NVIDIA GeForce GTX 570 vs NVIDIA P106-090 Comparison

The NVIDIA GeForce GTX 570 and NVIDIA P106-090 are two very different GPUs that land at nearly the same overall benchmark ranking, yet their performance profiles could not be more divergent. The data shows the P106-090 wins the only direct head-to-head benchmark by a massive 36.6%, but the GTX 570 holds its own in the aggregate scoring. This comparison is less about which card is "better" and more about which legacy workload each was designed to serve.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 570 has an average benchmark score of 13515, while the NVIDIA P106-090 scores 13470. The difference is a mere 0.3% in favor of the GTX 570, placing both cards at the 54th percentile of all GPUs.

Q: How do the two cards compare in the Geekbench OpenCL benchmark?

A: The NVIDIA P106-090 wins decisively with a score of 21304, compared to the GTX 570's 13515. This represents a 36.6% advantage for the P106-090, which is the only head-to-head benchmark available in the data.

Q: What are the closest rivals to these cards?

A: For the GTX 570, the nearest rival is the AMD Radeon HD 7770M with an average score of 13536 (0.2% ahead). For the P106-090, the closest competitor is the AMD Radeon Pro 555 at 13407 (0.5% behind). Both cards also sit near the AMD Radeon RX 9070 XT, which scores 13543.

Q: What is the power consumption difference between the two?

A: The GTX 570 has a TDP of 219 W and requires a 550 W suggested power supply, while the P106-090 has a TDP of just 75 W and only needs a 250 W suggested PSU. The P106-090 also uses a single 6-pin power connector versus the GTX 570's two 6-pin connectors.

Q: Do both cards support the same DirectX version?

A: No. The GTX 570 supports DirectX 12 (11_0), while the P106-090 supports DirectX 12 (12_1). The P106-090 also adds Vulkan 1.4 support, which the GTX 570 lacks entirely.

Q: Which card has more memory and bandwidth?

A: The P106-090 has 3 GB of GDDR5 memory on a 192-bit bus, delivering 192.2 GB/s of bandwidth. The GTX 570 has 1280 MB of GDDR5 on a 320-bit bus, providing 152.0 GB/s. Despite the narrower bus, the P106-090 offers more capacity and higher bandwidth.

The Verdict

The data splits this pairing into two clear camps. If the task is raw compute through OpenCL, the NVIDIA P106-090 is the unequivocal choice — its 21304 score versus 13515 represents a 36.6% lead that no other metric in the pack can offset. The P106-090 also brings modern API support with DirectX 12 (12_1) and Vulkan 1.4, along with triple the memory capacity at 3 GB.

However, the GTX 570 is not obsolete in comparison. Its average benchmark score of 13515 edges out the P106-090's 13470, and it matches the P106-090's 54th percentile ranking. The GTX 570 was a flagship product from the GeForce 500 generation with a 349 USD launch MSRP, whereas the P106-090 was a mining-specific card with no display outputs. For anyone needing video output, the GTX 570 is the only functional option, as the P106-090 has no outputs whatsoever.

The verdict is context-dependent. For compute-heavy, headless workloads — likely mining or server-side tasks — the P106-090 wins on performance and efficiency. For general desktop use, legacy gaming, or any scenario requiring a display connection, the GTX 570 is the only viable pick, despite its older architecture and higher power draw.

Head-to-Head Benchmarks

The single head-to-head benchmark in the data is Geekbench OpenCL, and it is a landslide. The NVIDIA P106-090 scores 21304, while the NVIDIA GeForce GTX 570 manages only 13515. The delta is -36.6% from the perspective of the GTX 570, meaning the P106-090 outperforms it by more than a third in this compute test.

This result aligns with the architectural gulf between the two cards. The P106-090 has 768 shading units versus 480 on the GTX 570, and its FP32 throughput is 2.352 TFLOPS compared to 1,405.4 GFLOPS. The P106-090 also benefits from a 16 nm process node versus 40 nm, which explains its ability to deliver higher clock speeds — 1354 MHz base and 1531 MHz boost — compared to the GTX 570's unspecified base and boost clocks.

Beyond the head-to-head, the aggregate scores tell a different story. The GTX 570's average of 13515 nearly matches the P106-090's 13470, suggesting that in other benchmark suites — not included in the head-to-head data — the older card is competitive. The P106-090's win is isolated to OpenCL, where its compute-oriented design shines.

Specification Differences

The two cards diverge sharply on nearly every specification. The GTX 570 uses the GF110 chip on a 40 nm process, while the P106-090 uses the GP106 chip on a 16 nm process. Transistor counts differ: 3,000 million for the GTX 570 versus 4,400 million for the P106-090, though the die sizes are 520 mm² and 200 mm² respectively, yielding transistor densities of 5.8M / mm² and 22.0M / mm².

Memory configurations are distinct. The GTX 570 has 1280 MB of GDDR5 on a 320-bit bus with 152.0 GB/s bandwidth, while the P106-090 has 3 GB on a 192-bit bus with 192.2 GB/s. The P106-090's memory clock is 2002 MHz (8 Gbps effective) versus the GTX 570's 950 MHz (3.8 Gbps effective).

Compute resources favor the P106-090 in shading units (768 vs 480) and ROPs (48 vs 40), but the GTX 570 has more TMUs (60 vs 48). Pixel rates are 73.49 GPixel/s for the P106-090 versus 21.96 GPixel/s for the GTX 570, and texture rates are 73.49 GTexel/s versus 43.92 GTexel/s.

Power requirements are drastically different: 75 W TDP with a 250 W PSU suggestion for the P106-090, versus 219 W TDP with a 550 W PSU for the GTX 570. The P106-090 uses one 6-pin connector; the GTX 570 uses two. The P106-090 also has no display outputs, while the GTX 570 offers 2x DVI and 1x mini-HDMI 1.3a.

Architecture Differences

The GTX 570 is built on Fermi 2.0 architecture, representing the GeForce 500 generation. The P106-090 is built on Pascal, classified under Mining GPUs. This is a generational leap: Fermi 2.0 was released in 2010, while Pascal arrived in 2017.

The process node difference is stark — 40 nm for Fermi versus 16 nm for Pascal. This explains the P106-090's higher clock speeds and dramatically better performance-per-watt, as evidenced by the 75 W TDP versus 219 W. The P106-090 also has a significantly higher transistor density at 22.0M / mm² versus 5.8M / mm².

API support reflects the architectural age. The GTX 570 tops out at DirectX 12 (11_0) and OpenGL 4.6, with no Vulkan support. The P106-090 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The P106-090 also has a 1:64 FP16 ratio at 36.74 GFLOPS, whereas the GTX 570 has no listed FP16 capability.

Bus interfaces differ as well: the GTX 570 uses PCIe 2.0 x16, while the P106-090 uses PCIe 1.0 x1 — a peculiar downgrade that reflects its mining-specific design, where bandwidth to the host is less critical. Physical dimensions are similar: 267 mm for the GTX 570 versus 250 mm for the P106-090, both dual-slot.

Where Each One Wins

The NVIDIA P106-090 wins in every compute-heavy scenario. Its 36.6% lead in OpenCL makes it the clear choice for general-purpose GPU compute, and its 3 GB memory capacity allows larger datasets than the GTX 570's 1280 MB. The P106-090's higher bandwidth (192.2 GB/s vs 152.0 GB/s) and faster pixel rate (73.49 GPixel/s vs 21.96 GPixel/s) reinforce its advantage in throughput-bound tasks. Its 75 W TDP also makes it far easier to power and cool, requiring only a 250 W PSU.

The NVIDIA GeForce GTX 570 wins in any scenario requiring display output. The P106-090 has no outputs, making it useless for desktop use, gaming, or any visual workload. The GTX 570's 2x DVI and mini-HDMI 1.3a outputs allow it to drive monitors, and its 480 shading units and 60 TMUs provide adequate legacy gaming performance. Its 349 USD launch MSRP positioned it as a high-end card in its era, and its higher average benchmark score (13515 vs 13470) suggests it remains competitive in non-OpenCL tests.

The GTX 570 also wins on memory bus width (320-bit vs 192-bit), which can benefit certain bandwidth-sensitive workloads, even though the P106-090's overall bandwidth is higher. For users with older PCIe slots, the GTX 570's PCIe 2.0 x16 interface is more universally compatible than the P106-090's PCIe 1.0 x1, which could bottleneck data transfer in some systems.

In summary: pick the P106-090 for headless compute and efficiency; pick the GTX 570 for anything requiring a display or legacy system compatibility.