Commit 9a2eef93 authored by Martin Larralde's avatar Martin Larralde
Browse files

Update `figure2.svg` to use latest benchmark results

parent 63fa930e
......@@ -58,13 +58,9 @@ for color, (backend, group) in zip(
Y = numpy.array([r["mean"] for r in group])
reg = scipy.stats.linregress(X, Y)
plt.plot([ 0, max(X) ], [ reg.intercept, reg.slope*max(X) + reg.intercept ], color=color, linestyle="--", marker="")
<<<<<<< HEAD
# ci = [1.96 * r["stddev"] / math.sqrt(len(r["times"])) for r in group]
plt.scatter(X, Y, marker="+", color=color, label=f"{backend} (R²={reg.rvalue**2:.3f})")
=======
ci = [1.96 * r["stddev"] / math.sqrt(len(r["times"])) for r in group]
plt.errorbar(X, Y, ci, linestyle='', marker="+", color=color, elinewidth=0.3, ecolor='black', label=f"{backend} ($R^2$={reg.rvalue**2:.3f})")
>>>>>>> d7a9c03 (Update connection scoring benchmark plot to show R² instead of R)
plt.legend()
plt.xlabel("Nucleotide count (Mbp)")
plt.ylabel("Time (s)")
......
../benches/connection_scoring/v0.6.4.svg
\ No newline at end of file
../benches/connection_scoring/v0.7.2.svg
\ No newline at end of file
......@@ -129,8 +129,9 @@ the heuristic filter in Pyrodigal saves about half of the time needed to score
connections between all the nodes of a sequence.
![Evaluation of the connection scoring performance with different heuristic
filter SIMD backends (SSE2 or AVX2) or without enabling the filter (None).
*Each sequence was processed 10 times on a quiet i7-8550U CPU @ 1.80GHz*. \label{fig:benchmark}](figure2.svg){width=100%}
filter SIMD backends (SSE2 or AVX2), with a generic backend (Generic) or without enabling
the filter (None).
*Each sequence was processed 10 times on a quiet i7-10710U CPU @ 1.10GHz*. \label{fig:benchmark}](figure2.svg){width=100%}
# Availability
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment