CUDA libraries
E758502
CUDA libraries are a collection of NVIDIA-provided GPU-accelerated software libraries that offer optimized routines for tasks such as linear algebra, deep learning, signal processing, and parallel algorithms on CUDA-enabled hardware.
All labels observed (4)
| Label | Occurrences |
|---|---|
| CUDA libraries canonical | 2 |
| CUDA Toolkit installer | 1 |
| NVIDIA CUDA Toolkit | 1 |
| cuSignal | 1 |
How this entity was disambiguated
This entity first appeared as the object of triple T8823361 — resolving that mention is where its identity was fixed. The disambiguator weighed these candidate entities and picked the highlighted one (or “None”, minting a new entity). This is how homonymy is resolved: the same surface form can point to different entities.
NED1
Entity disambiguation (via context triple)
gpt-5-mini-2025-08-07
Target entity: CUDA libraries Context triple: [CUDA Fortran, supports, CUDA libraries]
-
A.
cuBLAS
cuBLAS is NVIDIA’s GPU-accelerated implementation of the BLAS linear algebra library, providing high-performance matrix and vector operations for CUDA applications.
-
B.
cuDNN
cuDNN is NVIDIA’s GPU-accelerated library of optimized primitives for deep neural networks, widely used to speed up training and inference in frameworks like TensorFlow and PyTorch.
-
C.
NVIDIA CUDA
NVIDIA CUDA is a parallel computing platform and programming model that enables developers to use NVIDIA GPUs for general-purpose high-performance computing.
-
D.
CUDA Driver API
The CUDA Driver API is a low-level programming interface from NVIDIA that gives developers fine-grained control over GPU resources and execution for CUDA applications.
-
E.
cuSPARSE
cuSPARSE is NVIDIA’s GPU-accelerated library providing high-performance sparse linear algebra routines for CUDA applications.
- F. None of above. chosen
- G. Unsure - the case is ambiguous/there is not enough information to decide.
NED2
Entity disambiguation (via description)
gpt-5-mini-2025-08-07
Target entity: CUDA libraries Target entity description: CUDA libraries are a collection of NVIDIA-provided GPU-accelerated software libraries that offer optimized routines for tasks such as linear algebra, deep learning, signal processing, and parallel algorithms on CUDA-enabled hardware.
-
A.
cuBLAS
cuBLAS is NVIDIA’s GPU-accelerated implementation of the BLAS linear algebra library, providing high-performance matrix and vector operations for CUDA applications.
-
B.
cuDNN
cuDNN is NVIDIA’s GPU-accelerated library of optimized primitives for deep neural networks, widely used to speed up training and inference in frameworks like TensorFlow and PyTorch.
-
C.
NVIDIA CUDA
NVIDIA CUDA is a parallel computing platform and programming model that enables developers to use NVIDIA GPUs for general-purpose high-performance computing.
-
D.
CUDA Driver API
The CUDA Driver API is a low-level programming interface from NVIDIA that gives developers fine-grained control over GPU resources and execution for CUDA applications.
-
E.
cuSPARSE
cuSPARSE is NVIDIA’s GPU-accelerated library providing high-performance sparse linear algebra routines for CUDA applications.
- F. None of above. chosen
Statements (52)
| Predicate | Object |
|---|---|
| instanceOf |
GPU-accelerated library suite
ⓘ
software library collection ⓘ |
| component |
CUB
NERFINISHED
ⓘ
NCCL NERFINISHED ⓘ NPP (NVIDIA Performance Primitives) NERFINISHED ⓘ NVJPEG NERFINISHED ⓘ NVRTC NERFINISHED ⓘ TensorRT (CUDA-accelerated inference library) NERFINISHED ⓘ Thrust NERFINISHED ⓘ cuBLAS NERFINISHED ⓘ cuDNN NERFINISHED ⓘ cuFFT NERFINISHED ⓘ cuRAND NERFINISHED ⓘ cuSOLVER NERFINISHED ⓘ cuSPARSE NERFINISHED ⓘ cuSPARSELt NERFINISHED ⓘ cuTENSOR NERFINISHED ⓘ |
| computingPlatform | CUDA NERFINISHED ⓘ |
| developer | NVIDIA NERFINISHED ⓘ |
| domain |
Fourier transforms
ⓘ
collective communication ⓘ linear algebra ⓘ parallel algorithms ⓘ random number generation ⓘ sparse matrix computations ⓘ tensor operations ⓘ |
| integratesWith |
CUDA Toolkit
NERFINISHED
ⓘ
NVIDIA GPU drivers NERFINISHED ⓘ |
| license | proprietary with some components under permissive licenses ⓘ |
| optimization |
GPU acceleration
ⓘ
parallel computing ⓘ |
| platform |
Linux
ⓘ
Windows ⓘ macOS (limited and legacy support) NERFINISHED ⓘ |
| programmingModel | CUDA NERFINISHED ⓘ |
| supportsFeature |
distributed training (via NCCL)
ⓘ
mixed-precision computation ⓘ multi-GPU scaling ⓘ tensor cores (on supported GPUs) ⓘ |
| supportsLanguage |
C
ⓘ
C++ NERFINISHED ⓘ Fortran NERFINISHED ⓘ Python (via bindings) ⓘ |
| targetHardware | CUDA-enabled GPU ⓘ |
| useCase |
data analytics
ⓘ
deep learning ⓘ high-performance computing ⓘ image processing ⓘ machine learning ⓘ numerical linear algebra ⓘ scientific computing ⓘ signal processing ⓘ |
How these facts were elicited
The pipeline generated the facts above by prompting gpt-5.1 with this entity's name + description and the instruction below.
Instruction
You are a knowledge base construction expert. Given a subject entity and a description of it, return factual statements that you know for the subject as a JSON list of dictionaries(triples), where keys must be "subject", "predicate" and "object". The number of facts may be very high, between 25 to 50 or more, for very popular subjects. For less popular subjects, the number of facts can be very low, like 5 or 10. # Requirements - If you don't know the subject at all, return an empty list. - If the subject is not a named entity, return an empty list. - Include at least one triple where predicate is "instanceOf". - Do not get too wordy. - Separate several objects into multiple triples with one object.
Input
Subject: CUDA libraries Description of subject: CUDA libraries are a collection of NVIDIA-provided GPU-accelerated software libraries that offer optimized routines for tasks such as linear algebra, deep learning, signal processing, and parallel algorithms on CUDA-enabled hardware.
Referenced by (5)
Full triples — surface form annotated when it differs from this entity's canonical label.
this entity surface form:
CUDA Toolkit installer
subject surface form:
CUDA
this entity surface form:
cuSignal