AMD Radeon Pro Vega 16 vs NVIDIA GeForce RTX 3070 Mobile Comparison

Head-to-Head Benchmarks

The recorded data contains two direct comparisons between the AMD Radeon Pro Vega 16 and the NVIDIA GeForce RTX 3070 Mobile. In both cases, the NVIDIA part dominates decisively.

Starting with Geekbench OpenCL, the AMD Radeon Pro Vega 16 scores 18,268 points. The NVIDIA GeForce RTX 3070 Mobile scores 92,939 points. That is a delta of -80.3% for the AMD part, meaning the RTX 3070 Mobile delivers roughly five times the compute throughput in this workload. The margin is enormous, and it reflects the fundamental gap in raw execution resources between these two products.

The Geekbench Vulkan result shows a similar pattern, though slightly less lopsided. The AMD Radeon Pro Vega 16 scores 21,832 points, while the NVIDIA GeForce RTX 3070 Mobile scores 86,768 points. The delta here is -74.8%. While the AMD part narrows the relative gap compared to OpenCL, it still trails by a factor of roughly four in Vulkan performance.

The overall head-to-head record is 0 wins for the AMD Radeon Pro Vega 16 and 2 wins for the NVIDIA GeForce RTX 3070 Mobile. There is no benchmark category in the database where the AMD part comes out ahead.

The average benchmark score for the AMD Radeon Pro Vega 16 is 23,250, which places it at the 68th percentile among all GPUs in the database. The NVIDIA GeForce RTX 3070 Mobile has an average score of 20,534, putting it at the 65th percentile. This is an interesting nuance: despite the massive head-to-head losses, the AMD part has a higher average score and percentile ranking. The explanation lies in the benchmark mix. The AMD part's average draws from three Geekbench results (Metal, OpenCL, Vulkan), while the NVIDIA part's average includes a broader set of tests, many of which are Passmark sub-tests with lower absolute scores that drag its average down. The head-to-head comparisons, however, are direct apples-to-apples tests, and they are unambiguous in favor of NVIDIA.

Looking at nearest rivals in the database, the AMD Radeon Pro Vega 16 sits within a tight cluster. The NVIDIA P106-100 has an average score of 23,249, essentially identical with a delta of 0%. The AMD Radeon RX 6600M scores 23,273, a delta of -0.1% relative to the Vega 16. The AMD Radeon R9 M290X scores 23,276, also -0.1%. The AMD Radeon AI PRO R9700 scores 23,315, a -0.3% delta. In other words, the Vega 16 is statistically indistinguishable from its nearest neighbors in aggregate performance, despite the wide variation in their underlying architectures.

The NVIDIA GeForce RTX 3070 Mobile's nearest rivals tell a different story. The Intel Arc B570 scores 20,556, a delta of -0.1% relative to the RTX 3070 Mobile. The Intel Arc A750 scores 20,582, a -0.2% delta. The NVIDIA Quadro M4000M scores 20,480, a +0.3% delta (meaning the Quadro is slightly ahead). The AMD Radeon R9 M390X scores 20,662, a -0.6% delta. The RTX 3070 Mobile is thus bracketed by a mix of Intel, NVIDIA, and AMD parts, all within a single percentage point.

Architecture Differences

The two GPUs come from different architectural generations and are built for different purposes.

The AMD Radeon Pro Vega 16 uses the Vega 12 chip based on the GCN 5.0 architecture, fabricated on a 14 nm process at GlobalFoundries. It belongs to the Radeon Pro Mac (Vega Series) generation. The NVIDIA GeForce RTX 3070 Mobile uses the GA104 chip based on the Ampere architecture, fabricated on an 8 nm process at Samsung. It belongs to the GeForce 30 Mobile generation. The process node difference alone, 14 nm versus 8 nm, explains a substantial portion of the efficiency and density gap.

The NVIDIA chip carries 17,400 million transistors on a 392 mm² die, with a transistor density of 44.4 million per square millimeter. The AMD chip's transistor count and die size are not recorded in the database, but the architectural generational gap is clear.

Compute resources differ massively. The AMD Radeon Pro Vega 16 has 1,024 shading units, 64 texture mapping units, and 32 ROPs. The NVIDIA GeForce RTX 3070 Mobile has 5,120 shading units, 160 TMUs, and 80 ROPs. That is a 5x advantage in shading units, a 2.5x advantage in TMUs, and a 2.5x advantage in ROPs.

The NVIDIA part also includes hardware features that the AMD part lacks entirely. The RTX 3070 Mobile has 40 ray tracing cores and 160 tensor cores, enabling DirectX 12 Ultimate support. The AMD Radeon Pro Vega 16 has no ray tracing cores and no tensor cores. Its DirectX support is capped at 12 (12_1), while the NVIDIA part supports DirectX 12 Ultimate (12_2). Both support OpenGL 4.6, but the NVIDIA part supports Vulkan 1.4 versus Vulkan 1.3 on the AMD side.

Clock speeds also differ. The AMD Radeon Pro Vega 16 runs at a base clock of 815 MHz and a boost clock of 1190 MHz. The NVIDIA GeForce RTX 3070 Mobile runs at a base clock of 1110 MHz and a boost clock of 1560 MHz. The NVIDIA part is higher clocked and has far more execution units, so the throughput gap compounds.

Memory architecture is another major divergence. The AMD Radeon Pro Vega 16 uses 4 GB of HBM2 on a 1024-bit bus, running at 1200 MHz with 2.4 Gbps effective data rate, yielding 307.2 GB/s of bandwidth. The NVIDIA GeForce RTX 3070 Mobile uses 8 GB of GDDR6 on a 256-bit bus, running at 1750 MHz with 14 Gbps effective data rate, yielding 448.0 GB/s of bandwidth. The AMD part uses a very wide bus with slower HBM2, while the NVIDIA part uses a narrower bus with much faster GDDR6. The NVIDIA part ends up with more capacity and more bandwidth.

The measured rates reflect the hardware differences. The AMD part achieves a pixel rate of 38.08 GPixel/s and a texture rate of 76.16 GTexel/s, with FP32 throughput of 2.437 TFLOPS and FP16 throughput of 4.874 TFLOPS (2:1 ratio). The NVIDIA part achieves a pixel rate of 124.8 GPixel/s and a texture rate of 249.6 GTexel/s, with FP32 throughput of 15.97 TFLOPS and FP16 throughput of 15.97 TFLOPS (1:1 ratio). The NVIDIA part is roughly 3.3x faster in pixel rate, 3.3x faster in texture rate, and 6.5x faster in FP32 throughput.

Power consumption also differs. The AMD Radeon Pro Vega 16 has a TDP of 75 W and is designated as an integrated graphics package (IGP) with no dedicated power connectors. The NVIDIA GeForce RTX 3070 Mobile has a TDP of 115 W and also lists no power connectors, consistent with a mobile part.

The bus interface differs as well: PCIe 3.0 x16 on the AMD side versus PCIe 4.0 x16 on the NVIDIA side. Both parts list display outputs as "Portable Device Dependent," indicating they are intended for laptop or all-in-one implementations.

The Verdict

The data supports a clear split decision based on workload and context.

For anyone running OpenCL or Vulkan compute workloads, the NVIDIA GeForce RTX 3070 Mobile is the overwhelming choice. It is 80.3% ahead in OpenCL and 74.8% ahead in Vulkan. These are not marginal differences; they are generational gaps. The RTX 3070 Mobile also offers DirectX 12 Ultimate, hardware ray tracing (40 RT cores), and 160 tensor cores, features that are simply absent from the AMD part.

The AMD Radeon Pro Vega 16's case rests on its aggregate profile. Its average benchmark score of 23,250 is higher than the RTX 3070 Mobile's 20,534, and its 68th percentile ranking beats the NVIDIA part's 65th percentile. It also has a lower TDP of 75 W versus 115 W, which can matter in thermally constrained mobile chassis. For a system that prioritizes power efficiency and where the primary workloads are not the compute-heavy tests where the RTX 3070 Mobile excels, the Vega 16 remains a coherent option.

However, the direct head-to-head benchmarks are the most reliable indicator of relative performance in identical workloads, and they are lopsided. The RTX 3070 Mobile wins both recorded comparisons. The AMD part wins zero.

The nearest rival data reinforces that both parts sit in competitive clusters. The Vega 16 is flanked by the NVIDIA P106-100, AMD RX 6600M, AMD R9 M290X, and AMD AI PRO R9700, all within 0.3% of its average score. The RTX 3070 Mobile is flanked by Intel Arc B570, Intel Arc A750, NVIDIA Quadro M4000M, and AMD R9 M390X, all within 0.6%. Neither part is an outlier in its respective performance tier.

The production status for both parts is end-of-life, so neither is a forward-looking purchase. The AMD part was released on November 13, 2018; the NVIDIA part was released on January 11, 2021.

Specification Differences

The following fields differ between the two parts in the database:

• Architecture: GCN 5.0 (AMD) versus Ampere (NVIDIA)
• Process node: 14 nm (AMD) versus 8 nm (NVIDIA)
• Foundry: GlobalFoundries (AMD) versus Samsung (NVIDIA)
• Transistors: not recorded (AMD) versus 17,400 million (NVIDIA)
• Die size: not recorded (AMD) versus 392 mm² (NVIDIA)
• Transistor density: not recorded (AMD) versus 44.4M / mm² (NVIDIA)
• Base clock: 815 MHz (AMD) versus 1110 MHz (NVIDIA)
• Boost clock: 1190 MHz (AMD) versus 1560 MHz (NVIDIA)
• Memory clock: 1200 MHz, 2.4 Gbps effective (AMD) versus 1750 MHz, 14 Gbps effective (NVIDIA)
• Memory size: 4 GB (AMD) versus 8 GB (NVIDIA)
• Memory type: HBM2 (AMD) versus GDDR6 (NVIDIA)
• Memory bus width: 1024 bit (AMD) versus 256 bit (NVIDIA)
• Memory bandwidth: 307.2 GB/s (AMD) versus 448.0 GB/s (NVIDIA)
• Shading units: 1024 (AMD) versus 5120 (NVIDIA)
• TMUs: 64 (AMD) versus 160 (NVIDIA)
• ROPs: 32 (AMD) versus 80 (NVIDIA)
• RT cores: none (AMD) versus 40 (NVIDIA)
• Tensor cores: none (AMD) versus 160 (NVIDIA)
• Pixel rate: 38.08 GPixel/s (AMD) versus 124.8 GPixel/s (NVIDIA)
• Texture rate: 76.16 GTexel/s (AMD) versus 249.6 GTexel/s (NVIDIA)
• FP32: 2.437 TFLOPS (AMD) versus 15.97 TFLOPS (NVIDIA)
• FP16: 4.874 TFLOPS (2:1) (AMD) versus 15.97 TFLOPS (1:1) (NVIDIA)
• TDP: 75 W (AMD) versus 115 W (NVIDIA)
• Slot width: IGP (AMD) versus not recorded (NVIDIA)
• Bus interface: PCIe 3.0 x16 (AMD) versus PCIe 4.0 x16 (NVIDIA)
• DirectX: 12 (12_1) (AMD) versus 12 Ultimate (12_2) (NVIDIA)
• Vulkan: 1.3 (AMD) versus 1.4 (NVIDIA)
• Release date: November 13, 2018 (AMD) versus January 11, 2021 (NVIDIA)
• Predecessor: not recorded (AMD) versus GeForce 20 Mobile (NVIDIA)

Fields that are the same or not recorded for either part include: chip codename, game clock, display outputs (both "Portable Device Dependent"), dimensions, power connectors (both none), suggested PSU, and launch MSRP.

FAQ

Q: Which GPU wins the direct head-to-head benchmarks?

A: The NVIDIA GeForce RTX 3070 Mobile wins both recorded head-to-head tests. It scores 92,939 versus 18,268 in Geekbench OpenCL, a delta of -80.3% for the AMD part, and 86,768 versus 21,832 in Geekbench Vulkan, a delta of -74.8%.

Q: How do the two GPUs compare in average benchmark score?

A: The AMD Radeon Pro Vega 16 has an average benchmark score of 23,250, while the NVIDIA GeForce RTX 3070 Mobile has an average of 20,534. The AMD part ranks at the 68th percentile among all GPUs, and the NVIDIA part ranks at the 65th percentile.

Q: What are the nearest rivals of the AMD Radeon Pro Vega 16?

A: The nearest rivals are the NVIDIA P106-100 (23,249, 0% delta), the AMD Radeon RX 6600M (23,273, -0.1% delta), the AMD Radeon R9 M290X (23,276, -0.1% delta), and the AMD Radeon AI PRO R9700 (23,315, -0.3% delta).

Q: Does the AMD Radeon Pro Vega 16 support ray tracing?

A: No. The AMD Radeon Pro Vega 16 has no ray tracing cores and no tensor cores. The NVIDIA GeForce RTX 3070 Mobile has 40 ray tracing cores and 160 tensor cores.

Q: What are the memory configurations of the two GPUs?

A: The AMD Radeon Pro Vega 16 uses 4 GB of HBM2 on a 1024-bit bus with 307.2 GB/s of bandwidth. The NVIDIA GeForce RTX 3070 Mobile uses 8 GB of GDDR6 on a 256-bit bus with 448.0 GB/s of bandwidth.

Q: What is the power draw difference between the two parts?

A: The AMD Radeon Pro Vega 16 has a TDP of 75 W and is listed as an integrated graphics package. The NVIDIA GeForce RTX 3070 Mobile has a TDP of 115 W. Both parts list no power connectors.