Skip to main content

Unified Compute Platform - CPU, GPU, FPGA, Quantum Computing

Project description

Digital Soul

Unified Compute Platform - CPU, GPU, FPGA, Quantum Computing

DigitalSoul is a Python module designed to bridge the gap between classical, quantum, and potentially hardware-accelerated computation. It provides flexible data structures and a node-based execution model, allowing you to express computations that can be seamlessly executed across CPU, GPU, quantum simulators, and potentially FPGAs.

Key Features

  • Customizable Data Types: Define Boolean (Bool), integer (Int, UInt), floating-point (Float), quantum states (Qudit), quantum gates (QuantumGate), and multidimensional tensors (Tensor) to suit your computational needs.
  • Node-Based Computation: Build computational graphs using nodes that represent operations (e.g., LogicalAnd, LogicalOr). Nodes manage input/output data through "Edges".
  • Multi-Backend Execution: Execute computations using NumPy, Cupy (for GPU), TensorFlow, and internal quantum simulator
  • VHDL Transpilation: Translate computational graphs into VHDL code, opening the door for hardware synthesis on FPGAs.

Installation

From PyPI

pip install setuptools wheel pybind11
pip install DigitalSoul

Also download numpy, cupy and tensorflow if you want to access richer executors

Quick Example

import DigitalSoul as ds

e1=ds.Edge(ds.Bool(False))
e2=ds.Edge(ds.Bool(False))
e3=ds.Edge(ds.Bool(None))
e4=ds.Edge(ds.Bool(True))
e5=ds.Edge(ds.Bool(None))
e6=ds.Edge(ds.Bool(True))
e7=ds.Edge(ds.Bool(None))
e8=ds.Edge(ds.Bool(None))

print("\n"*4)
or_gate=ds.LogicalOr((e1,e2), e3)
xor_gate=ds.LogicalXor((e3,e4), e8)
not_gate=ds.LogicalNot(e8,e5)
and_gate1=ds.LogicalAnd((e5,e6), e7)
print(e7)
print("Executing function")
and_gate1.execute("cp")
print(e7)
and_gate1.q_stream()
print("\n",e7.sv)
print("\n"*3)
print(and_gate1.transpile("vhdl"))

output:

Edge_6 holding Bool_6 value=None entropy=1
Executing function
Edge_6 holding Bool_6 value=False entropy=0

 2-levelQudit_3 value=[1. 0.] entropy=0




library IEEE;
use IEEE.STD_LOGIC_1164.ALL;
entity main is
Port(
	Bool_1:in std_logic;
	Bool_0:in std_logic;
	Bool_3:in std_logic;
	Bool_5:in std_logic;
	Bool_6:out std_logic
);
end main;
architecture Behavioral of main is
	signal Bool_2:std_logic;
	signal Bool_4:std_logic;
	signal Bool_7:std_logic;
begin
	Bool_7<=Bool_2 xor Bool_3;
	Bool_2<=Bool_0 or Bool_1;
	Bool_4<=not(Bool_7);
	Bool_6<=Bool_4 and Bool_5;
end architecture Behavioral;

As you can see from output, the value of output edge(e7, shown as Edge_6) is uncertain before computation. Hence, it has maximum entropy. As soon as value is computed and certainly known, the entropy is zero. Then, program is capable of generating VHDL code of corresponding computational graph. Additionally, it simulated quantum equivalent of the computaitonal graph with Non-Hermetian Gates

Tree

|----Bool-------------------|
|                           |---__init__(value=None)
|                           |---entropy()
|                           |---name()
|                           |---__repr__()
|
|
|----Int--------------------|
|                           |---__init__(value=None,depth=32)
|                           |---bounds()
|                           |---entropy()
|                           |---name()
|                           |---__repr__()
|
|
|----UInt-------------------|
|                           |---__init__(value=None,depth=32)
|                           |---entropy()
|                           |---name()
|                           |---__repr__()
|
|
|----Float------------------|
|                           |---__init__(value=None,exponent=8,mantissa=23)
|                           |---float_info()
|                           |---entropy()
|                           |---name()
|                           |---__repr__()
|
|
|----Qudit------------------|
|                           |---__init__(value,num_levels=None,utol=1e-9)
|                           |---num_levels()
|                           |---entropy()
|                           |---name()
|                           |---__repr__()
|                           |---__and__(other)
|
|
|----QuantumGate(object)----|
|                           |---__init__(data,utol=1e-8)
|                           |---__repr__()
|                           |---data()
|                           |---value()
|                           |---set_data(data,utol)
|                           |---num_levels()
|                           |---entropy()
|                           |---name()
|                           |---__and__(other)
|                           |---__call__(sv)
|
|
|----NonHermitianGate-------|
|                           |---__init__(data)
|                           |---value()
|                           |---name()
|                           |---__call__(sv)
|                           |---__repr__()
|
|
|----Tensor(object)---------|
|                           |---__init__(value,dtype=Float(0),shape=(1,))
|                           |---entropy()
|                           |---__repr__()
|                           |---name()
|
|
|----Edge(object)-----------|
|                           |---__init__(sculk)
|                           |---vhdl()
|                           |---twoscpl(x
|                           |---name()
|                           |---__repr__()
|                           |---entropy()
|                           |---set_predecessor(node)
|                           |---unpack(executor="np")
|                           |---q_info()
|
|
|----Node(object)-----------|
|                           |---__init__(in_terminals,out_terminals,ops)
|                           |---execute(executor="np")
|                           |---edge_accumulator()
|                           |---input_accumulator()
|                           |---node_accumulator()
|                           |---transpile(target="vhdl")
|                           |---qv_contemplate()
|                           |---q_stream()
|
|
|----LogicalAnd(Node)-------|
|                           |---__init__(in_terminals,out_terminals,ops)
|
|----LogicalOr(Node)--------|
|                           |---__init__(in_terminals,out_terminals,ops)
|
|----LogicalXor(Node)--------|
|                            |---__init__(in_terminals,out_terminals,ops)
|
|----LogicalNot(Node)--------|
|                            |---__init__(in_terminals,out_terminals,ops)
|
|----QN---------------------|
                            |---i
                            |---x
                            |---y
                            |---z
                            |---h
                            |---cx
                            |---ccx

  

Roadmap

  • Implementing more nodes
  • Improved hardware synthesis flow with VHDL transpilation.
  • Custom node creation guide.

Contributing

We welcome contributions to DigitalSoul!

License

DigitalSoul is distributed under the MIT License (see LICENSE.md).

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

DigitalSoul-1.1.5.tar.gz (13.6 kB view details)

Uploaded Source

File details

Details for the file DigitalSoul-1.1.5.tar.gz.

File metadata

  • Download URL: DigitalSoul-1.1.5.tar.gz
  • Upload date:
  • Size: 13.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.0.0 CPython/3.11.5

File hashes

Hashes for DigitalSoul-1.1.5.tar.gz
Algorithm Hash digest
SHA256 538d3899a3007b6d03295358629e69aa2b34c420602f9a5269a89d8df8b895ff
MD5 777ab3c26987bc8d627f7ed9e5584e71
BLAKE2b-256 078d223d6e06106e98dc027b8ec70205e2ecc58b54cf21831f5a0b5194f8c258

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page