Add GF2Multiplication bloq for multiplication over GF($2^m$) (#1436)

* Add GF2Multiplication for multiplication over GF(2^m) * Fix formatting * Address nits
quantumlib · Oct 7, 2024 · 93b7d96 · 93b7d96
1 parent 4937e7d
commit 93b7d96
Show file tree

Hide file tree

Showing 9 changed files with 1,293 additions and 3 deletions.
diff --git a/dev_tools/autogenerate-bloqs-notebooks-v2.py b/dev_tools/autogenerate-bloqs-notebooks-v2.py
diff --git a/docs/bloqs/index.rst b/docs/bloqs/index.rst
@@ -87,6 +87,12 @@ Bloqs Library
     factoring/ecc/ec_add.ipynb
     factoring/ecc/ecc.ipynb
 
+.. toctree::
+    :maxdepth: 2
+    :caption: GF Arithmetic:
+
+    gf_arithmetic/gf2_multiplication.ipynb
+
 .. toctree::
     :maxdepth: 2
     :caption: Rotations:

diff --git a/qualtran/bloqs/bookkeeping/allocate.py b/qualtran/bloqs/bookkeeping/allocate.py
@@ -69,7 +69,7 @@ def adjoint(self) -> 'Bloq':
         return Free(self.dtype, self.dirty)
 
     def on_classical_vals(self) -> Dict[str, int]:
-        return {'reg': 0}
+        return {'reg': self.dtype.from_bits([0] * self.dtype.num_qubits)}
 
     def my_tensors(
         self, incoming: Dict[str, 'ConnectionT'], outgoing: Dict[str, 'ConnectionT']

diff --git a/qualtran/bloqs/gf_arithmetic/__init__.py b/qualtran/bloqs/gf_arithmetic/__init__.py
@@ -0,0 +1,15 @@
+#  Copyright 2024 Google LLC
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#
+#      https://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+from qualtran.bloqs.gf_arithmetic.gf2_multiplication import GF2Multiplication
diff --git a/qualtran/bloqs/gf_arithmetic/gf2_multiplication.ipynb b/qualtran/bloqs/gf_arithmetic/gf2_multiplication.ipynb
@@ -0,0 +1,182 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "87c95c4a",
+   "metadata": {
+    "cq.autogen": "title_cell"
+   },
+   "source": [
+    "# GF($2^m$) Multiplication"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "31c1f087",
+   "metadata": {
+    "cq.autogen": "top_imports"
+   },
+   "outputs": [],
+   "source": [
+    "from qualtran import Bloq, CompositeBloq, BloqBuilder, Signature, Register\n",
+    "from qualtran import QBit, QInt, QUInt, QAny\n",
+    "from qualtran.drawing import show_bloq, show_call_graph, show_counts_sigma\n",
+    "from typing import *\n",
+    "import numpy as np\n",
+    "import sympy\n",
+    "import cirq"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "307679ec",
+   "metadata": {
+    "cq.autogen": "GF2Multiplication.bloq_doc.md"
+   },
+   "source": [
+    "## `GF2Multiplication`\n",
+    "Out of place multiplication over GF($2^m$).\n",
+    "\n",
+    "The bloq implements out of place multiplication of two quantum registers storing elements\n",
+    "from GF($2^m$) using construction described in Ref[1], which extends the classical construction\n",
+    "of Ref[2].\n",
+    "\n",
+    "To multiply two m-bit inputs $a = [a_0, a_1, ..., a_{m-1}]$ and $b = [b_0, b_1, ..., b_{m-1}]$,\n",
+    "the construction computes the output vector $c$ via the following three steps:\n",
+    "    1. Compute $e = U.b$ where $U$ is an upper triangular matrix constructed using $a$.\n",
+    "    2. Compute $Q.e$ where $Q$ is an $m \\times (m - 1)$ reduction matrix that depends upon the\n",
+    "        irreducible polynomial $P(x)$ of the galois field $GF(2^m)$. The i'th column of the\n",
+    "        matrix corresponds to coefficients of the polynomial $x^{m + i} % P(x)$. This matrix $Q$\n",
+    "        is a linear reversible circuit that can be implemented only using CNOT gates.\n",
+    "    3. Compute $d = L.b$ where $L$ is a lower triangular matrix constructed using $a$.\n",
+    "    4. Compute $c = d + Q.e$ to obtain the final product.\n",
+    "\n",
+    "Steps 1 and 3 are performed using $n^2$ Toffoli gates and step 2 is performed only using CNOT\n",
+    "gates.\n",
+    "\n",
+    "#### Parameters\n",
+    " - `bitsize`: The degree $m$ of the galois field $GF(2^m)$. Also corresponds to the number of qubits in each of the two input registers $a$ and $b$ that should be multiplied. \n",
+    "\n",
+    "#### Registers\n",
+    " - `x`: Input THRU register of size $m$ that stores elements from $GF(2^m)$.\n",
+    " - `y`: Input THRU register of size $m$ that stores elements from $GF(2^m)$.\n",
+    " - `result`: Output RIGHT register of size $m$ that stores the product $x * y$ in $GF(2^m)$.  \n",
+    "\n",
+    "#### References\n",
+    " - [On the Design and Optimization of a Quantum Polynomial-Time Attack on Elliptic Curve Cryptography](https://arxiv.org/abs/0710.1093). \n",
+    " - [Low complexity bit parallel architectures for polynomial basis multiplication over GF(2m)](https://ieeexplore.ieee.org/abstract/document/1306989). \n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "872a44d1",
+   "metadata": {
+    "cq.autogen": "GF2Multiplication.bloq_doc.py"
+   },
+   "outputs": [],
+   "source": [
+    "from qualtran.bloqs.gf_arithmetic import GF2Multiplication"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "d0f0db7d",
+   "metadata": {
+    "cq.autogen": "GF2Multiplication.example_instances.md"
+   },
+   "source": [
+    "### Example Instances"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "131bc962",
+   "metadata": {
+    "cq.autogen": "GF2Multiplication.gf16_multiplication"
+   },
+   "outputs": [],
+   "source": [
+    "gf16_multiplication = GF2Multiplication(4)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "69f564d8",
+   "metadata": {
+    "cq.autogen": "GF2Multiplication.gf2_multiplication_symbolic"
+   },
+   "outputs": [],
+   "source": [
+    "import sympy\n",
+    "\n",
+    "m = sympy.Symbol('m')\n",
+    "gf2_multiplication_symbolic = GF2Multiplication(m)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2a62c2b8",
+   "metadata": {
+    "cq.autogen": "GF2Multiplication.graphical_signature.md"
+   },
+   "source": [
+    "#### Graphical Signature"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cf003e98",
+   "metadata": {
+    "cq.autogen": "GF2Multiplication.graphical_signature.py"
+   },
+   "outputs": [],
+   "source": [
+    "from qualtran.drawing import show_bloqs\n",
+    "show_bloqs([gf16_multiplication, gf2_multiplication_symbolic],\n",
+    "           ['`gf16_multiplication`', '`gf2_multiplication_symbolic`'])"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f14ef0c5",
+   "metadata": {
+    "cq.autogen": "GF2Multiplication.call_graph.md"
+   },
+   "source": [
+    "### Call Graph"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f4b7bf2c",
+   "metadata": {
+    "cq.autogen": "GF2Multiplication.call_graph.py"
+   },
+   "outputs": [],
+   "source": [
+    "from qualtran.resource_counting.generalizers import ignore_split_join\n",
+    "gf16_multiplication_g, gf16_multiplication_sigma = gf16_multiplication.call_graph(max_depth=1, generalizer=ignore_split_join)\n",
+    "show_call_graph(gf16_multiplication_g)\n",
+    "show_counts_sigma(gf16_multiplication_sigma)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "name": "python"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}