Add a graphs and stuff to benchmarking-dwarfs - not done, but progress

assets/benchmarking-dwarfs/data/benchmark-data.csv

@@ -0,0 +1,12 @@
+data/mountpoints/dwarfs,25G-null.bin,12.80490s,371.14600µs,96.32895ms,351.30788ms
+data/mountpoints/dwarfs,25G-random.bin,40.71916s,14.15143ms,109.78266ms,3.51396s
+data/mountpoints/dwarfs,100M-polygon.txt,19.11096s,2.95083ms,96.39260ms,480.97789ms
+data/mountpoints/dwarfs,kernel/linux-6.6.58.tar.xz,160.75466ms,1.52300µs,94.55468ms,882.57600µs
+data/mountpoints/fuse-archive-tar,25G-null.bin,24.88932s,393.56800µs,98.66828ms,0ms
+data/mountpoints/fuse-archive-tar,25G-random.bin,24.84052s,397.62600µs,94.52984ms,0ms
+data/mountpoints/fuse-archive-tar,100M-polygon.txt,26.63768s,77.50500µs,96.61561ms,0ms
+data/mountpoints/fuse-archive-tar,kernel/linux-6.6.58.tar.xz,121.50200ms,1.22300µs,93.25915ms,0ms
+data/datasets,25G-null.bin,25.54820s,27.92200µs,96.79632ms,5.51523ms
+data/datasets,25G-random.bin,16.91976s,290.90600µs,97.64200ms,91.13626ms
+data/datasets,100M-polygon.txt,17.98264s,140.88400µs,98.92292ms,94.05722ms
+data/datasets,kernel/linux-6.6.58.tar.xz,88.59571ms,1.39300µs,91.41823ms,949.77100µs

assets/benchmarking-dwarfs/data/printed.txt

@@ -0,0 +1,86 @@
+=== data/mountpoints/dwarfs/25G-null.bin ===
+Sequential read (complete file read): 12.80490s
+Sequential latency (1 byte read): 371.14600µs
+Random read (1024x 1 MiB): 96.32895ms
+Random latency (1024x 1 byte read): 351.30788ms
+
+=== data/mountpoints/dwarfs/25G-random.bin ===
+Sequential read (complete file read): 40.71916s
+Sequential latency (1 byte read): 14.15143ms
+Random read (1024x 1 MiB): 109.78266ms
+Random latency (1024x 1 byte read): 3.51396s
+
+=== data/mountpoints/dwarfs/100M-polygon.txt ===
+Sequential read (complete file read): 19.11096s
+Sequential latency (1 byte read): 2.95083ms
+Random read (1024x 1 MiB): 96.39260ms
+Random latency (1024x 1 byte read): 480.97789ms
+
+=== data/mountpoints/dwarfs/kernel/linux-6.6.58.tar.xz ===
+Sequential read (complete file read): 160.75466ms
+Sequential latency (1 byte read): 1.52300µs
+Random read (1024x 1 MiB): 94.55468ms
+Random latency (1024x 1 byte read): 882.57600µs
+
+[bulk] Testing data/mountpoints/dwarfs/small-files/null
+[bulk] Testing data/mountpoints/dwarfs/small-files/random
+
+=== === === === === === === === === === ===
+
+=== data/mountpoints/fuse-archive-tar/25G-null.bin ===
+Sequential read (complete file read): 24.88932s
+Sequential latency (1 byte read): 393.56800µs
+Random read (1024x 1 MiB): 98.66828ms
+Random latency (1024x 1 byte read): 0ms
+
+=== data/mountpoints/fuse-archive-tar/25G-random.bin ===
+Sequential read (complete file read): 24.84052s
+Sequential latency (1 byte read): 397.62600µs
+Random read (1024x 1 MiB): 94.52984ms
+Random latency (1024x 1 byte read): 0ms
+
+=== data/mountpoints/fuse-archive-tar/100M-polygon.txt ===
+Sequential read (complete file read): 26.63768s
+Sequential latency (1 byte read): 77.50500µs
+Random read (1024x 1 MiB): 96.61561ms
+Random latency (1024x 1 byte read): 0ms
+
+=== data/mountpoints/fuse-archive-tar/kernel/linux-6.6.58.tar.xz ===
+Sequential read (complete file read): 121.50200ms
+Sequential latency (1 byte read): 1.22300µs
+Random read (1024x 1 MiB): 93.25915ms
+Random latency (1024x 1 byte read): 0ms
+
+[bulk] Testing data/mountpoints/fuse-archive-tar/small-files/null
+[bulk] Testing data/mountpoints/fuse-archive-tar/small-files/random
+
+=== === === === === === === === === === ===
+
+=== data/datasets/25G-null.bin ===
+Sequential read (complete file read): 25.54820s
+Sequential latency (1 byte read): 27.92200µs
+Random read (1024x 1 MiB): 96.79632ms
+Random latency (1024x 1 byte read): 5.51523ms
+
+=== data/datasets/25G-random.bin ===
+Sequential read (complete file read): 16.91976s
+Sequential latency (1 byte read): 290.90600µs
+Random read (1024x 1 MiB): 97.64200ms
+Random latency (1024x 1 byte read): 91.13626ms
+
+=== data/datasets/100M-polygon.txt ===
+Sequential read (complete file read): 17.98264s
+Sequential latency (1 byte read): 140.88400µs
+Random read (1024x 1 MiB): 98.92292ms
+Random latency (1024x 1 byte read): 94.05722ms
+
+=== data/datasets/kernel/linux-6.6.58.tar.xz ===
+Sequential read (complete file read): 88.59571ms
+Sequential latency (1 byte read): 1.39300µs
+Random read (1024x 1 MiB): 91.41823ms
+Random latency (1024x 1 byte read): 949.77100µs
+
+[bulk] Testing data/datasets/small-files/null
+[bulk] Testing data/datasets/small-files/random
+
+=== === === === === === === === === === ===

assets/benchmarking-dwarfs/process-data.py

@@ -0,0 +1,144 @@
+#!/usr/bin/env python3
+import csv
+import re
+
+class HelperFunctions:
+    def get_fs(dir):
+        if dir.endswith('dwarfs'):
+            return 'DwarFS'
+        elif dir.endswith('fuse-archive-tar'):
+            return 'fuse-archive (tar)'
+
+        return 'Btrfs'
+
+    def get_label(filename):
+        if filename == '25G-null.bin':
+            return 'Null 25 GiB file'
+        elif filename == '25G-random.bin':
+            return 'Random 25 GiB file'
+        elif filename == '100M-polygon.txt':
+            return '100 million-sided polygon data'
+        elif filename.startswith('kernel'):
+            return 'Linux LTS kernel'
+
+    def convert_time(time: str, unit: str) -> int:
+        unit_exponents = ['ns', 'µs', 'ms', 's']
+
+        if time.endswith('ms'):
+            current_unit = 'ms'
+        elif time.endswith('µs'):
+            current_unit = 'µs'
+        elif time.endswith('ns'):
+            current_unit = 'ns'
+        else:
+            current_unit = 's'
+        
+        unit_multiplier = unit_exponents.index(current_unit) - unit_exponents.index(unit)
+        return HelperFunctions.time_int(time) * (1000 ** unit_multiplier)
+    
+    def time_int(time: str):
+        time = re.sub("[^0-9\\.]", "", time)
+        return float(time)
+
+
+def sequential_latency():
+    datasets = {'labels': []}
+    with open('assets/benchmarking-dwarfs/original-data/benchmark-data.csv', 'rt') as f:
+        for line in csv.reader(f):
+            fs = HelperFunctions.get_fs(line[0])
+            label = HelperFunctions.get_label(line[1])
+            datasets['labels'].append(label) if label not in datasets[
+                'labels'
+            ] else False
+            try:
+                datasets[fs].append(line[3])
+            except KeyError:
+                datasets[fs] = []
+                datasets[fs].append(line[3])
+
+    return datasets
+
+
+def singles():
+    pass
+
+
+def bulk():
+    pass
+
+
+if __name__ == '__main__':
+    # from https://github.com/chartjs/Chart.js/blob/master/docs/scripts/utils.js (CHART_COLORS)
+    # modified so similar color aren't adjacent
+    chart_colors = [
+        "'rgb(255, 99, 132)'",  # red
+        "'rgb(75, 192, 192)'",  # green
+        "'rgb(54, 162, 235)'",  # blue
+        "'rgb(255, 159, 64)'",  # orange
+        "'rgb(153, 102, 255)'",  # purple
+        "'rgb(255, 205, 86)'",  # yellow
+        "'rgb(201, 203, 207)'",  # grey
+    ]
+
+    print('Sequential latency:')
+    labels_code = 'const labels = $labels$'
+    dataset_code = '''
+    {
+      label: '$label$',
+      data: $data$,
+      backgroundColor: $color$,
+    },'''
+
+    config_code = '''
+let config = {
+    type: 'bar',
+    data: {
+        datasets: data,
+        labels
+    },
+    options: {
+      plugins: {
+        title: {
+          display: true,
+          text: '$title$ - in $timeunit$'
+        },
+      },
+      responsive: true,
+      interaction: {
+        intersect: false,
+      },
+    }
+  };
+'''
+
+    data = sequential_latency()
+    labels_code = labels_code.replace('$labels$', format(data['labels']))
+    print(labels_code)
+    data.pop('labels')
+    print('let data = [', end='')
+    largest_time_unit = 'ns'
+    for fs in data.keys():
+        for item in data[fs]:
+            if item.endswith('ms'):
+                largest_time_unit = 'ms'
+            elif item.endswith('µs') and largest_time_unit != 'ms':
+                largest_time_unit = 'µs'
+            elif item.endswith('ns') and largest_time_unit != 'ms' and largest_time_unit != 'µs':
+                largest_time_unit = 'ns'
+            elif re.sub('[0-9]', '', item) == 's':
+                largest_time_unit = 's'
+                break
+
+        for i in range(len(data[fs])):
+            data[fs][i] = HelperFunctions.convert_time(data[fs][i], largest_time_unit)
+        
+        print(
+            dataset_code.replace('$label$', fs)
+            .replace('$data$', format(data[fs]))
+            .replace('$color$', format(chart_colors[list(data.keys()).index(fs)])),
+            end=''
+        )
+    print('\n]\n')
+
+    title = 'Sequential Read Latency'
+    print(config_code.replace('$title$', title).replace('$timeunit$', largest_time_unit))