AMD Radeon Vega 11 vs NVIDIA P106-090 Comparison

AMD Radeon Vega 11 and NVIDIA P106-090 represent two very different approaches to graphics hardware: one is an integrated processor within a mobile/desktop APU package, while the other is a dedicated discrete card stripped of display outputs for mining workloads. The data shows a clear performance hierarchy, with the NVIDIA P106-090 winning both shared head-to-head benchmark tests by substantial margins. The AMD Vega 11, part of the Picasso generation on a 12 nm process, offers 704 shading units and a boost clock of 1400 MHz, while the NVIDIA P106-090, built on Pascal at 16 nm, provides 768 shading units and a boost clock of 1531 MHz. The P106-090 also holds a significant memory advantage with 3 GB of dedicated GDDR5 on a 192-bit bus, versus the Vega 11's system-shared memory. However, the Vega 11 counters with a much lower 15 W TDP, making it suitable for systems without discrete power connectors, whereas the P106-090 requires a 6-pin connector and a 250 W suggested PSU. This comparison ultimately hinges on whether raw compute performance or system integration is the priority.

The Verdict

From the benchmark data, the NVIDIA P106-090 is the superior performer in raw compute. In Geekbench OpenCL, it scores 21304 against the Vega 11's 13392, a 37.1% advantage. In Geekbench Vulkan, the P106-090 scores 18596 versus 12273, a 34% lead. The head-to-head record reflects this dominance with 2 wins for NVIDIA and 0 for AMD. The P106-090's average benchmark score of 13470, however, is slightly lower than the Vega 11's 14352, which is a notable inversion driven by the specific test suites each card was subjected to. The Vega 11's percentile rank of 56 places it above the P106-090's 54, indicating that across all GPUs in the database, the integrated part performs relatively better in the broader benchmark set. Therefore, the verdict is split: for users prioritizing peak compute in OpenCL or Vulkan workloads, the P106-090 is the clear choice, especially given its 3 GB of dedicated memory. For users who need a graphics solution that requires no additional power connectors, fits in an IGP slot, and offers a higher overall benchmark percentile, the Vega 11 is the safer pick. The P106-090's lack of display outputs also makes it unsuitable as a primary graphics card for typical desktop use, limiting its appeal to compute or mining scenarios.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon Vega 11 has a higher average benchmark score of 14352, compared to the NVIDIA P106-090's 13470. This is despite the P106-090 winning both head-to-head tests.

Q: What is the memory configuration difference?

A: The NVIDIA P106-090 has 3 GB of dedicated GDDR5 memory with a 192-bit bus and 192.2 GB/s bandwidth. The AMD Vega 11 uses system-shared memory, meaning its size, type, and bus width are all listed as "System Shared" and bandwidth is "System Dependent."

Q: Does the P106-090 support display outputs?

A: No, the NVIDIA P106-090 has no display outputs. The AMD Vega 11's display outputs are listed as "Motherboard Dependent," meaning they are provided by the host system's motherboard.

Q: What are the power requirements for each card?

A: The AMD Vega 11 has a 15 W TDP and requires no power connectors, fitting in an IGP slot. The NVIDIA P106-090 has a 75 W TDP, requires a single 6-pin power connector, a 250 W suggested PSU, and occupies a dual-slot width.

Q: Which architecture do these GPUs use?

A: The AMD Vega 11 uses GCN 5.0 architecture on a 12 nm process, while the NVIDIA P106-090 uses Pascal architecture on a 16 nm process.

Q: How do the FP16 compute capabilities differ?

A: The AMD Vega 11 offers 3.942 TFLOPS FP16 performance at a 2:1 ratio, while the NVIDIA P106-090 offers only 36.74 GFLOPS at a 1:64 ratio, making the Vega 11 vastly superior in this specific metric.

Architecture Differences

The two GPUs are built on fundamentally different architectures and process nodes. The AMD Radeon Vega 11 uses the GCN 5.0 architecture, manufactured by GlobalFoundries on a 12 nm process. It is part of the Picasso generation, which is a Vega IGP design. The chip itself is named Picasso and contains 4,940 million transistors on a 210 mm² die, resulting in a transistor density of 23.5M per mm². In contrast, the NVIDIA P106-090 uses the Pascal architecture, manufactured by TSMC on a 16 nm process. It belongs to the "Mining GPUs" generation and uses the GP106 chip, with 4,400 million transistors on a 200 mm² die, giving a density of 22.0M per mm². The Vega 11's newer 12 nm process offers a slight density advantage, but the P106-090's dedicated memory interface is a key architectural differentiator. The Vega 11 integrates its memory controller with the system, while the P106-090 has a dedicated 192-bit bus for its GDDR5 memory. The FP16 compute capability is starkly different: the Vega 11 supports 2:1 FP16 rates, achieving 3.942 TFLOPS, while the P106-090's FP16 is severely limited at 1:64 ratio, yielding only 36.74 GFLOPS. Both support DirectX 12 (12_1) and OpenGL 4.6, but the Vega 11 supports Vulkan 1.3 while the P106-090 supports Vulkan 1.4, indicating a newer API implementation on the NVIDIA side.

Specification Differences

The specification sheets reveal distinct design goals. The AMD Vega 11 has a base clock of 300 MHz and a boost clock of 1400 MHz, while the NVIDIA P106-090 starts at a much higher base of 1354 MHz and boosts to 1531 MHz. The P106-090's memory runs at 2002 MHz (8 Gbps effective), while the Vega 11's memory clock is listed as "System Shared." In terms of compute units, the Vega 11 has 704 shading units, 44 TMUs, and 8 ROPs. The P106-090 has 768 shading units, 48 TMUs, and 48 ROPs — the latter being a six-fold increase in ROP count. This translates to a pixel rate of 11.20 GPixel/s for the Vega 11 versus 73.49 GPixel/s for the P106-090, and a texture rate of 61.60 GTexel/s versus 73.49 GTexel/s. The FP32 performance is 1.971 TFLOPS for AMD and 2.352 TFLOPS for NVIDIA. Power consumption is a major split: the Vega 11 is rated at 15 W TDP with no power connectors, while the P106-090 is rated at 75 W TDP with a 1x 6-pin connector and a suggested PSU of 250 W. The Vega 11 is an IGP with no slot width and a bus interface of "IGP," whereas the P106-090 is a dual-slot card with a PCIe 1.0 x1 interface and a physical length of 250 mm (9.8 inches). The P106-090 has no display outputs; the Vega 11 relies on motherboard-dependent outputs. Production status for both is end-of-life, with release dates of 2019-09-29 for the Vega 11 and 2017-07-30 for the P106-090.

Head-to-Head Benchmarks

The shared benchmark suite consists of two tests: Geekbench OpenCL and Geekbench Vulkan. In Geekbench OpenCL, the NVIDIA P106-090 scores 21304 against the AMD Vega 11's 13392, yielding a delta of -37.1% from the perspective of the AMD card. This is a substantial lead for NVIDIA, reflecting its higher shading unit count and dedicated memory bandwidth. In Geekbench Vulkan, the P106-090 scores 18596 versus the Vega 11's 12273, a delta of -34%. While the margin narrows slightly in Vulkan, the NVIDIA card still maintains a commanding lead. The head-to-head record is 2-0 in favor of NVIDIA. However, the average benchmark scores tell a different story: the Vega 11's average of 14352 is higher than the P106-090's 13470. This discrepancy arises because the Vega 11 was also benchmarked with Geekbench Metal, scoring 17392, which is its highest result and boosts its average. The P106-090 also has a 3DMark Steel Nomad DX12 score of 509, which is relatively low and drags its average down. The Vega 11's nearest rivals include the NVIDIA GeForce GTX TITAN with an average score of 14373 (delta -0.1%) and the AMD Radeon RX Vega 11 at 14385 (delta -0.2%), showing it sits in a tight performance cluster. The P106-090's nearest rivals include the NVIDIA GeForce GTX 570 at 13515 (delta -0.3%) and the AMD Radeon Pro 555 at 13407 (delta 0.5%), indicating it is also closely matched with older discrete parts.

Where Each One Wins

The NVIDIA P106-090 wins decisively in both head-to-head compute benchmarks, making it the better choice for OpenCL and Vulkan compute tasks. Its 3 GB of GDDR5 memory with 192.2 GB/s bandwidth provides a significant advantage for workloads that require local memory access without system RAM contention. The higher ROP count (48 vs 8) and pixel rate (73.49 GPixel/s vs 11.20 GPixel/s) suggest it is also far better suited for rasterization-heavy tasks, although no direct gaming benchmarks are present in the data. The P106-090 is the pick for users who need a dedicated compute card and have the power budget (75 W TDP, 6-pin connector, 250 W PSU) and physical space (250 mm length, dual-slot) to accommodate it.

The AMD Radeon Vega 11 wins in the areas of power efficiency and system integration. Its 15 W TDP requires no external power connectors, making it usable in compact or low-power systems where a discrete card cannot be installed. Its IGP form factor means it uses system memory, eliminating the need for a dedicated VRAM allocation, and its display outputs are motherboard-dependent, which is standard for integrated graphics. The Vega 11 also has a higher average benchmark score (14352 vs 13470) and a higher percentile rank (56 vs 54), indicating it performs better relative to the full GPU landscape. Its FP16 performance is vastly superior at 3.942 TFLOPS compared to the P106-090's 36.74 GFLOPS, making it the better option for workloads that leverage half-precision arithmetic. Additionally, the Vega 11's Vulkan 1.3 support is older than the P106-090's Vulkan 1.4, but its overall benchmark diversity, including a strong Geekbench Metal score of 17392, shows versatility across different API environments. For users building a system without a discrete GPU slot or power budget, the Vega 11 is the only viable option in this comparison.