AI term
Accelerated computing
What is Accelerated computing?
Computing that uses specialized processors to perform selected workloads faster than general-purpose processors.
In other languages
- 한국어가속 컴퓨팅
- 특정 작업을 전용 하드웨어로 빠르게 처리하는 컴퓨팅 방식입니다.
- 日本語アクセラレーテッドコンピューティング
- 専用プロセッサーで計算処理を高速化する方式です。
Related Terms
- Custom silicon acceleratorSpecialized computer chip designed to run specific workloads faster or more efficiently than general-purpose processors
- Neural AcceleratorSpecialized hardware designed to speed up the processing of artificial intelligence tasks by efficiently managing matrix-based computations.
- AI acceleratorA specialized hardware component designed to speed up the complex mathematical calculations required for machine learning and neural network processing
- Heterogeneous ComputingThe use of multiple types of processors or specialized cores within a single system to optimize performance and energy efficiency for diverse computational tasks.
- HPCHigh-Performance Computing; the use of supercomputers and parallel processing techniques to solve complex computational problems at high speed
- Scientific computeShared high-performance computing resources used for research workloads requiring substantial processing power
- Tensor processing unitA specialized processor designed to accelerate machine-learning computations.
- Inference chipSpecialized processor designed to run trained AI models efficiently during prediction or generation workloads
- Host-to-device transferMovement of data from a computer’s CPU memory to memory accessible by an accelerator.
- Instruction cacheA small, fast memory that stores machine instructions near a processor for quick access.
- Sparse computingComputing methods that skip operations involving zero or inactive data to reduce processing work.
- Application-level optimizationImproving the speed or resource use of a complete software application by changing its code, configuration, or execution.
The fastest, most reliable AI news, straight to your inbox