[BUG]: Nx-cuGraph caching not working with documentation example

[beckernick]

trafficstars

Version

24.06

Which installation method(s) does this occur on?

Conda

Describe the bug.

When I ran the documentation example multiple times, I expected performance to be significantly better the second time with caching enabled. Instead, performance was the same and the CPU to GPU conversion was happening every time.

Minimum reproducible example

%env NETWORKX_BACKEND_PRIORITY=cugraph
%env CACHE_CONVERTED_GRAPH=True

import pandas as pd
import networkx as nx

url = "https://data.rapids.ai/cugraph/datasets/cit-Patents.csv"
df = pd.read_csv(url, sep=" ", names=["src", "dst"], dtype="int32")

G = nx.from_pandas_edgelist(df, source="src", target="dst")
env: NETWORKX_BACKEND_PRIORITY=cugraph
env: CACHE_CONVERTED_GRAPH=True


%time result = nx.betweenness_centrality(G, k=10)
CPU times: user 18.1 s, sys: 1.69 s, total: 19.8 s
Wall time: 20.7 s

%time result = nx.betweenness_centrality(G, k=10)
CPU times: user 19.6 s, sys: 859 ms, total: 20.4 s
Wall time: 20.6 s

%%snakeviz

result = nx.betweenness_centrality(G, k=10)

Relevant log output

No response

Environment details

!conda list | grep "cugraph\|networkx"

cugraph                   24.06.00a94     cuda12_py310_240604_g1667f7a43_94    rapidsai-nightly
libcugraph                24.06.00a94     cuda12_240604_g1667f7a43_94    rapidsai-nightly
libcugraph_etl            24.06.00a94     cuda12_240604_g1667f7a43_94    rapidsai-nightly
libcugraphops             24.06.00a24     cuda12_240603_gac69d0a6_24    rapidsai-nightly
networkx                  3.3                pyhd8ed1ab_1    conda-forge
nx-cugraph                24.06.00a94            py310_94    rapidsai-nightly
pylibcugraph              24.06.00a94     cuda12_py310_240604_g1667f7a43_94    rapidsai-nightly

Other/Misc.

No response

Code of Conduct

• [X] I agree to follow cuGraph's Code of Conduct
• [X] I have searched the open bugs and have found no duplicates for this bug report