AMD Radeon 840M vs AMD Ryzen Z2 A GPU Comparison

The Verdict

The data presents two integrated graphics solutions from AMD aimed at different implementation targets. The Radeon 840M, built on Krackan Point with RDNA 3.5 architecture, is a 4 nm part with a 400 MHz base clock and 2900 MHz boost clock. The Ryzen Z2 A GPU, built on Van Gogh with RDNA 2.0 architecture, is a 7 nm part with a 1000 MHz base clock and 1600 MHz boost clock. Both carry a 15 W TDP, placing them in the same power envelope but with different performance characteristics.

The Radeon 840M targets portable devices where the system memory is shared, and its 256 shading units operate at a much higher boost frequency. The Ryzen Z2 A GPU uses dedicated LPDDR5 memory with a 128-bit bus and 102.4 GB/s bandwidth, which eliminates the system memory dependency. The Z2 A GPU also has twice the shading units (512 versus 256), twice the texture mapping units (32 versus 16), and twice the render output units (16 versus 8). It also includes 8 RT cores versus 4 on the 840M.

For users prioritizing raw throughput in a shared-memory design, the Radeon 840M delivers a higher boost clock and newer architecture. For users needing dedicated memory bandwidth and more compute units, the Ryzen Z2 A GPU offers a more substantial hardware configuration. The recorded data shows both GPUs occupy the 50th percentile among all GPUs in the database, with identical average benchmark scores of zero, meaning the database contains no direct performance measurements for either part.

Architecture Differences

The Radeon 840M uses the Krackan Point chip built on TSMC's 4 nm process, while the Ryzen Z2 A GPU uses the Van Gogh chip built on TSMC's 7 nm process. This process difference is significant: the 4 nm node allows the 840M to reach a 2900 MHz boost clock, whereas the 7 nm Van Gogh chip peaks at 1600 MHz. The newer RDNA 3.5 architecture in the 840M contrasts with the RDNA 2.0 architecture in the Z2 A GPU.

The Radeon 840M belongs to the Navi III IGP generation under the Strix Point Mobile family, while the Ryzen Z2 A GPU is categorized as a Console GPU under AMD's generation naming. The 840M has a transistor count listed as unknown and a die size listed as unknown. The Z2 A GPU has a documented 2,400 million transistors on a 163 mm² die, yielding a transistor density of 14.7 million per square millimeter.

Memory architecture separates these two parts fundamentally. The 840M uses system-shared memory with a system-dependent bandwidth, meaning its performance scales with the host platform's memory configuration. The Z2 A GPU has 16 GB of LPDDR5 memory on a 128-bit bus with 102.4 GB/s fixed bandwidth. The 840M's memory clock is listed as "System Shared" with no dedicated memory clock, while the Z2 A GPU runs its memory at 800 MHz with 6.4 Gbps effective data rate.

Both GPUs support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The 840M uses a PCIe 4.0 x8 bus interface, while the Z2 A GPU's bus interface is not specified in the database. The 840M is an IGP with no power connectors and a slot width of "IGP," while the Z2 A GPU has null values for slot width and power connectors, indicating a different physical integration approach. Display outputs differ: the 840M is portable-device dependent, while the Z2 A GPU provides a single USB Type-C output.

Head-to-Head Benchmarks

The database records no direct head-to-head benchmark results between these two GPUs, and neither part has individual benchmark scores or nearest rivals listed. The average benchmark score for both is zero, and the wins count for each is zero. This absence of measured data means the comparison must rely entirely on the specification fields provided.

The arithmetic rate calculations from the recorded clocks and unit counts show specific deltas. The Radeon 840M achieves a pixel rate of 23.20 GPixel/s, while the Ryzen Z2 A GPU achieves 25.60 GPixel/s, a difference of 2.40 GPixel/s in favor of the Z2 A GPU. The texture rate shows a smaller gap: 46.40 GTexel/s for the 840M versus 51.20 GTexel/s for the Z2 A GPU, a 4.80 GTexel/s advantage for the latter. These differences stem from the Z2 A GPU having double the ROPs and TMUs, which offsets its lower boost clock.

The FP32 throughput tells a different story. The 840M delivers 1,484.8 GFLOPS, while the Z2 A GPU delivers 1.638 TFLOPS, which converts to 1,638 GFLOPS. This is a 153.2 GFLOPS advantage for the Z2 A GPU, roughly 10% higher. The FP16 figures show a larger divergence: the 840M posts 1,484.8 GFLOPS with a 1:1 ratio, while the Z2 A GPU posts 3.277 TFLOPS with a 2:1 ratio. The Z2 A GPU's FP16 throughput is more than double that of the 840M, which reflects the RDNA 2 architecture's packed math capability versus the RDNA 3.5 design's conservative 1:1 approach.

Clock speed differences are substantial. The 840M boosts to 2900 MHz, which is 81.25% higher than the Z2 A GPU's 1600 MHz boost. The base clocks show the opposite relationship: the Z2 A GPU runs at 1000 MHz base, while the 840M idles at 400 MHz base. This indicates the 840M relies more heavily on boosting to reach its performance targets, while the Z2 A GPU maintains a higher floor.

Specification Differences

The following fields differ between the two parts, with no other specification differences recorded:

• Chip: Krackan Point versus Van Gogh
• Architecture: RDNA 3.5 versus RDNA 2.0
• Generation: Navi III IGP (Strix Point Mobile) versus Console GPU (AMD)
• Process Node: 4 nm versus 7 nm
• Transistors: unknown versus 2,400 million
• Die Size: unknown versus 163 mm²
• Transistor Density: null versus 14.7M / mm²
• Base Clock: 400 MHz versus 1000 MHz
• Boost Clock: 2900 MHz versus 1600 MHz
• Memory Clock: System Shared versus 800 MHz 6.4 Gbps effective
• Memory Size: System Shared versus 16 GB
• Memory Type: System Shared versus LPDDR5
• Memory Bus Width: System Shared versus 128 bit
• Memory Bandwidth: System Dependent versus 102.4 GB/s
• Shading Units: 256 versus 512
• TMUs: 16 versus 32
• ROPs: 8 versus 16
• RT Cores: 4 versus 8
• Pixel Rate: 23.20 GPixel/s versus 25.60 GPixel/s
• Texture Rate: 46.40 GTexel/s versus 51.20 GTexel/s
• FP32: 1,484.8 GFLOPS versus 1.638 TFLOPS
• FP16: 1,484.8 GFLOPS (1:1) versus 3.277 TFLOPS (2:1)
• Slot Width: IGP versus null
• Power Connectors: None versus null
• Bus Interface: PCIe 4.0 x8 versus null
• Display Outputs: Portable Device Dependent versus 1x USB Type-C
• Release Date: 2025-02-28 versus 2024-12-31
• Predecessor: Navi II IGP versus null

FAQ

Q: Which GPU has the higher boost clock?

A: The Radeon 840M has a 2900 MHz boost clock, while the Ryzen Z2 A GPU has a 1600 MHz boost clock.

Q: Which GPU has more shading units?

A: The Ryzen Z2 A GPU has 512 shading units, double the 256 shading units of the Radeon 840M.

Q: How does memory configuration differ between the two?

A: The Radeon 840M uses system-shared memory with bandwidth described as system dependent. The Ryzen Z2 A GPU has 16 GB of LPDDR5 memory on a 128-bit bus with 102.4 GB/s fixed bandwidth.

Q: What is the FP32 performance difference?

A: The Radeon 840M delivers 1,484.8 GFLOPS, while the Ryzen Z2 A GPU delivers 1.638 TFLOPS (1,638 GFLOPS), a 153.2 GFLOPS advantage for the Z2 A GPU.

Q: Do both GPUs support the same API levels?

A: Yes, both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Q: Which GPU has a higher pixel rate?

A: The Ryzen Z2 A GPU achieves 25.60 GPixel/s, higher than the Radeon 840M's 23.20 GPixel/s.

Where Each One Wins

The Radeon 840M wins on clock speed and process efficiency. Its 2900 MHz boost clock is 1300 MHz higher than the Z2 A GPU's boost, and its 4 nm node represents two generations of process advancement over the 7 nm Van Gogh chip. The 840M also carries the RDNA 3.5 architecture, which is newer than RDNA 2.0. For workloads that scale with clock frequency, such as lightly threaded rendering tasks or latency-sensitive operations, the 840M's higher boost provides an advantage. Its PCIe 4.0 x8 interface also gives it a defined host connection, whereas the Z2 A GPU's bus interface is unspecified.

The Ryzen Z2 A GPU wins on compute unit count, memory bandwidth, and fixed memory capacity. With 512 shading units, 32 TMUs, 16 ROPs, and 8 RT cores, it has double the resources of the 840M in every unit category. Its 16 GB of dedicated LPDDR5 memory with 102.4 GB/s bandwidth removes dependency on system memory performance, which is critical for sustained workloads where the host platform's memory speed could bottleneck the 840M. The Z2 A GPU also delivers higher pixel rate (25.60 GPixel/s versus 23.20 GPixel/s) and higher texture rate (51.20 GTexel/s versus 46.40 GTexel/s) despite its lower clock.

The FP32 comparison favors the Z2 A GPU at 1.638 TFLOPS versus 1,484.8 GFLOPS, and the FP16 comparison favors it even more strongly at 3.277 TFLOPS versus 1,484.8 GFLOPS. The 2:1 FP16 ratio on the Z2 A GPU indicates support for packed math operations that the 840M's 1:1 ratio does not provide. For applications using FP16 compute, the Z2 A GPU offers more than double the throughput.

The release dates show the 840M launched on 2025-02-28, while the Z2 A GPU launched on 2024-12-31. Both are marked as Active production status. The 840M has a predecessor listed as Navi II IGP, while the Z2 A GPU has no predecessor recorded. Neither part has a successor listed.

For systems where the host memory is fast and the workload is clock-sensitive, the 840M's RDNA 3.5 architecture and 2900 MHz boost make it the stronger choice. For systems requiring predictable memory bandwidth, double the compute units, and a fixed 16 GB memory pool, the Z2 A GPU provides a more substantial configuration. The data does not include benchmark scores or rival comparisons, so the analysis rests solely on the recorded specifications. Both GPUs sit at the 50th percentile in the database, indicating no measured performance differentiation at this time.