#cuSPARSE - Twitter Hashtag

about 3 years ago

Such a pleasure to see great team-work move #MLIR sparsification forward with elegant solutions. For example, in this posting we show how progressive lowering can accelerate an idiomatic SDDMM with #cuSparse. https://t.co/l0Atxub8Hh

0

10

0

816

R. M. Caplan

@sumseq

over 4 years ago

We will be presenting at #GTC22 very soon! U can register for free at https://t.co/yWOxi1GqFt, and check out our presentations on recent numerical advances of @predsci model software. @NVIDIAGTC #cuSPARSE #OpenACC @NVIDIAHPCDev

sumseq's tweet photo. We will be presenting at #GTC22 very soon! U can register for free at https://t.co/yWOxi1GqFt, and check out our presentations on recent numerical advances of @predsci model software. @NVIDIAGTC #cuSPARSE #OpenACC @NVIDIAHPCDev https://t.co/GqecTm5aLL

0

4

2

1

0

R. M. Caplan

@sumseq

over 4 years ago

#GTC22 @predsci @NVIDIAHPCDev @NVIDIAGTC @OpenACCorg #OpenACC #cuSparse @fortranlang

0

1

0

Predictive Science Inc. @predsci

over 4 years ago

Next week, some of us will be attending (and presenting at) #GTC22! Register for the meeting at https://t.co/ALGzBSWxyM (for free), and check out these presentations by @sumseq and @Mstulajter on recent numerical advances of PSI model software. @NVIDIAGTC #cuSPARSE

predsci's tweet photo. Next week, some of us will be attending (and presenting at) #GTC22! Register for the meeting at https://t.co/ALGzBSWxyM (for free), and check out these presentations by @sumseq and @Mstulajter on recent numerical advances of PSI model software. @NVIDIAGTC #cuSPARSE https://t.co/RPZP3w4xIP

0

4

2

1

0

Darrin Johnson @darrinpjohnson

over 5 years ago

Starting from #CUDA 11.2.1, #cuSPARSE provides Tensor Core-Accelerated Block-Sparse Matrix Multiplication. It enables near-linear speedup compared to dense computation and mixed-data type computation on NVIDIA Volta, Turing, and Ampere #GPUs. https://t.co/c3OzYqqTgT

0

7

0

NVIDIA HPC Developer

@NVIDIAHPCDev

over 5 years ago

Starting from #CUDA 11.2.1, #cuSPARSE provides Tensor Core-Accelerated Block-Sparse Matrix Multiplication. It enables near-linear speedup compared to dense computation and mixed-data type computation on NVIDIA Volta, Turing, and Ampere #GPUs.