AMD Ryzen Z2 Go GPU vs Intel Arc B370 Comparison

The Verdict

The data presents two very different integrated-class graphics solutions. The AMD Ryzen Z2 Go GPU, built on RDNA 2.0, offers a fixed 16 GB of LPDDR5 memory, a 128-bit bus, and a 102.4 GB/s bandwidth, which is a substantial advantage for local data throughput. Its 768 shading units and 48 texture mapping units suggest a design centered on consistent, sustained rendering. Conversely, the Intel Arc B370 uses Xe3-LPG architecture with 1280 shading units and 40 TMUs, which gives it a theoretical raw compute advantage in FP32 (6.144 TFLOPS vs 4.147 TFLOPS), but it relies entirely on system-shared memory with dependent bandwidth, and its measured benchmark score is very low.

The recorded database shows a single benchmark result for the Intel Arc B370: a 3DMark Steel Nomad DX12 score of 1184. This places it at the 5th percentile of all GPUs. Its nearest rivals are all older discrete cards: the ATI Mobility Radeon HD 5570 (score 1186, 0.2% faster), ATI Radeon HD 5770 (score 1190, 0.5% faster), AMD Radeon HD 7650M (score 1192, 0.7% faster), and AMD FirePro M2000 (score 1168, 1.4% slower). The AMD Ryzen Z2 Go GPU has no recorded benchmark scores and no nearest rivals in the database, meaning its percentile rank of 50 is an unverified placeholder rather than a measured result.

For a buyer, the choice depends on whether they value a known, self-contained memory subsystem or a higher theoretical shader count. The AMD part is the safer pick for tasks where memory bandwidth and capacity are fixed and predictable. The Intel part is a gamble: its compute density is higher, but the only recorded performance data shows it trading blows with decade-old discrete cards. No one should select the Intel Arc B370 based on its current benchmark standing; it sits at the 5th percentile, which indicates severe underperformance in the tested workload. The AMD Ryzen Z2 Go GPU, with no recorded data, cannot be recommended on measured performance, but its architecture and memory configuration are more conventional for a portable device.

FAQ

Q: Which GPU has a higher measured benchmark score?

A: Only the Intel Arc B370 has a recorded score in the database: 1184 in 3DMark Steel Nomad DX12. The AMD Ryzen Z2 Go GPU has no benchmark entries, so no direct measured comparison can be made from the recorded data.

Q: How does the Intel Arc B370 compare to its closest rivals in the database?

A: The Intel Arc B370 scores 1184, which is 0.2% behind the ATI Mobility Radeon HD 5570 (1186), 0.5% behind the ATI Radeon HD 5770 (1190), and 0.7% behind the AMD Radeon HD 7650M (1192). It is 1.4% ahead of the AMD FirePro M2000 (1168).

Q: What memory configurations do the two parts use?

A: The AMD Ryzen Z2 Go GPU uses a fixed 16 GB of LPDDR5 across a 128-bit bus, providing 102.4 GB/s of bandwidth. The Intel Arc B370 uses system-shared memory, with a system-dependent bus width and bandwidth, meaning its memory performance depends entirely on the host platform.

Q: Which GPU has a higher FP32 compute rating?

A: The Intel Arc B370 delivers 6.144 TFLOPS FP32, while the AMD Ryzen Z2 Go GPU delivers 4.147 TFLOPS. The Intel part also has a higher FP16 rating at 12.29 TFLOPS (2:1) compared to 8.294 TFLOPS (2:1) for AMD.

Q: What are the process nodes for each chip?

A: The AMD Ryzen Z2 Go GPU uses a 6 nm process from TSMC, with a die size of 208 mm² and 13,100 million transistors. The Intel Arc B370 uses a 3 nm process from Intel, with transistor count and die size listed as unknown.

Q: What is the power draw for each GPU?

A: The AMD Ryzen Z2 Go GPU has a TDP of 28 W. The Intel Arc B370 has a TDP of 25 W. Both use no external power connectors.

Architecture Differences

The AMD Ryzen Z2 Go GPU is based on the Rembrandt+ chip, using the RDNA 2.0 architecture. It is manufactured on a 6 nm process at TSMC, with a die size of 208 mm² and a transistor count of 13,100 million. The transistor density is 63.0 million per square millimeter. The Intel Arc B370 uses the Panther Lake chip with the Xe3-LPG architecture. It is built on a 3 nm process at Intel, though its transistor count and die size are unknown. The process node difference is significant: 6 nm versus 3 nm, which affects density and power characteristics, though the AMD part has a larger die and more transistors than the Intel part has been recorded with.

The memory architecture differs fundamentally. The AMD GPU has a dedicated 16 GB of LPDDR5 on a 128-bit bus, yielding a fixed 102.4 GB/s bandwidth. The Intel GPU uses system-shared memory, with the bus width and bandwidth being system dependent. This means the AMD part has a guaranteed memory pool, while the Intel part can vary dramatically depending on the host device's memory configuration.

In terms of compute resources, the Intel Arc B370 has more shading units (1280 vs 768), but fewer texture mapping units (40 vs 48) and fewer render output units (20 vs 32). The AMD part has 12 ray tracing cores, while the Intel part has 10. Neither has tensor cores listed. The pixel rate for AMD is 86.40 GPixel/s, while Intel is 48.00 GPixel/s. The texture rate is 129.6 GTexel/s for AMD and 96.00 GTexel/s for Intel. These numbers indicate that AMD's architecture is more focused on fill-rate and rasterization throughput, while Intel's design leans toward compute-heavy workloads with more shader units.

Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, so the API feature set is identical. The Intel part is marked as an IGP (integrated graphics processor), with a slot width of IGP and a bus interface of IGP. The AMD part has no listed slot width or bus interface, but its display output is a single USB Type-C. The Intel part's display outputs are listed as portable device dependent.

Specification Differences

The two GPUs differ across nearly every specification field. The AMD Ryzen Z2 Go GPU has a base clock of 800 MHz and a boost clock of 2700 MHz, while the Intel Arc B370 has a base clock of 300 MHz and a boost clock of 2400 MHz. The memory clock for AMD is 800 MHz with 6.4 Gbps effective, while Intel's memory clock is system shared.

Memory size: AMD has 16 GB of LPDDR5; Intel has system shared memory. Memory bus width: AMD has 128 bit; Intel is system shared. Memory bandwidth: AMD has 102.4 GB/s; Intel is system dependent. Shading units: AMD has 768; Intel has 1280. TMUs: AMD has 48; Intel has 40. ROPs: AMD has 32; Intel has 20. Ray tracing cores: AMD has 12; Intel has 10. Pixel rate: AMD has 86.40 GPixel/s; Intel has 48.00 GPixel/s. Texture rate: AMD has 129.6 GTexel/s; Intel has 96.00 GTexel/s. FP32: AMD has 4.147 TFLOPS; Intel has 6.144 TFLOPS. FP16: AMD has 8.294 TFLOPS; Intel has 12.29 TFLOPS. TDP: AMD has 28 W; Intel has 25 W. Process node: AMD is 6 nm; Intel is 3 nm. Foundry: AMD is TSMC; Intel is Intel. Transistors: AMD has 13,100 million; Intel is unknown. Die size: AMD is 208 mm²; Intel is unknown. Display outputs: AMD has 1x USB Type-C; Intel has portable device dependent.

The release dates also differ: AMD was released on 2024-12-31, while Intel was released on 2026-01-26. Both are marked as Active production status. Neither has a launch MSRP recorded.

Head-to-Head Benchmarks

There are no direct head-to-head benchmark entries in the database, and the win counts are zero for both parts. However, the Intel Arc B370 has a single recorded benchmark: 3DMark Steel Nomad DX12 with a score of 1184. This score places it at the 5th percentile of all GPUs. Its nearest rival comparisons are telling: it is essentially tied with the ATI Mobility Radeon HD 5570 (1186, 0.2% faster), ATI Radeon HD 5770 (1190, 0.5% faster), and AMD Radeon HD 7650M (1192, 0.7% faster). The only GPU it beats is the AMD FirePro M2000 (1168, 1.4% slower). These are all old, low-end discrete mobile or desktop parts from the early 2010s. The data indicates that the Intel Arc B370, in its only measured test, performs at the level of a legacy entry-level card.

The AMD Ryzen Z2 Go GPU has no benchmark scores, no percentile data (the 50th percentile is a placeholder), and no nearest rivals. This means the database contains zero measured performance information for the AMD part. Any comparison between the two on raw benchmark scores is impossible, as only one side has data. The head-to-head section must therefore rely on the architectural and specification differences, which show AMD with a massive fixed memory bandwidth advantage and higher pixel and texture rates, while Intel has a higher shader count and FP32 throughput.

The recorded data for Intel shows a GPU that, despite having 1280 shading units and 6.144 TFLOPS of FP32, only manages a score that rivals a 2010-era ATI Radeon HD 5770. This suggests severe inefficiency or a bottleneck elsewhere, likely in memory bandwidth or driver overhead. The AMD part, with fewer shaders but a dedicated 102.4 GB/s memory path, is structurally better positioned for sustained rendering workloads, even without a recorded score to confirm it.

Where Each One Wins

The AMD Ryzen Z2 Go GPU wins in every category that has recorded data, except for raw compute throughput. It has a higher pixel rate (86.40 GPixel/s vs 48.00 GPixel/s), a higher texture rate (129.6 GTexel/s vs 96.00 GTexel/s), more ROPs (32 vs 20), more TMUs (48 vs 40), and more ray tracing cores (12 vs 10). It also has a dedicated 16 GB memory pool with 102.4 GB/s of bandwidth, whereas the Intel part depends on system memory. For any workload that relies on fill rate, rasterization, or memory bandwidth, the AMD architecture is superior based on the specification data.

The Intel Arc B370 wins on pure compute metrics. It has 1280 shading units versus 768, delivering 6.144 TFLOPS FP32 versus 4.147 TFLOPS. Its FP16 rating is 12.29 TFLOPS versus 8.294 TFLOPS. This gives it a theoretical advantage in compute-heavy tasks such as shader processing or general-purpose GPU compute, if the memory system can feed it. However, its only measured benchmark score, 1184 in 3DMark Steel Nomad DX12, does not reflect this advantage. It sits at the 5th percentile and matches the ATI Radeon HD 5770, a card with far fewer resources. The data shows that the Intel part's compute lead does not translate into benchmark success in the recorded test.

For usage cases, the AMD part is the choice for any scenario requiring consistent memory access: texture-heavy rendering, high-resolution pixel output, or ray tracing workloads, where its 12 RT cores and higher pixel rate matter. The Intel part is the choice only for theoretical compute throughput, and even then, the measured performance does not support it. There is no benchmark win for Intel in the database, so no real-world workload can be credited to it. The AMD part, despite having no recorded scores, wins by default on every measured or specified performance category where a direct comparison exists. The only field where Intel leads is FP32 and FP16 throughput, shading units, and process node (3 nm vs 6 nm), but none of those translate into a recorded benchmark victory.