db

Database benchmark

Description

# Python Code for Description
# import ...

def write():
    data_dir = "data/numpy"
    # ...
    m_vec = np.random.rand(100 * 500)
    m_vec[9301] = 0.1
    for i in range(100):
        for j in range(500):
            m = m_vec[i * 500 + j]
            matrix = np.random.rand(500, 4)
            np.savez(f"{data_dir}/run_{i:03d}/trial_{j:04d}/data", m=m, matrix=matrix)

def update():
    data_dir = "data/numpy"
    for i in range(100):
        for j in range(500):
            data = np.load(f"{data_dir}/run_{i:03d}/trial_{j:04d}/data.npz")
            if data["m"] == 0.1:
                matrix = np.zeros((500, 4))
                np.savez(f"{data_dir}/run_{i:03d}/trial_{j:04d}/data", m=0.1, matrix=matrix)

def read():
    dir = "data/numpy"
    for i in range(100):
        for j in range(500):
            npz_dir = f"{dir}/run_{i:03d}/trial_{j:04d}/data.npz"
            data = np.load(npz_dir)
            if data["m"] == 0.1:
                print(f"id: {i:03d}, j: {j:04d}, m: {data['m']}")
                print(data["matrix"][0,:])

Hardware specification

• CPU: 13th Gen Intel i5-1340P (16) @ 4.600GHz
• Kernel: 6.5.13-7-MANJARO

Compilation option

• Rust : cargo build --release

Specific version

• Rust : rustc 1.75.0
  • peroxide : 0.34.4
  • serde : 1.0.196
  • nativedb : 0.5.3
  • jammdb : 0.11.0
• Python : Python 3.11.6
  • numpy : 1.26.2

Result

Result (Add HashMap for Parameter Scan)