bench(gfql): fair same-HW cross-DB competitor matrix runners (Kuzu/Memgraph/Neo4j/Pokec)#1704
Open
lmeyerov wants to merge 2 commits into
Open
bench(gfql): fair same-HW cross-DB competitor matrix runners (Kuzu/Memgraph/Neo4j/Pokec)#1704lmeyerov wants to merge 2 commits into
lmeyerov wants to merge 2 commits into
Conversation
lmeyerov
force-pushed
the
bench/gfql-fair-matrix
branch
4 times, most recently
from
2645ed8 to
33e5a58
Compare
lmeyerov
changed the base branch from
feat/gfql-index-degrees
to
perf/gfql-row-pushdown
lmeyerov
force-pushed
the
perf/gfql-row-pushdown
branch
from
90db1e3 to
ed6e12c
Compare
lmeyerov
force-pushed
the
bench/gfql-fair-matrix
branch
2 times, most recently
from
62ffb7f to
25bfd21
Compare
lmeyerov
force-pushed
the
bench/gfql-fair-matrix
branch
from
25bfd21 to
481e66c
Compare
lmeyerov
changed the base branch from
perf/gfql-row-pushdown
to
perf/gfql-olap-real-q1q9
lmeyerov
force-pushed
the
bench/gfql-fair-matrix
branch
from
481e66c to
42c8726
Compare
lmeyerov
force-pushed
the
perf/gfql-olap-real-q1q9
branch
from
dc01f38 to
c7b5c71
Compare
lmeyerov
force-pushed
the
bench/gfql-fair-matrix
branch
from
42c8726 to
da2cd7a
Compare
lmeyerov
force-pushed
the
perf/gfql-olap-real-q1q9
branch
2 times, most recently
from
ea3fcd9 to
36cf2ff
Compare
lmeyerov
force-pushed
the
bench/gfql-fair-matrix
branch
from
da2cd7a to
9f10199
Compare
lmeyerov
force-pushed
the
perf/gfql-olap-real-q1q9
branch
from
36cf2ff to
cc531c4
Compare
lmeyerov
force-pushed
the
bench/gfql-fair-matrix
branch
from
9f10199 to
ed8a3ba
Compare
lmeyerov
force-pushed
the
perf/gfql-olap-real-q1q9
branch
from
cc531c4 to
180ad49
Compare
lmeyerov
force-pushed
the
bench/gfql-fair-matrix
branch
from
ed8a3ba to
ed09beb
Compare
lmeyerov
force-pushed
the
perf/gfql-olap-real-q1q9
branch
from
180ad49 to
f7cc095
Compare
lmeyerov
force-pushed
the
bench/gfql-fair-matrix
branch
from
ed09beb to
a7fb73f
Compare
lmeyerov
force-pushed
the
perf/gfql-olap-real-q1q9
branch
from
f7cc095 to
b09397f
Compare
lmeyerov
force-pushed
the
bench/gfql-fair-matrix
branch
from
a7fb73f to
81d0ed9
Compare
lmeyerov
force-pushed
the
perf/gfql-olap-real-q1q9
branch
from
b09397f to
321fe95
Compare
…mgraph/Neo4j/Pokec)
Consolidates the competitor benchmark harnesses onto the index/bench stack (were
misplaced on the cypher-conformance branch). Each competitor runs via its OWN
graph-benchmark repo query (the fairness fix — hand-written queries undersold
Kuzu ~90x). New runners: graph_benchmark_{kuzu,memgraph_q1_q9,neo4j_q1_q9,pokec,
memgraph_neighbors}.py + {compare,polars_q5_q7}.py + {expansion,ldbc,neighbors,
path,pattern,planner}_probe.py + run_graph_benchmark_memgraph_dgx.sh +
docs/source/gfql/benchmark_memgraph.rst. graph_benchmark_q1_q9.py gains the q1
columnar in-degree shortcut + per-query policy metadata. README/RESULTS/
graph_benchmark.md union-merged with the stack's index-harness docs.
Bench-only (no graphistry/ source); the shortestPath-cache perf commit is split
out for a separate PR. Fair-matrix real medians (dgx, q1-q9 100k/2.4M): Kuzu q8
10.4ms / Ladybug q8 19.5ms are genuine WINs; GFQL wins q1(cuDF 19.7ms via the
shortcut)/q2/q4-q7/q9(GPU); Pokec seeded GFQL 8/8 via native indexed hop.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012rRBmezaBEi2CJgDGkMNgR
lmeyerov
force-pushed
the
bench/gfql-fair-matrix
branch
from
81d0ed9 to
b35a0ec
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Consolidates the competitor benchmark harnesses onto the index/bench stack (they were misplaced on the cypher-conformance branch). Each competitor runs via its own
prrao87/graph-benchmarkrepo query — the fairness fix (our hand-written queries undersold Kuzu ~90×).New runners
graph_benchmark_{kuzu,memgraph_q1_q9,memgraph_neighbors,neo4j_q1_q9,pokec}.py,graph_benchmark_{compare,polars_q5_q7}.py,graph_benchmark_{expansion,ldbc,neighbors,path,pattern,planner}_probe.py,run_graph_benchmark_memgraph_dgx.sh,docs/source/gfql/benchmark_memgraph.rst.graph_benchmark_q1_q9.pygains the q1 columnar in-degree shortcut + per-query policy metadata. README/RESULTS/graph_benchmark.md union-merged with the stack's index-harness docs.Scope
Bench-only — no
graphistry/source (the shortestPath-cache perf commit is split to a separate PR). Runnable on dgx viadgx-guard/safe_run.sh.Fair-matrix findings (dgx, q1–q9, 100k/2.42M FOLLOWS)
🤖 Generated with Claude Code