IBM在最新的量子开发者大会上更新了其技术路线图,明确目标是在本世纪末之前制造出具备纠错能力的量子计算机 1。为了实现这一里程碑,公司计划于2027年将两个下一代Nighthawk r3处理器连接起来,构建一个拥有超过1,000个量子比特的系统 2。尽管实现路径有所调整,计划互连的Nighthawk处理器数量从原定的9个减少至2个,但明年预计提供的量子比特总数保持不变 1。
在硬件规格方面,IBM扩大了Nighthawk通用处理器的性能指标,新增了一个支持超过500个可编程量子比特的版本 1。具体而言,即将推出的Nighthawk r3芯片预计将支持520个可编程量子比特(基于2,034个物理量子比特),并能够执行10,000次量子操作 2。作为对比,现有的Nighthawk r2处理器已展现出高达每秒10万样本的处理速度,支持在120个量子比特上执行超过7,500次操作 2。
实际应用数据显示了新一代处理器的显著加速效果。合作伙伴E.ON在电网优化场景中报告了15倍的计算加速,而与橡树岭国家实验室的合作实验则实现了12倍的速度提升 2。在材料模拟领域,Algorithmiq的实验表明,Nighthawk r2比上一代Heron处理器快8倍,尽管这导致经典计算机模拟所需的内存增加了240倍 2。此外,IBM通过与RIKEN合作,使SQD工作流速度提高了约5倍,同时计算量增加了100倍;Cleveland Clinic和RIKEN还利用自动化处理6.08亿个样本,速度比无自动化时快了33倍 2。
为了进一步优化混合计算工作流,IBM正通过量子平台整合经典高性能计算资源。公司已与橡树岭国家实验室及AMD启动联合设计平台,将Nighthawk处理器与ORNL的Umbra HPC系统共置,以提升整体效率 2。
IBM has updated its quantum computing roadmap at the IBM Quantum Developer's Conference, announcing plans to offer access to a dual-processor quantum system by 2027 12. The company stated that its goal is to deliver an error-corrected quantum computer before the end of the current decade 1. Although initial plans involved interconnecting nine Nighthawk processors, IBM adjusted this architecture to link just two processors instead 1. This revised plan involves connecting two Nighthawk r3 units to form a system with over 1,000 qubits, specifically targeting 1,040 qubits 2. Despite the reduction in the number of interconnected processors, the total number of qubits IBM plans to provide next year remains unchanged 1.
The specifications for the upcoming general-purpose Nighthawk r3 processor have been expanded to include a version with more than 500 programmable qubits 1. This r3 variant supports 520 programmable qubits and 2,034 physical qubits, capable of executing 10,000 quantum operations 2. In comparison, the existing Nighthawk r2 processor has demonstrated processing speeds of up to 100,000 samples per second and can perform more than 7,500 operations on 120 qubits 2.
IBM is also integrating classical high-performance computing resources into its quantum platform to enhance hybrid workflows 2. The company launched joint design platforms with Oak Ridge National Laboratory and AMD, which co-locate Nighthawk systems with ORNL’s Umbra HPC infrastructure 2. Collaborators have reported significant performance gains using these technologies. E.ON achieved a 15-fold acceleration in grid optimization tasks using Nighthawk r2 technology 2. Joint experiments with Oak Ridge National Laboratory resulted in a 12-fold speedup, while Algorithmiq observed an eightfold improvement in material simulations on the r2 compared to previous Heron processors 2. Additionally, Cleveland Clinic and RIKEN utilized automated processing to handle 608 million samples, achieving a 33-fold increase in speed relative to non-automated methods 2.
评论
还没有评论,欢迎留下第一条。