AMD Ryzen Z2 A GPU vs AMD Ryzen Z2 Extreme GPU Comparison

FAQ

Q: What are the two processors compared in this database entry?

A: The AMD Ryzen Z2 A GPU, based on the Van Gogh chip with RDNA 2.0 architecture, and the AMD Ryzen Z2 Extreme GPU, based on the Strix Point chip with RDNA 3.5 architecture.

Q: Which processor has a higher boost clock?

A: The AMD Ryzen Z2 Extreme GPU has a boost clock of 2700 MHz, while the AMD Ryzen Z2 A GPU boosts to 1600 MHz. The Extreme part also has a lower base clock at 800 MHz versus 1000 MHz for the Z2 A.

Q: How do the memory subsystems differ between the two?

A: Both use a 128-bit bus and 16 GB of memory, but the Z2 A uses LPDDR5 at 6.4 Gbps effective, delivering 102.4 GB/s. The Z2 Extreme uses LPDDR5X at 8 Gbps effective, delivering 128.0 GB/s.

Q: What is the transistor count and process node for each?

A: The Z2 A has 2,400 million transistors on a 7 nm process with a die size of 163 mm². The Z2 Extreme has 34,000 million transistors on a 4 nm process with a die size of 233 mm².

Q: Which processor has more shading units and ray tracing cores?

A: The AMD Ryzen Z2 Extreme GPU has 1024 shading units and 16 ray tracing cores. The AMD Ryzen Z2 A GPU has 512 shading units and 8 ray tracing cores.

Q: What is the average benchmark score for each in the database?

A: The Z2 Extreme has an average benchmark score of 516 in the 3DMark Steel Nomad DX12 test, while the Z2 A has no recorded benchmark scores in the database. The Z2 Extreme sits at the 1st percentile of all GPUs.

Where Each One Wins

The AMD Ryzen Z2 Extreme GPU is the clear performance leader in every measurable category. Its FP32 compute of 5.530 TFLOPS is more than triple the 1.638 TFLOPS of the Z2 A. The pixel rate of 129.6 GPixel/s versus 25.60 GPixel/s represents a massive advantage in fill-rate-bound scenarios. The texture rate of 172.8 GTexel/s compared to 51.20 GTexel/s gives the Extreme part a substantial edge in texture-heavy workloads.

The Z2 A does hold one advantage: power consumption. Its 15 W TDP is significantly lower than the 28 W TDP of the Z2 Extreme. This makes the Z2 A the more appropriate choice for designs where thermal and power budgets are extremely constrained. The Z2 A also has a higher base clock (1000 MHz versus 800 MHz), which means it maintains a higher floor for lightly threaded or low-utilization tasks.

For handheld gaming and portable console use cases, the Z2 Extreme delivers the performance headroom needed for higher fidelity settings and smoother frame rates. The Z2 A, by contrast, prioritizes efficiency and battery life over raw output. The data indicates a trade-off: the Extreme part wins on all performance metrics, while the Z2 A wins on power draw.

Architecture Differences

The two GPUs come from different architectural generations. The Z2 A uses RDNA 2.0, while the Z2 Extreme uses RDNA 3.5. This generational gap explains several key differences in feature support and efficiency.

The Z2 A is built on TSMC's 7 nm process, while the Z2 Extreme uses a 4 nm process. The newer process node allows the Z2 Extreme to pack far more transistors into its die: 34,000 million versus 2,400 million. The transistor density tells a similar story, with the Z2 Extreme at 145.9M transistors per square millimeter versus 14.7M for the Z2 A. The Z2 Extreme's die is larger at 233 mm² compared to 163 mm², but the density increase is far more dramatic than the physical size increase.

The shading units, texture mapping units, and render output units all scale accordingly. The Z2 Extreme doubles the shading units (1024 versus 512), doubles the TMUs (64 versus 32), and triples the ROPs (48 versus 16). Ray tracing cores also double from 8 to 16.

Memory technology differs as well. The Z2 A uses LPDDR5 with a 6.4 Gbps effective data rate, while the Z2 Extreme uses LPDDR5X with an 8 Gbps effective rate. Both run on a 128-bit bus, but the higher data rate on the Extreme part yields 128.0 GB/s of bandwidth versus 102.4 GB/s.

FP16 processing differs in ratio. The Z2 A delivers FP16 at a 2:1 ratio relative to FP32, meaning 3.277 TFLOPS. The Z2 Extreme delivers FP16 at a 1:1 ratio, matching its FP32 at 5.530 TFLOPS. This indicates the Extreme part has more balanced half-precision throughput.

Both GPUs support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Both feature a single USB Type-C display output. The Z2 Extreme does not require any power connectors, while the Z2 A's connector information is not listed.

Specification Differences

The two GPUs differ across nearly every core specification. The Z2 A has a base clock of 1000 MHz and a boost clock of 1600 MHz. The Z2 Extreme has a base clock of 800 MHz and a boost clock of 2700 MHz. Memory clocks also differ: 800 MHz with 6.4 Gbps effective for the Z2 A, and 1000 MHz with 8 Gbps effective for the Z2 Extreme.

Shading units: 512 for the Z2 A, 1024 for the Z2 Extreme. TMUs: 32 versus 64. ROPs: 16 versus 48. Ray tracing cores: 8 versus 16. FP32 performance: 1.638 TFLOPS versus 5.530 TFLOPS. FP16 performance: 3.277 TFLOPS (2:1) versus 5.530 TFLOPS (1:1).

Pixel rate: 25.60 GPixel/s versus 129.6 GPixel/s. Texture rate: 51.20 GTexel/s versus 172.8 GTexel/s. TDP: 15 W versus 28 W. Process node: 7 nm versus 4 nm. Transistors: 2,400 million versus 34,000 million. Die size: 163 mm² versus 233 mm². Transistor density: 14.7M per mm² versus 145.9M per mm².

Memory type: LPDDR5 versus LPDDR5X. Memory bandwidth: 102.4 GB/s versus 128.0 GB/s. Both have 16 GB capacity and a 128-bit bus. Both have one USB Type-C display output. The Z2 Extreme lists no power connectors, while the Z2 A does not list its connector configuration.

Release dates differ as well. The Z2 A entered production with a release date at the end of 2024, while the Z2 Extreme followed in mid-2025. Both are listed as Active in production status.

Head-to-Head Benchmarks

The database contains a single benchmark result for the Z2 Extreme and none for the Z2 A. The Z2 Extreme scores 516 in the 3DMark Steel Nomad DX12 test. With no recorded scores for the Z2 A, a direct head-to-head comparison from benchmark data is not possible.

The Z2 Extreme's nearest rivals in the database provide context. The Intel HD Graphics P4000 scores 534, which is 3.4% higher than the Z2 Extreme. The AMD Radeon HD 6870 scores 536, 3.7% higher. The AMD Radeon HD 6750M scores 484, which is 6.6% lower than the Z2 Extreme. The AMD Radeon HD 6770M scores 569, 9.3% higher.

These results place the Z2 Extreme in a narrow performance band alongside older discrete and integrated graphics solutions. The 1st percentile ranking among all GPUs confirms that this is a low-end performer in absolute terms, despite its modern architecture and relatively high transistor count.

The FP32 compute gap between the two chips is the most telling number. At 5.530 TFLOPS, the Z2 Extreme delivers 3.38 times the FP32 throughput of the Z2 A's 1.638 TFLOPS. The pixel rate gap is even larger: 5.06 times. The texture rate gap is 3.38 times.

These differences suggest that in any GPU-bound workload, the Z2 Extreme will substantially outperform the Z2 A. The Z2 A's lower TDP of 15 W versus 28 W means it operates in a different efficiency class, but the performance disparity is so large that the Extreme part's higher power draw is justified for performance-oriented designs.

The Verdict

The data clearly separates these two parts by intended use. The AMD Ryzen Z2 Extreme GPU is the performance option. It delivers roughly triple the FP32 compute, more than five times the pixel fill rate, and 25% more memory bandwidth compared to the Z2 A. Its 1024 shading units, 64 TMUs, and 48 ROPs provide a balanced configuration for modern game rendering. The 4 nm process and RDNA 3.5 architecture represent a newer generation with higher transistor density.

The AMD Ryzen Z2 A GPU is the efficiency option. Its 15 W TDP is nearly half that of the Z2 Extreme, making it suitable for ultra-portable or passively cooled designs where power draw is the primary constraint. Its RDNA 2.0 architecture on 7 nm is older, and its 512 shading units and 16 ROPs limit its performance ceiling. The 102.4 GB/s memory bandwidth is sufficient for light workloads but will bottleneck in demanding scenarios.

For a handheld gaming device where frame rate and graphical fidelity matter most, the Z2 Extreme is the obvious choice. Its benchmark score of 516 in Steel Nomad, while modest in absolute terms, represents the higher-performing of the two chips. For a device prioritizing battery life and minimal heat output, the Z2 A fits that role. The absence of benchmark data for the Z2 A means its real-world performance cannot be quantified from the database, but its specifications indicate a much lower performance tier.

Buyers choosing between these two should base the decision on the power envelope and performance requirements of the target device. The Z2 Extreme spends its 28 W budget to produce 5.530 TFLOPS. The Z2 A uses 15 W to produce 1.638 TFLOPS. The efficiency ratio favors the Extreme part slightly, but the absolute numbers tell the story: the Z2 Extreme is the only one of the two designed for serious gaming performance.