AMD Radeon Pro Vega 16 vs NVIDIA GeForce GTX 780 Comparison

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon Pro Vega 16 records an average benchmark score of 23250, while the NVIDIA GeForce GTX 780 scores 19164. This places the AMD part at the 68th percentile of all GPUs, compared to the 64th percentile for the NVIDIA card.

Q: How does the Radeon Pro Vega 16 compare to its nearest rivals?

A: The database shows the AMD card sits within 0.3% of the AMD Radeon AI PRO R9700 (23315), AMD Radeon R9 M290X (23276), and AMD Radeon RX 6600M (23273). It is essentially tied with the NVIDIA P106-100, which scores 23249.

Q: What are the biggest performance gaps in the head-to-head tests?

A: The AMD Radeon Pro Vega 16 leads by 193.2% in Geekbench Metal, scoring 29650 versus 10114. Conversely, the NVIDIA GeForce GTX 780 leads by 20.1% in Geekbench OpenCL (22863 versus 18268) and by 10.9% in Geekbench Vulkan (24514 versus 21832).

Q: Which card has a higher pixel rate?

A: The NVIDIA GeForce GTX 780 achieves 43.30 GPixel/s, while the AMD Radeon Pro Vega 16 reaches 38.08 GPixel/s. The NVIDIA card also has a higher texture rate at 173.2 GTexel/s versus 76.16 GTexel/s.

Q: What is the memory configuration difference?

A: The AMD Radeon Pro Vega 16 uses 4 GB of HBM2 memory on a 1024-bit bus, delivering 307.2 GB/s of bandwidth. The NVIDIA GeForce GTX 780 uses 3 GB of GDDR5 on a 384-bit bus, providing 288.4 GB/s of bandwidth.

Q: Which GPU supports newer graphics APIs?

A: The AMD Radeon Pro Vega 16 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.3. The NVIDIA GeForce GTX 780 supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175.

Architecture Differences

The AMD Radeon Pro Vega 16 is built on the Vega 12 chip using GCN 5.0 architecture, fabricated on a 14 nm process at GlobalFoundries. The NVIDIA GeForce GTX 780 uses the GK110 chip with Kepler architecture, produced on a 28 nm process at TSMC. The process node difference is significant: 14 nm versus 28 nm allows the AMD chip to pack more functionality into a smaller area, though the database does not list transistor counts for the AMD part. The NVIDIA chip contains 7,080 million transistors on a 561 mm² die, with a transistor density of 12.6M per mm².

The AMD card integrates 1024 shading units, 64 texture mapping units, and 32 raster output units. The NVIDIA card has 2304 shading units, 192 texture mapping units, and 48 raster output units. Despite having fewer shading units, the AMD card operates at higher clock speeds in some respects: its boost clock reaches 1190 MHz versus 902 MHz for the NVIDIA card. The base clocks are 815 MHz for AMD and 863 MHz for NVIDIA.

Memory architecture differs fundamentally. The AMD Radeon Pro Vega 16 uses 4 GB of HBM2 with a 1024-bit memory bus and 2.4 Gbps effective speed, producing 307.2 GB/s of bandwidth. The NVIDIA GeForce GTX 780 uses 3 GB of GDDR5 on a 384-bit bus with 6 Gbps effective speed, yielding 288.4 GB/s. The AMD memory bus is substantially wider, but the NVIDIA memory runs at higher effective clock speeds.

The AMD card is an integrated GPU (IGP) with a 75 W TDP, while the NVIDIA card is a dual-slot discrete solution with a 250 W TDP, requiring a 6-pin plus 8-pin power connector and a 600 W suggested power supply. The NVIDIA card measures 267 mm in length, 111 mm in height, and 38 mm in width. The AMD card's display outputs are portable device dependent, whereas the NVIDIA card provides 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2 outputs.

The AMD card supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.3. The NVIDIA card supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The AMD card also supports FP16 computation at 4.874 TFLOPS with a 2:1 ratio, while no FP16 figure is recorded for the NVIDIA card. The AMD card's FP32 performance is 2.437 TFLOPS, compared to 4.156 TFLOPS for the NVIDIA card.

Head-to-Head Benchmarks

The head-to-head results reveal a split personality. In Geekbench Metal, the AMD Radeon Pro Vega 16 achieves 29650, which is 193.2% higher than the NVIDIA GeForce GTX 780's 10114. This is the largest margin in any test. The AMD card's Metal performance is exceptional, reflecting its design for Apple's ecosystem, as it belongs to the Radeon Pro Mac generation.

The NVIDIA GeForce GTX 780 wins the other two tests. In Geekbench OpenCL, the NVIDIA card scores 22863, which is 20.1% higher than the AMD card's 18268. In Geekbench Vulkan, the NVIDIA card scores 24514, beating the AMD card's 21832 by 10.9%. The NVIDIA card wins two out of three head-to-head tests, giving it a 2-1 victory in the benchmark suite.

The average benchmark score tells a different story. The AMD Radeon Pro Vega 16 averages 23250 across all recorded tests, while the NVIDIA GeForce GTX 780 averages 19164. This 21.3% advantage in average score stems from the massive Metal result, which outweighs the NVIDIA card's narrower wins in OpenCL and Vulkan. The AMD card's average score places it at the 68th percentile of all GPUs, versus the 64th percentile for the NVIDIA card.

The delta percentages in the nearest rivals lists provide context. The AMD card's closest rival, the NVIDIA P106-100, scores 23249 with a 0% delta. The AMD Radeon RX 6600M and Radeon R9 M290X both sit 0.1% lower, and the AMD Radeon AI PRO R9700 is 0.3% lower. The NVIDIA GTX 780's nearest rival, the NVIDIA TITAN Xp, scores 19177 with a -0.1% delta, meaning the GTX 780 is essentially tied. The Tesla K20m is 0.4% lower, the GeForce RTX 4050 Mobile is 0.6% lower, and the AMD Radeon RX 6600 is 0.7% lower.

The raw compute throughput figures support the NVIDIA card's OpenCL and Vulkan wins. The NVIDIA card has 4.156 TFLOPS of FP32 performance, while the AMD card has 2.437 TFLOPS. The NVIDIA card also has higher pixel and texture rates: 43.30 GPixel/s versus 38.08 GPixel/s, and 173.2 GTexel/s versus 76.16 GTexel/s respectively. These figures align with the NVIDIA card's advantage in general-purpose compute workloads.

The Verdict

The data presents two distinct profiles. The AMD Radeon Pro Vega 16 excels in Metal-based workloads, where it is nearly three times faster than the NVIDIA GeForce GTX 780. Its 193.2% lead in Geekbench Metal is the single largest performance gap in the head-to-head comparison. This makes it the clear choice for applications that leverage Apple's Metal API, which aligns with its positioning in the Radeon Pro Mac generation.

The NVIDIA GeForce GTX 780 dominates in OpenCL and Vulkan. Its 20.1% lead in OpenCL and 10.9% lead in Vulkan indicate stronger performance in cross-platform compute and graphics workloads. The NVIDIA card also has a significant advantage in raw compute throughput: 4.156 TFLOPS of FP32 versus 2.437 TFLOPS, along with higher pixel and texture fill rates. Users prioritizing OpenCL or Vulkan performance should favor the NVIDIA card.

The average benchmark score favors the AMD card, 23250 versus 19164, which reflects the outsized impact of the Metal result. The AMD card also holds a higher percentile ranking, 68th versus 64th. However, this aggregate measure hides the fact that the NVIDIA card wins two of the three individual tests.

The AMD card consumes 75 W versus 250 W for the NVIDIA card, and it is an integrated GPU rather than a dual-slot discrete card. The NVIDIA card requires a 600 W power supply and measures 267 mm in length. The AMD card's lower power draw and integrated form factor make it suitable for portable devices, while the NVIDIA card is a traditional desktop component.

The release dates differ by over five years: the AMD card launched on November 13, 2018, while the NVIDIA card launched on May 22, 2013. The NVIDIA card has a documented launch MSRP of 649 USD, while no launch MSRP is recorded for the AMD card. Both are end-of-life products.

For Metal-centric workflows, the AMD Radeon Pro Vega 16 is the superior choice. For OpenCL, Vulkan, or raw compute throughput, the NVIDIA GeForce GTX 780 is the better option. The AMD card's average score advantage suggests it is the better all-rounder in this specific benchmark suite, but the NVIDIA card's wins in two of three tests indicate it remains competitive in general-purpose scenarios. The choice ultimately depends on the target API and workload profile.

Specification Differences

| Specification | AMD Radeon Pro Vega 16 | NVIDIA GeForce GTX 780 |

|---|---|---|

| Chip | Vega 12 | GK110 |

| Architecture | GCN 5.0 | Kepler |

| Process Node | 14 nm | 28 nm |

| Foundry | GlobalFoundries | TSMC |

| Transistors | Not listed | 7,080 million |

| Die Size | Not listed | 561 mm² |

| Transistor Density | Not listed | 12.6M / mm² |

| Base Clock | 815 MHz | 863 MHz |

| Boost Clock | 1190 MHz | 902 MHz |

| Memory Clock | 1200 MHz, 2.4 Gbps effective | 1502 MHz, 6 Gbps effective |

| Memory Size | 4 GB | 3 GB |

| Memory Type | HBM2 | GDDR5 |

| Memory Bus Width | 1024 bit | 384 bit |

| Memory Bandwidth | 307.2 GB/s | 288.4 GB/s |

| Shading Units | 1024 | 2304 |

| TMUs | 64 | 192 |

| ROPs | 32 | 48 |

| Pixel Rate | 38.08 GPixel/s | 43.30 GPixel/s |

| Texture Rate | 76.16 GTexel/s | 173.2 GTexel/s |

| FP32 Performance | 2.437 TFLOPS | 4.156 TFLOPS |

| FP16 Performance | 4.874 TFLOPS (2:1) | Not listed |

| TDP | 75 W | 250 W |

| Slot Width | IGP | Dual-slot |

| Power Connectors | Not listed | 1x 6-pin + 1x 8-pin |

| Suggested PSU | Not listed | 600 W |

| Display Outputs | Portable Device Dependent | 2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 |

| DirectX Support | 12 (12_1) | 12 (11_0) |

| OpenGL Support | 4.6 | 4.6 |

| Vulkan Support | 1.3 | 1.2.175 |

| Dimensions | Not listed | 267 mm x 111 mm x 38 mm |

| Release Date | November 13, 2018 | May 22, 2013 |

| Launch MSRP | Not listed | 649 USD |