NVIDIA GeForce GTX 770 vs NVIDIA P106-090 Comparison

The NVIDIA GeForce GTX 770 and the NVIDIA P106-090 represent two very different approaches to GPU design, separated by four years of architecture evolution. The GTX 770 is a traditional, fully-featured consumer graphics card from the Kepler era, built for gaming and general-purpose compute. The P106-090 is a Pascal-based mining-specific card, stripped of display outputs and designed solely for computational throughput. Benchmark data shows a surprisingly close contest in average scores, but the two cards win in distinctly different workloads, making the choice heavily dependent on the intended use case.

FAQ

Q: Which card has the higher average benchmark score?

A: The NVIDIA GeForce GTX 770 has a higher average benchmark score of 14,747, compared to the NVIDIA P106-090's 13,470. This places the GTX 770 at the 57th percentile of all GPUs, while the P106-090 sits at the 54th percentile.

Q: How do the two cards compare in OpenCL performance?

A: The NVIDIA P106-090 is significantly faster in OpenCL, scoring 21,304 versus the GTX 770's 15,707. This represents a 26.3% lead for the P106-090, making it the clear winner in compute-heavy OpenCL workloads.

Q: Which card wins in Vulkan performance?

A: The NVIDIA GeForce GTX 770 wins in the Vulkan benchmark, scoring 19,166 against the P106-090's 18,596. The margin is narrow at 3.1%, showing that the older Kepler architecture can still hold its own in this modern API.

Q: What are the closest rivals to the GTX 770 based on average score?

A: The closest rival is the AMD Radeon RX 5500 XT, with an average score of 14,692, which is just 0.4% behind the GTX 770. Other close competitors include the NVIDIA Quadro M2000 at 14,532 (1.5% behind) and the NVIDIA GeForce GTX 660 Ti at 15,063 (2.1% ahead).

Q: What rivals does the P106-090 compare most closely to?

A: The P106-090's nearest rival is the NVIDIA GeForce GTX 570, which scores 13,515, a marginal 0.3% ahead. The AMD Radeon Pro 555 is 0.5% behind at 13,407, and the AMD Radeon HD 7770M is 0.5% ahead at 13,536.

Q: Does the P106-090 have any display outputs?

A: No, the NVIDIA P106-090 has no display outputs. It is designed exclusively for mining and compute tasks, whereas the GTX 770 offers a full set of display outputs including 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2.

Architecture Differences

The two cards are built on entirely different architectures from different process nodes. The GTX 770 uses the GK104 chip based on the Kepler architecture, manufactured on a 28 nm process at TSMC. The P106-090 uses the GP106 chip based on the Pascal architecture, also fabricated by TSMC but on a much more modern 16 nm process. This process shrink allows the Pascal chip to pack 4,400 million transistors into a 200 mm² die, achieving a transistor density of 22.0M per mm². In contrast, the Kepler chip contains 3,540 million transistors spread across a larger 294 mm² die, resulting in a lower density of 12.0M per mm².

The compute configurations differ dramatically. The GTX 770 features 1,536 shading units, 128 texture mapping units, and 32 render output units. The P106-090 has fewer shading units at 768, and only 48 texture mapping units, but it compensates with 48 render output units. This gives the P106-090 a higher pixel rate of 73.49 GPixel/s compared to the GTX 770's 34.72 GPixel/s. However, the GTX 770 has a much higher texture rate of 138.9 GTexel/s versus 73.49 GTexel/s for the P106-090. In raw FP32 compute, the GTX 770 delivers 3.333 TFLOPS, while the P106-090 provides 2.352 TFLOPS. The P106-090 also has a negligible FP16 capability of 36.74 GFLOPS at a 1:64 ratio.

Clock speeds show the architectural advantage of Pascal. The P106-090 has a base clock of 1354 MHz and a boost clock of 1531 MHz, significantly higher than the GTX 770's base of 1046 MHz and boost of 1085 MHz. API support also differs: the GTX 770 supports DirectX 12 (11_0), while the P106-090 supports DirectX 12 (12_1). Both cards support OpenGL 4.6, but the P106-090 supports Vulkan 1.4, whereas the GTX 770 tops out at Vulkan 1.2.175.

Where Each One Wins

The data clearly separates these two cards by workload type. The P106-090 dominates in OpenCL compute, with a 26.3% advantage over the GTX 770. This makes it the superior choice for general-purpose GPU computing, mining operations, or any task that leverages OpenCL for parallel processing. Its higher pixel rate and modern Pascal architecture contribute to this compute advantage.

The GTX 770, despite its age, wins in Vulkan performance by 3.1%. This suggests it is the better option for gaming and graphics workloads that utilize the Vulkan API. The GTX 770 also has the benefit of display outputs, making it a usable card for actual gaming systems, whereas the P106-090 cannot output video at all. For users who need a functional graphics card for gaming, the GTX 770 is the only viable option between the two.

In terms of memory configuration, the P106-090 has a larger memory capacity at 3 GB, but the GTX 770 has a wider 256-bit memory bus and higher bandwidth of 224.4 GB/s. The GTX 770's memory bandwidth advantage of 224.4 GB/s versus 192.2 GB/s for the P106-090 may help in texture-heavy gaming scenarios, even though the P106-090 has more memory total.

Specification Differences

The most striking difference is the display outputs. The GTX 770 has 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2, while the P106-090 has no outputs at all. This alone defines their intended purposes: one is a consumer graphics card, the other is a compute-only mining card.

Power requirements also diverge significantly. The GTX 770 has a TDP of 230 W and requires a 550 W suggested power supply, along with both a 6-pin and an 8-pin power connector. The P106-090 is far more efficient, with a TDP of just 75 W, a 250 W suggested power supply, and only a single 6-pin connector. This makes the P106-090 a much easier card to install in power-constrained systems.

Memory configuration differs in capacity, bus width, and speed. The GTX 770 has 2 GB of GDDR5 memory on a 256-bit bus, while the P106-090 has 3 GB of GDDR5 on a 192-bit bus. Memory clocks also differ, with the GTX 770 running at 1753 MHz (7 Gbps effective) and the P106-090 at 2002 MHz (8 Gbps effective). The resulting bandwidth is 224.4 GB/s for the GTX 770 and 192.2 GB/s for the P106-090.

The bus interface is another key differentiator. The GTX 770 uses PCIe 3.0 x16, while the P106-090 is limited to PCIe 1.0 x1. This makes the P106-090 severely bandwidth-constrained for data transfer to and from the host system, which could bottleneck many workloads despite its compute strengths.

Head-to-Head Benchmarks

The two head-to-head benchmarks available show a split decision. In Geekbench OpenCL, the P106-090 scores 21,304 against the GTX 770's 15,707, giving the Pascal card a decisive 26.3% victory. This is the largest performance gap between the two cards and highlights the P106-090's compute-focused design.

In Geekbench Vulkan, the GTX 770 fights back with a score of 19,166, edging out the P106-090's 18,596. The 3.1% margin is modest, but it demonstrates that the older Kepler architecture can still outperform the newer Pascal chip in certain graphics-oriented APIs. This result is particularly notable given that the P106-090 has higher clock speeds and a more modern architecture.

The average benchmark scores also tell a story. The GTX 770's average of 14,747 places it 9.5% ahead of the P106-090's 13,470. This suggests that across a broader range of tests, the GTX 770 is the more consistent performer, despite losing the OpenCL test by a wide margin. The GTX 770's nearest rival in terms of average score is the AMD Radeon RX 5500 XT at 14,692, just 0.4% behind, while the P106-090's closest competitor is the NVIDIA GeForce GTX 570 at 13,515, which is 0.3% ahead.

The Verdict

The choice between these two cards is dictated entirely by the use case, as the data shows they excel in different domains. The NVIDIA GeForce GTX 770 is the card to pick for anyone needing a functional graphics card with display outputs. It wins the Vulkan benchmark by 3.1%, has a higher average benchmark score of 14,747, and offers full display connectivity with 2x DVI, HDMI, and DisplayPort. Its 224.4 GB/s memory bandwidth and 3.333 TFLOPS FP32 compute make it a capable gaming and general-purpose GPU, despite its higher 230 W power draw.

The NVIDIA P106-090 is the specialist tool for compute-heavy OpenCL workloads. Its 26.3% victory in OpenCL shows a clear advantage for parallel compute tasks, mining, or any headless compute environment. The 75 W TDP makes it far easier to power, and the 3 GB memory capacity is larger, but the lack of display outputs and the severely limited PCIe 1.0 x1 interface restrict its utility to dedicated compute tasks. Users who need raw OpenCL performance and do not require video output should choose the P106-090.

For gaming or any task requiring a display, the GTX 770 is the only rational choice. For pure compute, the P106-090's OpenCL advantage is compelling, but the PCIe 1.0 x1 interface may negate some of that advantage in system-level workloads. The GTX 770, with its higher average score, wider memory bus, and display support, is the more versatile card overall. The P106-090 wins where compute density and power efficiency matter most, but its single-purpose design limits its appeal.