# Size
if False or PLOT_ALL:
with plot_settings(half_width=True):
d = read_data(input_dir + f"{name}.csv", 1, 5)
d = convert_to_gb(d)
theo = {}
for x in d:
theo[x] = [bit_to_gb(comp_bloom_bits(x, 10**-20))] * 2
error_plot_mult(
[d, theo],
output_dir + f'size{EXTENSION}',
100000000,
0,
20,
4,
"Capacity [#]",
"Size [GB]",
r"Bloom Filter Size (FP Rate: $1^{{-20}}$, Stored: To Cap., "
r"10 Reps)",
legend=Legend(['Measured', 'Theoretic'], location="upper left"),
adjust=(LEFT_ADJUST, RIGHT_ADJUST, TOP_ADJUST, BOTTOM_ADJUST),
xticks=xticks[:],
xlabels=xlabels[:],
x_label_step=2)
# Query Time
if False or PLOT_ALL:
with plot_settings(half_width=True):
d = read_data(input_dir + f"{name}.csv", 1, 7)
# d = convert_to_minutes(d)
error_plot_mult(
[d],
# Fit lengths
# new_data = {}
# step = xticks[-1] / len(data.keys())
# for i, key in enumerate(sorted(data.keys())):
# new_data[(i + 1) * step] = data[key]
# data = new_data
for i in range(len(xticks[:]) - 1, -1, -1):
if xticks[i] < 0:
del xticks[i]
del x_labels[i]
xstep = 0.5
min_y = 0
max_y = 20
y_step = 5
print(xticks, x_labels)
error_plot_mult([data],
output_file,
xstep,
min_y,
max_y,
y_step,
xlabel,
ylabel,
title,
xticks=xticks,
xlabels=x_labels,
ver_grid=True,
x_rotation=90,
auto_ylabels=True)
x_is_float=eval['x_float'])
with open(input_dir + f"{eval['name']}.pyc", 'wb') as f:
pickle.dump(d, f)
for eval in evals:
with plot_settings(half_width=eval['half_width']):
with open(input_dir + f"{eval['name']}.pyc", 'rb') as f:
d = pickle.load(f)
error_plot_mult(
[d],
output_dir + f"{eval['out']}{EXTENSION}",
eval['x_step'],
0,
eval['y_max'] if 'y_max' in eval else 0,
eval['y_step'] if 'y_step' in eval else 0,
x_axis[eval['x']],
y_label,
eval['title'],
adjust=eval['adjust'] if 'adjust' in eval else None,
y_log=eval['y_log'],
auto_ylabels=True if 'auto_y' not in eval else eval['auto_y'],
y_lim=eval['y_lim'] if 'y_lim' in eval else None,
y_lim_bottom=eval['y_lim_bottom']
if 'y_lim_bottom' in eval else None,
xticks=eval['xticks'] if 'xticks' in eval else None,
xlabels=eval['xlabels'] if 'xlabels' in eval else None,
minor_xticks=eval['minor_xticks']
if 'minor_xticks' in eval else None,
minor_xlabels=eval['minor_xlabels']
if 'minor_xlabels' in eval else None,
)
hatches=hatches,
y_lim=62)
# RAM Plot
if 0 or PLOT_ALL:
output_file = output_dir + f'provider_ram{EXTENSION}'
xlabel = "Uploads [#]"
title = "Title"
d = read_ram_max(input_dir + 'butthead_provider_uploads_ram' + '.csv', 2,
4)
convert_mib_to_gb(d)
error_plot_mult(
[d],
output_file,
100,
0,
35,
5,
xlabel,
"RAM Usage [GB]",
r"Title",
auto_ylabels=True,
)
# -----------------------------------------------------------------------------
# -----------------------------------------------------------------------------
# -----------------------------------------------------------------------------
# Random Length Dependence
# -----------------------------------------------------------------------------
# -----------------------------------------------------------------------------
name = "butthead_provider_record_length"
output_dir = output_dir_provider + "rand_length/"
os.makedirs(output_dir, exist_ok=True)
output_dir = OUTPUT_DIR + 'bloom_insert_dep/'
input_dir = INPUT_DIR + 'bloom_full_insert/'
name = "butthead_bloom_insert"
xticks = [i for i in range(0, 10**9 + 1, 10**8)]
xlabels: List[str] = [0] + [f"{round(i / 10 ** 9, 1)} Bil" for i in xticks[1:]]
BOTTOM_ADJUST = 0.22
RIGHT_ADJUST = 0.935
TOP_ADJUST = 0.97
LEFT_ADJUST = 0.15
# Insert Time
if False or PLOT_ALL:
with plot_settings(half_width=True):
d = read_data(input_dir + f"{name}.csv", 3, 6)
d = convert_to_minutes(d)
error_plot_mult([d],
output_dir + f'insert_time{EXTENSION}',
100000000,
0,
350,
50,
"Inserted Elements [#]",
"Insertion Time [min]",
r"",
x_label_step=2,
auto_ylabels=True,
adjust=(LEFT_ADJUST, RIGHT_ADJUST, TOP_ADJUST,
BOTTOM_ADJUST),
xticks=xticks,
xlabels=xlabels)
f"{round(i / 10 ** 9, 1)} Bil" for i in
xticks[1:]]
BOTTOM_ADJUST = 0.22
RIGHT_ADJUST = 0.935
TOP_ADJUST = 0.97
LEFT_ADJUST = 0.11
# Query Time
if False or PLOT_ALL:
with plot_settings(half_width=True):
d = read_data(input_dir + f"{name}.csv", 4, 7)
d = convert_to_minutes(d)
error_plot_mult(
[d],
output_dir + f'query_time{EXTENSION}',
100000000,
0,
350,
50,
"Queries [#]",
"Query Time [min]",
r"Time for Query (FP Rate: $1^{{-20}}$, Stored: To Cap., "
r"1Bil Queries, "
"10 Reps)",
x_label_step=2,
auto_ylabels=True,
adjust=(LEFT_ADJUST, RIGHT_ADJUST, TOP_ADJUST, BOTTOM_ADJUST),
xticks=xticks,
xlabels=xlabels
)
d = read_data(input_dir + f"{name}.csv", 2, 5, x_is_float=True)
d = convert_to_gb(d)
theo = {}
for x in d:
theo[x] = [bit_to_gb(comp_bloom_bits(100000000, x))] * 2
error_plot_mult(
[d, theo],
output_dir + f'size{EXTENSION}',
100000000,
0,
1.5,
0.5,
"FP Rate",
"Size [GB]",
"Bloom Filter Size (Capacity: 100Mio, Stored: To Cap., 10 Reps)",
legend=Legend(['Measured', 'Theoretic'], location='upper left'),
adjust=(LEFT_ADJUST, RIGHT_ADJUST, TOP_ADJUST, BOTTOM_ADJUST),
xticks=xticks,
xlabels=xlabels,
minor_xticks=minor_xticks,
minor_xlabels=minor_xlabels,
log_base=10,
x_log=True,
x_sync=True,
reverseX=True,
)
# Query Time
if False or PLOT_ALL:
with plot_settings(half_width=True):
d = read_data(input_dir + f"{name}.csv", 2, 7, x_is_float=True)
print("Runtime at 300ms: ",
sum(data_list[300][10**7]) / len(data_list[300][10**7]), "min")
latencies = sorted(data_list.keys(), reverse=True)
legend_labels = [f"{i}ms" for i in latencies]
legend_labels[-2] = " 50ms"
legend_labels[-1] = " 0ms"
xticks = list(range(0, 10**7 + 1, 2 * 10**6))
error_plot_mult(
[data_list[i] for i in latencies],
output_dir + f'{name}{EXTENSION}',
2 * 10**6,
0,
5,
1,
"PSI Set Size [#]",
"Time [min]",
"PSI Execution Time w/ Latency(No TLS/MAL)",
adjust=(0.11, 0.935, 0.76, 0.21),
# for 2/3 Height (0.19, 0.93, 0.98, 0.16),
legend=Legend(legend_labels, location='above'),
x_label_step=1,
y_lim=5,
xticks=xticks,
xlabels=[0] + [f"{i // 10 ** 6} Mio" for i in xticks[1:]])
# Bandwidth WITHOUT TLS -------------------------------------------------------
if False or PLOT_ALL:
with plot_settings(half_width=True):
name = "butthead_psi_bandwidth"
data_list = read_data_mult(input_dir + f"bandwidth/{name}.csv", 2, 10,
8)
bws = sorted(data_list.keys())[1:]