NVIDIA GeForce GTX 780M vs NVIDIA P106-090 Comparison

FAQ

Q: Which GPU is faster in OpenCL compute workloads?

A: The NVIDIA P106-090 is substantially ahead. In Geekbench OpenCL, it scores 21304 versus 12769 for the GeForce GTX 780M, a 66.8% advantage.

Q: How do the two compare in Vulkan performance?

A: The P106-090 wins again. It records 18596 in Geekbench Vulkan, while the GTX 780M scores 12696, putting the P106-090 46.5% higher.

Q: What is the average benchmark score for each GPU?

A: The P106-090 has a higher average benchmark score at 13470, compared to 11261 for the GTX 780M. The P106-090 sits at the 54th percentile of all GPUs in the database, while the GTX 780M sits at the 50th percentile.

Q: Which GPU has more memory and a wider memory bus?

A: The GTX 780M has 4 GB of GDDR5 memory on a 256-bit bus. The P106-090 has 3 GB on a 192-bit bus. However, the P106-090's memory bandwidth is higher at 192.2 GB/s versus 160.0 GB/s.

Q: What are the architectural generations of these two cards?

A: The P106-090 uses the Pascal architecture with the GP106 chip, built on TSMC's 16 nm process. The GTX 780M uses the older Kepler architecture with the GK104 chip, built on a 28 nm process.

Q: Which GPU has a higher transistor density?

A: The P106-090 packs 4,400 million transistors into a 200 mm² die, resulting in a density of 22.0M transistors per mm². The GTX 780M contains 3,540 million transistors on a larger 294 mm² die, giving a density of 12.0M per mm².

Architecture Differences

The two cards come from different NVIDIA design eras. The P106-090 is built on the Pascal architecture, using the GP106 chip, while the GTX 780M is a Kepler part based on the GK104 chip. This generational gap shows up in nearly every architectural metric.

The process node tells a clear story. The P106-090 is fabricated on TSMC's 16 nm process, whereas the GTX 780M uses the older 28 nm node. That process advantage allows the P106-090 to fit 4,400 million transistors into a 200 mm² die. The GTX 780M, by contrast, has 3,540 million transistors spread across a much larger 294 mm² die. Transistor density reflects this: the P106-090 reaches 22.0M transistors per mm², while the GTX 780M manages only 12.0M per mm².

Clock speeds differ dramatically. The P106-090 runs at a base clock of 1354 MHz and a boost clock of 1531 MHz. The GTX 780M is far lower, with a base of 771 MHz and a boost of 797 MHz. This clock advantage is a major reason the newer card dominates in compute benchmarks.

Shader configuration is another contrast. The GTX 780M has more shading units, 1536 versus 768 for the P106-090, and more texture mapping units, 128 versus 48. However, the P106-090 has more ROPs, 48 versus 32. The GTX 780M's higher pixel rate, 25.50 GPixel/s, does not match the P106-090's 73.49 GPixel/s, which indicates the clock speed advantage more than compensates for lower raw unit counts.

Memory architecture differs as well. The GTX 780M uses a 256-bit bus with 4 GB of GDDR5, while the P106-090 uses a 192-bit bus with 3 GB. Yet the P106-090 achieves higher bandwidth, 192.2 GB/s versus 160.0 GB/s, because its memory runs at 8 Gbps effective versus 5 Gbps effective on the older card.

Feature support shows the newer architecture. The P106-090 supports DirectX 12 (12_1) and Vulkan 1.4, while the GTX 780M tops out at DirectX 12 (11_0) and Vulkan 1.2.175. The P106-090 also has a 1:64 FP16 ratio, delivering 36.74 GFLOPS, while the GTX 780M has no recorded FP16 capability.

Form factor and power are also distinct. The P106-090 is a dual-slot card with a 75 W TDP and a single 6-pin power connector. The GTX 780M is an MXM module with a 122 W TDP and no power connectors, designed for portable devices. The P106-090 has no display outputs, while the GTX 780M's outputs are dependent on the host laptop.

Head-to-Head Benchmarks

The database records two shared benchmark tests between these GPUs, and the P106-090 wins both. The margin is not close.

In Geekbench OpenCL, the P106-090 scores 21304 against the GTX 780M's 12769. That is a 66.8% difference, a massive gap for a compute workload. The P106-090's higher clock speeds and newer architecture clearly drive this result. The GTX 780M's higher shading unit count, 1536 versus 768, cannot overcome the clock and process advantages of the Pascal part.

In Geekbench Vulkan, the P106-090 again leads, scoring 18596 versus 12696. The delta here is 46.5%. Vulkan is a lower-level API, and the P106-090's newer driver support and architectural efficiency show through. The GTX 780M's older Kepler design, with its lower Vulkan version support, trails by a wide margin.

The overall benchmark averages line up with these individual results. The P106-090 has an average score of 13470, while the GTX 780M averages 11261. That is a 19.6% gap in aggregate performance.

Looking at the nearest rivals in the database adds context. The P106-090 sits within 0.5% of the AMD Radeon Pro 555 and within 0.3% of the NVIDIA GeForce GTX 570. The GTX 780M, by contrast, is within 0.3% of the AMD Radeon Pro WX 3200 and the AMD FirePro W4300. The P106-090 is competitive with desktop-class cards, while the GTX 780M aligns more closely with workstation mobile parts.

The percentile ranking reinforces this. The P106-090 is at the 54th percentile of all GPUs, above the GTX 780M's 50th percentile. Neither card is a top-tier performer, but the P106-090 holds a clear edge in the database's measured tests.

Specification Differences

The two GPUs differ across nearly every specification field recorded in the database.

• Architecture: P106-090 uses Pascal; GTX 780M uses Kepler.
• Chip: P106-090 uses GP106; GTX 780M uses GK104.
• Process Node: P106-090 is 16 nm; GTX 780M is 28 nm.
• Transistors: P106-090 has 4,400 million; GTX 780M has 3,540 million.
• Die Size: P106-090 is 200 mm²; GTX 780M is 294 mm².
• Transistor Density: P106-090 is 22.0M per mm²; GTX 780M is 12.0M per mm².
• Base Clock: P106-090 is 1354 MHz; GTX 780M is 771 MHz.
• Boost Clock: P106-090 is 1531 MHz; GTX 780M is 797 MHz.
• Memory Clock: P106-090 is 8 Gbps effective; GTX 780M is 5 Gbps effective.
• Memory Size: P106-090 has 3 GB; GTX 780M has 4 GB.
• Memory Bus: P106-090 is 192-bit; GTX 780M is 256-bit.
• Memory Bandwidth: P106-090 is 192.2 GB/s; GTX 780M is 160.0 GB/s.
• Shading Units: P106-090 has 768; GTX 780M has 1536.
• TMUs: P106-090 has 48; GTX 780M has 128.
• ROPs: P106-090 has 48; GTX 780M has 32.
• Pixel Rate: P106-090 is 73.49 GPixel/s; GTX 780M is 25.50 GPixel/s.
• Texture Rate: P106-090 is 73.49 GTexel/s; GTX 780M is 102.0 GTexel/s.
• FP32: P106-090 is 2.352 TFLOPS; GTX 780M is 2.448 TFLOPS.
• FP16: P106-090 is 36.74 GFLOPS; GTX 780M has no recorded value.
• TDP: P106-090 is 75 W; GTX 780M is 122 W.
• Slot Width: P106-090 is dual-slot; GTX 780M is MXM Module.
• Power Connectors: P106-090 has 1x 6-pin; GTX 780M has none.
• Bus Interface: P106-090 is PCIe 1.0 x1; GTX 780M is MXM-B (3.0).
• Display Outputs: P106-090 has none; GTX 780M is portable device dependent.
• DirectX Support: P106-090 is 12 (12_1); GTX 780M is 12 (11_0).
• Vulkan Support: P106-090 is 1.4; GTX 780M is 1.2.175.
• Dimensions: P106-090 is 250 mm long; GTX 780M has no recorded length.

Where Each One Wins

The P106-090 is the clear winner in compute-oriented tasks. Its 66.8% lead in OpenCL and 46.5% lead in Vulkan make it the better choice for any workload that relies on those APIs. The newer Pascal architecture, much higher clock speeds, and superior memory bandwidth all contribute to this dominance. The card's higher average benchmark score, 13470 versus 11261, confirms it as the stronger all-around performer in the database's measurements.

The P106-090 also wins on efficiency. Its 75 W TDP is significantly lower than the GTX 780M's 122 W, despite delivering better benchmark results. This makes it the more sensible option for systems where power draw matters. Its higher pixel rate, 73.49 GPixel/s, also suggests strong fill-rate-dependent performance.

The GTX 780M has some advantages, though they are narrow. It offers 4 GB of memory versus 3 GB, which could matter for workloads with larger memory footprints, though its lower bandwidth may offset that benefit. It also has more shading units, 1536 versus 768, and more TMUs, 128 versus 48. Its texture rate of 102.0 GTexel/s is higher than the P106-090's 73.49 GTexel/s, which could help in texture-heavy rendering scenarios. Its FP32 output, 2.448 TFLOPS, is slightly higher than the P106-090's 2.352 TFLOPS, though real-world tests show the newer card pulls ahead.

The GTX 780M is an MXM module, which makes it suitable for laptop upgrades and portable systems. The P106-090 is a dual-slot desktop card with no display outputs, which means it is not suited for standard graphics output. The GTX 780M's display support is portable device dependent, giving it more flexibility in mobile environments.

For gaming, the picture is mixed. The P106-090's higher clocks and newer architecture likely help in most scenarios, but the GTX 780M's larger memory and wider bus could benefit high-resolution textures. The database does not record gaming-specific tests for these cards, so the compute results are the primary measurable evidence.

For compute workloads, the P106-090 wins decisively. For mobile integration and larger memory capacity, the GTX 780M has a role. The P106-090 is the stronger GPU overall, as its benchmark wins and higher percentile ranking demonstrate.