Qualcomm, a company well-known for its mobile processors, is expanding its reach beyond smartphones. While primarily recognized for chips found in mid-range and high-end mobile devices, and recently venturing into the laptop market with their Snapdragon X Elite processors, Qualcomm has now announced a significant new CPU specifically for Artificial Intelligence (AI) tasks, dubbed the Qualcomm Dragonfly C1000.
In a world increasingly focused on AI, it’s no surprise that manufacturers are orienting their products towards this growing field. Companies like NVIDIA, traditionally known for their gaming graphics cards, have shifted their focus to prioritizing AI graphics, a move that has proven immensely profitable. This trend is not limited to NVIDIA; other companies are also making similar strategic adjustments to their business models.
Qualcomm Announces the Dragonfly C1000 CPU with Over 250 Oryon Cores
Companies such as Qualcomm, which have historically specialized in mobile phones and tablets, have identified a significant market opportunity in AI. Their foray into this sector began with the Snapdragon X Elite processors for laptops. These chips were designed to accelerate AI tasks, featuring a 40 TOPs Neural Processing Unit (NPU) and integration into the PC Copilot+ ecosystem with exclusive Windows 11 features. Not content with this niche, Qualcomm has now developed the Dragonfly C1000 CPU, intended not only for AI tasks but also described as a general-purpose CPU.
While exact specifications are still under wraps, Qualcomm has revealed that the C1000 will boast over 250 Oryon CPU Cores, utilizing a custom architecture of high-performance and high-efficiency cores. These cores are said to be capable of exceeding 5 GHz frequencies, which will significantly benefit tasks like agent orchestration and server/data center workloads.
The C1000 CPU Will Support PCIe 7.0 and Launch in 2028
Beyond its substantial core count and high clock speeds, the Dragonfly C1000 features an architecture prepared for PCIe Gen 7 and CXL, promising data transfer speeds of up to 2 TB/s. This is a remarkable leap, as PCIe 7.0 is still a future technology; current hardware largely supports PCIe 5.0. The reason for this advanced planning is that the CPU itself is not yet ready for market, with its release slated for 2028. At that time, companies like Meta are expected to utilize this new processor.
A notable aspect of Qualcomm’s announcement is the claim of double the performance per watt compared to competing server CPUs, especially significant for power-hungry AI tasks in data centers.
Original Article Source: El Chapuzas Informático
