NVIDIA GeForce GTX 1080 vs NVIDIA P106-090 Comparison

Head-to-Head Benchmarks

The recorded data leaves little room for ambiguity. Across every shared benchmark test, the NVIDIA GeForce GTX 1080 posts a decisive victory over the NVIDIA P106-090. The largest margin appears in the 3DMark Steel Nomad DX12 test, where the GTX 1080 scores 1560 against the P106-090’s 509, a gap of 67.4%. That is not a close contest; it is a full performance tier apart. The P106-090 trails by more than two-thirds in this modern DirectX 12 workload, suggesting the GTX 1080’s additional shading units and memory bandwidth translate directly into frame-rendering capability.

The Geekbench OpenCL test tells a similar story, though the percentage gap narrows slightly. The GTX 1080 records 51204 points, while the P106-090 manages 21304, a delta of 58.4%. This workload exercises general-purpose compute rather than graphics-specific paths, and the GTX 1080 still maintains a commanding lead. The P106-090’s 2.352 TFLOPS of FP32 throughput simply cannot match the GTX 1080’s 8.873 TFLOPS, and the benchmark results reflect that raw arithmetic advantage.

The closest margin of the three appears in Geekbench Vulkan, where the GTX 1080 scores 30398 and the P106-090 scores 18596. That is a 38.8% difference, still a substantial victory but notably smaller than the other two tests. Vulkan’s lower-level API overhead may allow the P106-090’s architecture to express itself relatively better, though it remains firmly behind. The P106-090 wins zero head-to-head comparisons; the GTX 1080 wins all three. When interpreting these numbers, the pattern is consistent: the GTX 1080 is categorically faster in every measured dimension, with the smallest delta still representing a significant performance gap.

Architecture Differences

Both cards are built on NVIDIA’s Pascal architecture and fabricated on TSMC’s 16 nm process, but the silicon inside is quite different. The P106-090 uses the GP106 chip, a die measuring 200 mm² with 4,400 million transistors. The GTX 1080 uses the GP104 chip, a substantially larger 314 mm² die with 7,200 million transistors. Transistor density is nearly identical, 22.0M per mm² for the P106-090 and 22.9M per mm² for the GTX 1080, indicating the process node is being used with similar efficiency. The difference lies in scale, not density: the GP104 packs roughly 64% more transistors into a die that is 57% larger.

The compute resources diverge sharply. The P106-090 has 768 shading units, 48 texture mapping units, and 48 ROPs. The GTX 1080 has 2560 shading units, 160 TMUs, and 64 ROPs. That is more than triple the shader count and more than triple the texture units. Pixel rate and texture rate follow accordingly: the GTX 1080 reaches 110.9 GPixel/s and 277.3 GTexel/s, while the P106-090 posts 73.49 GPixel/s and 73.49 GTexel/s. The texture rate gap is particularly telling, 277.3 versus 73.49, a 3.8x difference, because it reflects the GTX 1080’s far larger TMU array.

Memory configuration is another fundamental split. The P106-090 carries 3 GB of GDDR5 on a 192-bit bus, yielding 192.2 GB/s of bandwidth. The GTX 1080 has 8 GB of GDDR5X on a 256-bit bus, delivering 320.3 GB/s. The memory type also matters: GDDR5X is a higher-bandwidth technology than GDDR5, and the GTX 1080’s effective memory clock is 10 Gbps versus the P106-090’s 8 Gbps. The P106-090’s smaller bus and slower memory cap its ability to feed the shading units, which is likely why its texture rate lags so far behind despite a similar ROP count.

Clock speeds also favor the GTX 1080. Its base clock is 1607 MHz and boost clock is 1733 MHz, while the P106-090 runs at 1354 MHz base and 1531 MHz boost. The GTX 1080’s higher clocks compound with its larger shader array to produce a much higher FP32 throughput, 8.873 TFLOPS versus 2.352 TFLOPS. FP16 performance is similarly lopsided: 138.6 GFLOPS for the GTX 1080 versus 36.74 GFLOPS for the P106-090, both at a 1:64 ratio. Neither card has ray tracing cores or tensor cores, so those fields are absent from both.

The most striking architectural difference is the P106-090’s interface and output situation. It uses PCIe 1.0 x1 and has no display outputs whatsoever. The GTX 1080 uses PCIe 3.0 x16 and offers 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.4a. The P106-090 is explicitly a mining GPU, designed to compute without rendering to a screen. Its PCIe 1.0 x1 interface is a bottleneck for data transfer in general-purpose computing, but it is sufficient for mining workloads where the GPU is mostly self-contained. The GTX 1080, by contrast, is a fully-featured consumer graphics card with standard connectivity.

Where Each One Wins

The P106-090’s only winning scenarios are those where display output is irrelevant and power draw is a primary concern. Its TDP is 75 W, compared to the GTX 1080’s 180 W, and its suggested PSU is 250 W versus 450 W. For a mining rig that prioritizes efficiency per watt, the P106-090 could be attractive, though the benchmark data does not include a power efficiency metric. The card’s 3 GB memory and 192-bit bus are enough for certain compute tasks, but its PCIe 1.0 x1 interface and lack of outputs mean it cannot serve as a general-purpose desktop GPU.

The GTX 1080 wins every measured benchmark, and its use cases follow from that dominance. It is a gaming card with full display outputs, high FP32 throughput, and 8 GB of GDDR5X memory. The 3DMark Steel Nomad DX12 result, 1560 versus 509, indicates strong modern gaming performance. The Geekbench OpenCL and Vulkan scores, 51204 and 30398 respectively, point to capable compute performance in APIs that are common in both gaming and professional applications. The GTX 1080’s 64 ROPs and 320.3 GB/s bandwidth also make it better suited for high-resolution rendering and texture-heavy workloads.

The P106-090’s average benchmark score of 13470 places it at the 54th percentile of all GPUs, while the GTX 1080’s average score of 11960 places it at the 51st percentile. This is a curious inversion: the GTX 1080 wins every head-to-head test decisively, yet its overall average is lower and its percentile rank is lower. The explanation lies in the benchmark pool. The GTX 1080 has many more recorded tests, including Passmark DirectX 9, 10, 11, 12, G2D, G3D, and GPU compute results, some of which are low (Passmark DirectX 12 scores only 55). The P106-090 has only three benchmark entries, all of which are relatively strong, so its average is skewed upward by the limited sample. The head-to-head comparisons are the more reliable indicator of relative performance.

FAQ

Q: Which card has a higher average benchmark score?

A: The P106-090 has an average benchmark score of 13470, while the GTX 1080 has an average of 11960. However, the GTX 1080 wins all three shared head-to-head tests, indicating its lower average is due to a broader set of benchmark entries, including older DirectX tests.

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

A: In the 3DMark Steel Nomad DX12 test, the GTX 1080 scores 1560 against the P106-090’s 509, a delta of 67.4%. This is the largest margin recorded in the head-to-head data.

Q: Do both cards support the same graphics APIs?

A: Yes, both list DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. Neither card has ray tracing or tensor cores.

Q: Can the P106-090 be used for display output?

A: No. The P106-090 has no display outputs, while the GTX 1080 offers 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.4a.

Q: How do the memory configurations differ?

A: The P106-090 has 3 GB of GDDR5 on a 192-bit bus with 192.2 GB/s bandwidth. The GTX 1080 has 8 GB of GDDR5X on a 256-bit bus with 320.3 GB/s bandwidth.

Q: Which card has more shading units?

A: The GTX 1080 has 2560 shading units, while the P106-090 has 768. The GTX 1080 also has 160 TMUs versus 48, and 64 ROPs versus 48.

Specification Differences

The two cards differ across nearly every major specification field. The GTX 1080 uses the GP104 chip with 7,200 million transistors on a 314 mm² die, while the P106-090 uses the GP106 chip with 4,400 million transistors on a 200 mm² die. The GTX 1080’s base clock is 1607 MHz and boost clock is 1733 MHz; the P106-090’s base is 1354 MHz and boost is 1531 MHz. The GTX 1080 has 8 GB of GDDR5X memory on a 256-bit bus with 320.3 GB/s bandwidth; the P106-090 has 3 GB of GDDR5 on a 192-bit bus with 192.2 GB/s bandwidth.

The compute unit counts are the most lopsided difference: 2560 shading units, 160 TMUs, and 64 ROPs for the GTX 1080 versus 768 shading units, 48 TMUs, and 48 ROPs for the P106-090. Pixel rate is 110.9 GPixel/s versus 73.49 GPixel/s, and texture rate is 277.3 GTexel/s versus 73.49 GTexel/s. FP32 throughput is 8.873 TFLOPS versus 2.352 TFLOPS, and FP16 is 138.6 GFLOPS versus 36.74 GFLOPS.

Power requirements also differ substantially. The GTX 1080 has a TDP of 180 W and requires a 450 W PSU and a 1x 8-pin connector. The P106-090 has a TDP of 75 W, needs a 250 W PSU, and uses a 1x 6-pin connector. The bus interface is another stark split: PCIe 3.0 x16 for the GTX 1080 versus PCIe 1.0 x1 for the P106-090. The GTX 1080 has full display outputs and a launch MSRP of 599 USD. The P106-090 has no display outputs and no launch MSRP in the database. Physical dimensions differ as well: the GTX 1080 is 267 mm long, 112 mm high, and 40 mm wide; the P106-090 is 250 mm long.

The Verdict

The data is unambiguous. The NVIDIA GeForce GTX 1080 is the superior graphics card in every measured benchmark, with margins ranging from 38.8% to 67.4%. It has more shading units, more texture units, more ROPs, more memory, higher bandwidth, higher clocks, and higher FP32 throughput. It also provides standard display outputs and a normal PCIe 3.0 x16 interface. Any user needing a card for gaming, rendering, or general compute should select the GTX 1080 without hesitation.

The NVIDIA P106-090 is a niche product. Its only advantages are lower power draw (75 W versus 180 W) and a smaller physical footprint (250 mm versus 267 mm). Its lack of display outputs makes it unsuitable as a desktop GPU, and its PCIe 1.0 x1 interface limits data transfer speeds. The benchmark results show it trailing badly in every shared test, and its nearest rivals include the GTX 570 and Radeon Pro 555, cards from an older era. The GTX 1080’s nearest rivals, by contrast, are the GTX 1660 and RX 6500 XT, indicating it competes in a higher performance bracket despite its lower average score.

The only scenario where the P106-090 makes sense is a dedicated compute setup where power efficiency is paramount and display output is unnecessary. Even then, the GTX 1080’s 3.8x texture rate advantage and 3.8x FP32 advantage suggest it would complete compute tasks far faster, potentially offsetting its higher power draw. For virtually all purposes, the GTX 1080 is the clear choice. The P106-090 is a specialized mining artifact, while the GTX 1080 is a general-purpose performer that wins every head-to-head test recorded in the database.