AMD Ryzen Embedded R1606G

The AMD Ryzen Embedded R1606G is a 2-core, 4-thread mobile processor built on the 14 nm Zen architecture, released on 2020-02-24. With a base clock of 2.60 GHz and a boost clock of 3.50 GHz, this chip sits in the 35th percentile of all CPUs, carrying an average benchmark score of 1215. Its benchmark results place it in a tight cluster with several Intel rivals, where the data shows a near-perfect statistical tie, making its positioning a matter of workload nuance rather than raw dominance.

How It Compares

Against the Intel Core i5-3570, the R1606G posts an average score of 1215 versus 1214, a delta of 0.1% in favor of the AMD part. This is effectively a dead heat, meaning the two chips deliver interchangeable aggregate performance. The i5-3570 is a desktop-class quad-core from an older generation, but the data does not show the R1606G trailing it in any meaningful way — the Zen architecture’s efficiency compensates for the core-count deficit.

The Intel Core i3-7300T is another near-identical rival, scoring 1213 against the R1606G’s 1215, again a 0.1% gap. The i3-7300T is a low-power dual-core with hyper-threading, similar in thread count to the R1606G, yet the AMD chip edges ahead in the aggregate. This suggests that clock-for-clock and thread-for-thread, the R1606G holds a slight performance lead, though the difference is within measurement noise.

The Intel Xeon E5-1607 v3 scores 1217, giving it a -0.2% delta relative to the R1606G, meaning the Xeon is marginally faster. The E5-1607 v3 is a server-grade quad-core without hyper-threading, and its higher average score likely stems from a stronger multi-core ceiling. However, the 0.2% gap is trivial, so the R1606G remains competitive even against a chip from a higher-tier market segment.

The Intel Pentium Gold G6500 also scores 1217, matching the Xeon’s delta of -0.2%. This is a modern desktop dual-core with hyper-threading, and the data shows it outperforming the R1606G by a whisker. The G6500’s newer architecture and higher boost clock likely explain the small edge, but again, the difference is not statistically compelling.

Power and Thermals

The R1606G carries a TDP class of 25 watts, which places it firmly in the ultra-low-power segment. This TDP figure is a defining characteristic, as it allows the processor to operate in fanless or passively cooled designs, thin-and-light embedded systems, and industrial PCs where thermal headroom is scarce. A 25-watt TDP implies that a capable air cooler — even a small heatsink with a low-profile fan — is sufficient for sustained operation. The data does not provide specific temperature figures, but the power envelope suggests that thermal throttling should be minimal under typical embedded workloads, provided the chassis allows basic airflow. The 14 nm process node from GlobalFoundries, with a die size of 148 mm² and 3,500 million transistors, contributes to this efficiency, though the architectural design is more decisive than the node alone.

Single-Thread vs Multi-Thread Behavior

The benchmark split reveals a striking imbalance: in Cinebench R20, the single-core score is 209, while the multi-core score is 1482, a ratio of roughly 7.1x. This means the R1606G scales exceptionally well across its two cores, with multi-threaded performance far exceeding what the single-thread result would predict. In Cinebench R23, the single-core score is 498 and multi-core is 3530, a ratio of 7.1x again, confirming consistent scaling. For real workloads, this implies that the processor is disproportionately strong in multi-threaded tasks like compiling, rendering, or batch processing, whereas single-threaded responsiveness — such as UI interaction or lightly threaded games — will lag behind its multi-core capability. The 2-core, 4-thread configuration allows the scheduler to extract near-linear gains from parallel code, but the low single-core absolute scores (498 in R23) indicate that latency-sensitive tasks will not feel snappy. The boost clock of 3.50 GHz is modest, and the lack of a higher frequency ceiling shows in the single-thread results, which sit near the bottom of modern desktop chips.

FAQ

Q: Does the R1606G support ECC memory?

A: No, the FACT PACK lists ECC memory as false, so it does not support error-correcting code memory.

Q: What is the memory bus width and bandwidth?

A: The R1606G uses dual-channel DDR4 memory with a bandwidth of 38.4 GB/s.

Q: How many PCIe lanes does the CPU provide?

A: The CPU provides 8 PCIe Gen 3 lanes, as stated in the FACT PACK.

Q: What is the integrated graphics solution?

A: It includes Radeon Vega 3 integrated graphics, which is part of the processor package.

Q: Is the multiplier unlocked for overclocking?

A: No, the multiplier is not unlocked, so overclocking is not supported.

Q: What is the production status of this processor?

A: The production status is active, meaning it is currently available for embedded designs.

Benchmark Performance

The Cinebench R15 multicore score is 355, which is a modest figure for a modern CPU but consistent with a 25-watt TDP. The Cinebench R20 multicore score jumps to 1482, and the single-core score is 209. In Cinebench R23, the multicore score reaches 3530, while single-core is 498. These numbers, when compared to the nearest rivals, show that the R1606G is not a performance leader but a balanced performer. The average score of 1215 places it in the 35th percentile of all CPUs, meaning it outperforms roughly a third of the database. Against the Intel Core i5-3570, the delta is +0.1%, so the R1606G is effectively identical in aggregate performance. The Intel Xeon E5-1607 v3 is -0.2% faster, a gap that is negligible in real-world use. The Pentium Gold G6500 also holds a -0.2% advantage, again trivial. The key takeaway is that the R1606G’s multi-core scaling does not translate into a lead over these rivals; it merely keeps pace. The single-core scores, however, are where the R1606G loses ground — the Cinebench R20 single-core of 209 is far below what a modern desktop chip would post, meaning any workload that cannot parallelize will expose the processor’s limitations.

Platform and Compatibility

The R1606G uses the AMD Socket FP5, which is a BGA-style socket designed for embedded and mobile applications. This means the processor is soldered to the board, so there is no upgrade path for the CPU itself — the entire board must be replaced to move to a faster chip. Memory support is DDR4 in a dual-channel configuration, with a bandwidth of 38.4 GB/s, which is adequate for embedded workloads but not exceptional. ECC memory is not supported, which may be a consideration for reliability-critical applications, though the FACT PACK does not elaborate on the implications. The PCIe support is Gen 3 with 8 lanes from the CPU, which is sufficient for a single GPU or a few NVMe drives, but limited compared to desktop platforms with more lanes. The integrated Radeon Vega 3 graphics provides basic display output and light GPU acceleration, reducing the need for a discrete card in many embedded designs. The production status is active, meaning the chip is still available for new designs, but the socket’s soldered nature means the upgrade path is essentially zero — you buy the R1606G for a specific board and stick with it.

Who Should Consider It

The R1606G is best suited for multi-threaded embedded workloads where the 25-watt TDP and compact footprint are paramount. Its Cinebench R23 multicore score of 3530 is respectable for a dual-core, and the 7.1x scaling ratio from single to multi-core indicates that software designed to use all four threads will run efficiently. For industrial automation, digital signage, or network appliances that run parallel processing tasks, the R1606G offers a solid balance of performance and power consumption. Gamers should be cautious: the single-core scores (209 in R20, 498 in R23) are low, and the Radeon Vega 3 iGPU is not designed for modern titles, so this is not a gaming processor. The 0.1% delta over the Intel Core i5-3570 and 0.1% over the Core i3-7300T means that for general office tasks — word processing, spreadsheets, web browsing — the R1606G is indistinguishable from those older chips, but its 25-watt TDP makes it far more thermally manageable in fanless enclosures. Content creation, such as video encoding or 3D rendering, would benefit from the multi-core scaling, but the absolute scores are too low for professional use; the 3530 in R23 multicore is below what a modern quad-core would achieve. The R1606G is a niche part, and its data shows it excels in scenarios where power efficiency trumps peak performance. If the workload is predominantly single-threaded, the data suggests looking elsewhere, as the 498 single-core score in R23 will feel sluggish. For embedded designers who need a reliable, active-production chip with a proven architecture, the R1606G is a defensible choice, provided the expectations align with its mid-pack percentile standing.