Add: Benchmark for SciPy

python/benchmark.ipynb

@@ -0,0 +1,177 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import simsimd as simd\n",
+    "from scipy.spatial.distance import cosine, sqeuclidean"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "count = 1000 # Common number of documents in a batch\n",
+    "ndim = 1536 # OpenAI Ada v2 embeddings API returns 1536-dim vectors\n",
+    "A = np.random.randn(count, ndim).astype(np.float32)\n",
+    "B = np.random.randn(count, ndim).astype(np.float32)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "3.35 ms ± 1.59 ms per loop (mean ± std. dev. of 10 runs, 1 loop each)\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%timeit -n 1 -r 10\n",
+    "result_np = [sqeuclidean(A[i], B[i]) for i in range(count)]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "418 µs ± 8.86 µs per loop (mean ± std. dev. of 10 runs, 1 loop each)\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%timeit -n 1 -r 10\n",
+    "result_simd = simd.sqeuclidean(A, B)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "A = np.random.randn(count, ndim).astype(np.float16)\n",
+    "B = np.random.randn(count, ndim).astype(np.float16)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "11.4 ms ± 387 µs per loop (mean ± std. dev. of 10 runs, 1 loop each)\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%timeit -n 1 -r 10\n",
+    "result_np = [sqeuclidean(A[i], B[i]) for i in range(count)]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "442 µs ± 9.22 µs per loop (mean ± std. dev. of 10 runs, 1 loop each)\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%timeit -n 1 -r 10\n",
+    "result_simd = simd.sqeuclidean(A, B)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "A = np.random.randint(-100, 100, (count, ndim), np.int8)\n",
+    "B = np.random.randint(-100, 100, (count, ndim), np.int8)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "5.41 ms ± 1.27 ms per loop (mean ± std. dev. of 10 runs, 1 loop each)\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%timeit -n 1 -r 10\n",
+    "result_np = [sqeuclidean(A[i], B[i]) for i in range(count)]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "241 µs ± 3.79 µs per loop (mean ± std. dev. of 10 runs, 1 loop each)\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%timeit -n 1 -r 10\n",
+    "result_simd = simd.sqeuclidean(A, B)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "base",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.11"
+  },
+  "orig_nbformat": 4
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

python/test.py

@@ -59,7 +59,7 @@ def test_cosine(ndim, dtype):
 
 
 @pytest.mark.parametrize("ndim", [3, 97, 1536])
-@pytest.mark.parametrize("dtype", [np.float32])
+@pytest.mark.parametrize("dtype", [np.float32, np.float16])
 def test_batch(ndim, dtype):
     """Compares the simd.simd.sqeuclidean() function with scipy.spatial.distance.sqeuclidean() for a batch of vectors, measuring the accuracy error for f16, and f32 types."""