Tech News
← Home  ·  All topics

Neoverse

2 GoKawiil briefs on this topic

New NEON kernel speeds up llama.cpp prompt processing on Arm CPUs up to 6.5x

A community patch to llama.cpp (164 lines across two files) adds an Arm NEON inner kernel using the SDOT instruction for the tiled K-quant matmul routine, which previously lacked an Arm-specific path for formats like Q3_K, Q2_K and the IQ family. Benchmarks on Qwen2.5 1.5B across three Arm Neoverse server CPUs (N1, V1, V2) show prompt-processing speedups ranging from about 1.7x to 6.5x for low-bit quantized formats, with the largest gain on IQ3_M on Neoverse N1; formats that already had an Arm kernel, such as Q4_K_M and Q8_0, show no change, and token generation speed is unaffected.

Arm unveils Neoverse CSS N4 'Ranger' with up to 128 cores on TSMC N3P

Arm has introduced its next-generation Neoverse CSS N4 semi-custom compute subsystem, codenamed Ranger, built on TSMC's N3P process node. The platform allows cloud and chip partners to configure between 8 and 128 Arm Neoverse N4 cores per die, clocked up to 3.8 GHz, with support for DDR5 or LPDDR6 memory, up to 256 MB of L3 cache, and 128 lanes of PCIe 7/6 or CXL 4.0.