The AMD Ryzen AI 5 Pro 340 is designed for both laptop and desktop platforms, built on a 4nm semiconductor process that helps keep power consumption in check at a TDP of 28W. It includes integrated graphics and fully supports 64-bit computing, while the maximum rated CPU temperature sits at 100°C. Connectivity is handled through PCIe 4.0, offering a solid interface for compatible expansion and storage devices.
The processor operates across six cores using big.LITTLE technology, pairing two groups of cores running at base speeds of 2GHz each, with a turbo clock reaching up to 4.8GHz for demanding workloads. It offers 12 threads in total and carries a clock multiplier of 20, though the multiplier is locked, meaning manual overclocking is not supported. On the cache side, 6MB of L2 and 16MB of L3 cache are available to help reduce memory latency during sustained workloads.
In PassMark testing, the processor achieves a multi-threaded score of 19,500, reflecting its capacity for parallel workloads across all available threads. The single-threaded result of 3,830 indicates how it performs when only one core is engaged, which is relevant for tasks that rely heavily on per-core throughput rather than parallelism.
The integrated graphics solution is the Radeon 840M, which reaches a turbo frequency of 2900MHz and can drive up to four displays simultaneously. It supports DirectX 12 for modern rendering workloads alongside OpenGL 4.6 and OpenCL 2.1, the latter being relevant for GPU-accelerated compute tasks.
The processor supports DDR5 memory running at speeds of up to 8000MHz across a dual-channel configuration, allowing for balanced bandwidth across both channels. It can address a maximum of 256GB of RAM, and notably includes support for ECC memory, which provides error-correcting capability useful in reliability-sensitive environments.
The processor supports multithreading and includes the NX bit for hardware-level memory protection against certain classes of malicious code execution. Its instruction set support spans MMX, SSE 4.1, SSE 4.2, AVX, AVX2, F16C, FMA3, and AES, covering a broad range of workloads from media processing and floating-point operations to hardware-accelerated encryption, with AVX2 in particular enabling wider vectorized data processing across compatible applications.