDR5 on the RX 570.

Compute resources are distributed differently. The P104-100 has 1,920 shading units, 120 texture mapping units, and 64 ROPs. The RX 570 has 2,048 shading units, 128 TMUs, and 32 ROPs. Despite having fewer shaders and TMUs, the P104-100 achieves higher pixel and texture rates: 110.9 GPixel/s and 208.0 GTexel/s, versus 39.81 GPixel/s and 159.2 GTexel/s for the RX 570. The FP32 throughput also favors NVIDIA at 6.655 TFLOPS versus 5.095 TFLOPS. In FP16, the P104-100 is severely limited at 104.0 GFLOPS (1:64 ratio), while the RX 570 delivers full-rate FP16 at 5.095 TFLOPS (1:1 ratio).

API support shows minor differences. Both support DirectX 12, but the P104-100 supports version 12_1 while the RX 570 supports 12_0. OpenGL is 4.6 on both. Vulkan support is 1.4 on the P104-100 and 1.3 on the RX 570. Neither card has ray tracing cores or tensor cores. The bus interface differs notably: the P104-100 uses PCIe 1.0 x4, while the RX 570 uses PCIe 3.0 x16. This is a significant bottleneck for the NVIDIA card in terms of data transfer, though it does not appear to hamper its benchmark results.

Specification Differences

The following specifications differ between the two cards. The P104-100 uses the GP104 chip, while the RX 570 uses Polaris 20. The P104-100 is on a 16 nm process from TSMC, the RX 570 on 14 nm from GlobalFoundries. Transistor counts are 7,200 million versus 5,700 million, and die sizes are 314 mm² versus 232 mm². Transistor density is 22.9M per mm² versus 24.6M per mm².

Base clocks are 1,607 MHz versus 1,168 MHz, and boost clocks are 1,733 MHz versus 1,244 MHz. Memory clock is 1,251 MHz with 10 Gbps effective for the P104-100, versus 1,750 MHz with 7 Gbps effective for the RX 570. Memory type is GDDR5X versus GDDR5. Bandwidth is 320.3 GB/s versus 224.0 GB/s. Shading units are 1,920 versus 2,048, TMUs are 120 versus 128, and ROPs are 64 versus 32. Pixel rate is 110.9 GPixel/s versus 39.81 GPixel/s, and texture rate is 208.0 GTexel/s versus 159.2 GTexel/s. FP32 is 6.655 TFLOPS versus 5.095 TFLOPS, and FP16 is 104.0 GFLOPS (1:64) versus 5.095 TFLOPS (1:1).

TDP is not listed for the P104-100, while the RX 570 has a TDP of 150 W. Power connectors are 1x 8-pin versus 1x 6-pin. Suggested PSU is 200 W versus 450 W. Bus interface is PCIe 1.0 x4 versus PCIe 3.0 x16. Display outputs are none versus 1x DVI, 1x HDMI 2.0b, and 3x DisplayPort 1.4a. DirectX support is 12 (12_1) versus 12 (12_0), and Vulkan is 1.4 versus 1.3. Dimensions are 267 mm (10.5 inches) versus 241 mm (9.5 inches). Release dates are 2017-12-11 versus 2017-04-17. The RX 570 has a launch MSRP of 169 USD, while the P104-100 has none listed. The RX 570 lists a predecessor (Arctic Islands) and successor (Vega), while the P104-100 has neither.