Check out what you can do with Wukong
Wukong's pay-per-use pricing keeps costs low.
Wukong can be easily deployed in seconds on home computers or in the cloud.
Wukong is open-source. Start contributing today.
Wukong uses serverless computing to scale to thousands of executors in seconds.
From machine learning & linear algebra to data analytics, Wukong can do it all.
Wukong delivers best-in-class end-to-end performance for a variety of workloads.
Wukong is open source. Head over to GitHub to get started today.
# Generate random input data. X = da.random.random((128000, 100), chunks = (10000, 100)) # Prepare SVD computation. u, s, v = da.linalg.svd(X) # Begin execution. result = u.compute() result = s.compute() result = v.compute()
# Generate random input data. X = da.random.random((10000, 10000), chunks = (2000, 2000)) # Prepare SVD computation. u, s, v = da.linalg.svd_compressed(X, k = 5) # Begin execution. result = u.compute() result = s.compute() result = v.compute()
# Generate random input data. X = da.random.random((10000, 10000), chunks = (1000, 1000)) # Prepare GEMM computation. XX = da.matmul(X, X) # Begin execution. result = XX.compute()
# Generate random input data. X = da.random.random((32768, 128), chunks = (8192, 128)) # Prepare GEMM computation. q,r = da.linalg.tsqr(X) # Begin execution. res_q = q.compute() res_r = r.compute()
# Prepare classifier. X, y = sklearn.datasets.make_classification(n_samples=1000) clf = ParallelPostFit(SVC(gamma='scale')) clf.fit(X, y) # Prepare the workload and begin execution. X, y = dask_ml.datasets.make_classification(n_samples = 1024000, random_state = 1024000, chunks = 10240) result = clf.predict(X).compute()
import pandas as pd # Read dataframe from file. df = pd.read_csv('employees.csv') # Create Dask dataframe. dask_df = dask.dataframe.from_pandas(df, npartitions = 64) # Prepare and perform operation. filtered_df = dask_df[dask_df['salary'] > 50000] filtered_df.compute()