### Run Slothy Examples

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/general.md

General command pattern to execute specific assembly optimization examples provided in the Slothy repository. Users can list all available examples by passing the help flag.

```bash
python3 example.py --examples={YOUR_EXAMPLE}
python3 example.py --help
```

--------------------------------

### Run SLOTHY Examples

Source: https://github.com/slothy-optimizer/slothy/blob/main/README.md

Executes example assembly snippets using the SLOTHY optimizer via its Python script. Replace YOUR_EXAMPLE with the desired example name.

```bash
python3 example.py --examples={YOUR_EXAMPLE}
```

--------------------------------

### Execute SLOTHY optimization example

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/index.md

Runs a specific assembly optimization example using the Python CLI. This command verifies the installation and invokes the constraint solver to minimize stalls in the provided assembly snippet.

```bash
python3 example.py --examples aarch64_simple0_a55
```

--------------------------------

### Install Development Requirements

Source: https://github.com/slothy-optimizer/slothy/blob/main/README.md

Installs the project's dependencies using pip from the requirements.txt file. It's recommended to use a virtual environment for this.

```bash
python3 -m venv venv
./venv/bin/python3 -m pip install -r requirements.txt
```

--------------------------------

### Install SLOTHY via Command Line

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/tutorial/README.md

This sequence of commands clones the SLOTHY repository, initializes a Python virtual environment, and installs the required dependencies. It ensures a clean, isolated environment for running the optimizer.

```bash
git clone https://github.com/slothy-optimizer/slothy
cd slothy
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
```

--------------------------------

### Example Assembly Code for SLOTHY

Source: https://github.com/slothy-optimizer/slothy/blob/main/README.md

A minimal example of assembly code that can be optimized by SLOTHY. It defines start and end labels for the optimization process.

```assembly
start_label:
    ldr x0, [x1]
    ldr x2, [x3]
    add x4, x0, x2
    str x4, [x5]
end_label:
```

--------------------------------

### Install SLOTHY using pip

Source: https://github.com/slothy-optimizer/slothy/blob/main/README.md

Installs the SLOTHY optimizer package from the Python Package Index (PyPI) using pip. This is the recommended method for most users.

```bash
pip install slothy
```

--------------------------------

### Complete NTT Optimization Example (Python)

Source: https://context7.com/slothy-optimizer/slothy/llms.txt

A comprehensive example demonstrating the optimization of a Number Theoretic Transform (NTT) implementation for cryptographic applications like Kyber. It involves loading the source, configuring various optimization parameters like software pipelining and register reservation, and then optimizing multiple layers of the NTT.

```python
import slothy
import slothy.targets.aarch64.aarch64_neon as AArch64_Neon
import slothy.targets.aarch64.cortex_a55 as Target_CortexA55

# Create optimizer instance
s = slothy.Slothy(AArch64_Neon, Target_CortexA55)
s.load_source_from_file("examples/naive/aarch64/kyber/ntt_kyber_123_4567.s")

# Configure for loop optimization
s.config.sw_pipelining.enabled = True
s.config.sw_pipelining.optimize_preamble = True
s.config.sw_pipelining.optimize_postamble = True
s.config.sw_pipelining.minimize_overlapping = False

s.config.inputs_are_outputs = True
s.config.variable_size = True
s.config.constraints.stalls_first_attempt = 64

# Reserve registers used by calling convention
s.config.reserved_regs = [f"x{i}" for i in range(0, 7)] + ["x30", "sp"]

# Optimize both NTT layers
s.optimize_loop("layer123_start")  # Layers 1, 2, 3
s.optimize_loop("layer4567_start")  # Layers 4, 5, 6, 7
```

--------------------------------

### Configure Input and Output Renaming

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.core.config.md

Examples of setting up rename_inputs and rename_outputs dictionaries to control how symbolic and architectural registers are mapped during optimization.

```python
config.rename_inputs = {"in": "r0", "arch": "static", "symbolic": "any"}
config.rename_outputs = {"other": "any"}
```

--------------------------------

### GET /slothy/targets/riscv/rv32_64_b_instructions/generate

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.riscv.rv32_64_b_instructions.md

Generates all instruction classes for the RV32/64 pseudo instructions.

```APIDOC
## GET /slothy/targets/riscv/rv32_64_b_instructions/generate

### Description
Generates all instruction classes for the RV32/64 pseudo instructions, enabling support for the bit manipulation extension set.

### Method
GET

### Endpoint
/slothy/targets/riscv/rv32_64_b_instructions/generate

### Parameters
None

### Request Example
{}

### Response
#### Success Response (200)
- **instructions** (list) - A list of generated instruction classes for the RV32/64-B extension.

#### Response Example
{
  "status": "success",
  "data": ["rol", "ror", "andn", "orn", "xnor", "pack", "packh", "rori", "brev8"]
}
```

--------------------------------

### Verify Development Installation

Source: https://github.com/slothy-optimizer/slothy/blob/main/README.md

Runs a test script to verify the development installation of SLOTHY. This command executes a specific test case for AArch64 architecture.

```bash
python3 test.py --tests aarch64_simple0_a55
```

--------------------------------

### Assembly Rescheduling Example

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/tutorial/README.md

This snippet shows an example of assembly code before and after optimization by Slothy. The comments indicate the original source and a visual representation of instruction movement. Slothy aims to reduce stalls by rescheduling instructions, potentially renaming registers to enable better interleaving.

```assembly
// sqrdmulh v11.8h, v11.8h, v0.h[1]       // .......*............
// mls v24.8h, v11.8h, v1.h[0]            // ...........*........
// sub     v11.8h,    v10.8h, v24.8h      // .................*..
// add     v10.8h,    v10.8h, v24.8h      // ...............*....
// str q8,  [x0], #4*16                   // ..............*.....
// str q9,  [x0, #-3*16]                  // ................*...
// str q10, [x0, #-2*16]                  // ..................*.
// str q11, [x0, #-1*16]                  // ...................*
```

--------------------------------

### Verify SLOTHY Installation

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/tutorial/README.md

Executes the test suite to verify that the SLOTHY installation is functional. It runs a specific test case to ensure all dependencies and core logic are correctly configured.

```bash
python3 test.py --tests simple0_a55
```

--------------------------------

### Configure SLOTHY Optimization Parameters

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.core.config.md

Example of how to structure the configuration object for the SLOTHY optimizer. This includes settings for LLVM MCA integration, solver timeouts, and heuristic splitting.

```json
{
  "sw_pipelining": {
    "enabled": true
  },
  "with_llvm_mca": true,
  "compiler_binary": "/usr/bin/clang",
  "timeout": 30,
  "split_heuristic": true,
  "split_heuristic_factor": 4,
  "do_address_fixup": true
}
```

--------------------------------

### Stepwise Assembly Performance Profiling

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/tutorial/README.md

An example of a performance profiling output generated by the pqax framework. It maps individual assembly instructions to their execution timing, allowing for visual analysis of instruction latency and dependency chains.

```asm
ldr q0, [x1, #0]
ldr q8, [x0, #0*16]
ldr q9, [x0, #1*16]
ldr q10, [x0, #2*16]
ldr q11, [x0, #3*16]
mul v12.8h, v9.8h, v0.h[0]
sqrdmulh v9.8h, v9.8h, v0.h[1]
mls v12.8h, v9.8h, v1.h[0]
sub v9.8h, v8.8h, v12.8h
add v8.8h, v8.8h, v12.8h
mul v12.8h, v11.8h, v0.h[0]
sqrdmulh v11.8h, v11.8h, v0.h[1]
mls v12.8h, v11.8h, v1.h[0]
sub v11.8h, v10.8h, v12.8h
add v10.8h, v10.8h, v12.8h
str q8, [x0], #4*16
str q9, [x0, #-3*16]
```

--------------------------------

### Define ARM v8.1-M MVE Instruction Classes

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.arm_v81m.arch_v81m.md

Example implementation of instruction classes for ARM v8.1-M, defining patterns and register dependencies for MVE operations.

```python
class veor(MVEInstruction):
    pattern = 'veor.<dt> <Qd>, <Qn>, <Qm>'
    inputs = ['Qn', 'Qm']
    outputs = ['Qd']

class vstrw(MVEInstruction):
    pattern = 'vstrw.<dt> <Qd>, [<Rn>, <imm>]'
    inputs = ['Qd', 'Rn']
    @classmethod
    def make(cls, src):
        pass
    def write(self):
        pass

class vstrw_with_writeback(MVEInstruction):
    pattern = 'vstrw.<dt> <Qd>, [<Rn>, <imm>]!'
    inputs = ['Qd']
    in_outs = ['Rn']
```

--------------------------------

### Define ARMv7-M Instruction Classes

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.arm_v7m.arch_v7m.md

Example of defining an instruction class for ARMv7-M in Slothy, specifying the pattern, inputs, and outputs.

```python
class pkhbt(Armv7mLogical):
    pattern = 'pkhbt<width> <Rd>, <Ra>, <Rb>'
    inputs = ['Ra', 'Rb']
    outputs = ['Rd']

class ldr_with_imm(Armv7mLoadInstruction):
    pattern = 'ldr<width> <Rd>, [<Ra>, <imm>]'
    inputs = ['Ra']
    outputs = ['Rd']
    @classmethod
    def make(cls, src):
        pass
    def write(self):
        pass
```

--------------------------------

### Initialize SLOTHY Optimizer

Source: https://context7.com/slothy-optimizer/slothy/llms.txt

Demonstrates how to instantiate the Slothy class with specific architecture and microarchitecture models, configure parameters, and perform optimization on a source file.

```python
import slothy
import slothy.targets.aarch64.aarch64_neon as AArch64_Neon
import slothy.targets.aarch64.cortex_a55 as Target_CortexA55

s = slothy.Slothy(AArch64_Neon, Target_CortexA55)
s.load_source_from_file("input.s")
s.config.variable_size = True
s.config.constraints.stalls_first_attempt = 32
s.optimize(start="start_label", end="end_label")
s.write_source_to_file("output_optimized.s")
```

--------------------------------

### Clone SLOTHY Repository for Development

Source: https://github.com/slothy-optimizer/slothy/blob/main/README.md

Clones the SLOTHY optimizer repository from GitHub. This is necessary for development installations or running provided examples.

```bash
git clone https://github.com/slothy-optimizer/slothy.git
```

--------------------------------

### Complete Slothy Docker Setup Sequence

Source: https://github.com/slothy-optimizer/slothy/blob/main/paper/artifact/README.md

A consolidated sequence of bash commands to build the Slothy Docker image, verify its creation, run a container, check the running container, and access its shell. This provides a complete workflow for setting up the Slothy environment via Docker.

```bash
docker build -f slothy.Dockerfile -t slothy_image .
docker image ls
docker run --name slothy_container -d -it slothy_image /bin/bash
docker container ls
docker exec -it slothy_container /bin/bash
```

--------------------------------

### RISC-V BranchLoop Class: Start and End Loop

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.riscv.riscv.md

Implements loop starting and ending logic for RISC-V architecture using the BranchLoop class. The start method emits initial instructions and jump labels, while the end method handles the compare-and-branch instruction, which is part of the loop body in this experimental loop type.

```python
class BranchLoop(Loop):
    """More general loop type that just considers the branch instruction as part
    of the boundary.
    This can help to improve performance as the instructions that belong to
    handling the loop can be considered by SLOTHY as well.

    NOTE
    This loop type is still rather experimental. It has a lot of logic
    inside as it needs to be able to “understand” a variety of different
    ways to express loops, e.g., how counters get incremented, how
    registers marking the end of the loop need to be modified in case of
    software pipelining etc.

    Example:

    ```asm
    loop_lbl:
       {code}
       bltu <cnt>, <end>, loop_lbl
    ```

    where cnt is the loop counter that gets incremented within the loop body.
    """
    def start(loop_cnt, indentation=0, fixup=0, unroll=1, jump_if_empty=None, preamble_code=None, body_code=None, postamble_code=None, register_aliases=None):
        """Emit starting instruction(s) and jump label for loop"""
        pass

    def end(other, indentation=0):
        """Nothing to do here - the branch instruction is already part of the body"""
        pass
```

--------------------------------

### Neoverse N1 Performance Metrics Functions

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.aarch64.neoverse_n1_experimental.md

Provides functions to retrieve performance-related metrics for the Neoverse N1 target. This includes functions to add constraints, check for min/max objectives, get latency, get execution units, and get inverse throughput. These functions are crucial for the Slothy optimizer to model CPU behavior accurately.

```python
def add_further_constraints(slothy):
    pass

def has_min_max_objective(config):
    pass

def get_min_max_objective(slothy):
    pass

def get_latency(src, out_idx, dst):
    pass

def get_units(src):
    pass

def get_inverse_throughput(src):
    pass
```

--------------------------------

### POST /instruction/build

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.aarch64.aarch64_neon.md

Attempts to parse a raw assembly string into a specific instruction instance using the instruction factory.

```APIDOC
## POST /instruction/build

### Description
Attempts to parse a string as an instance of a target instruction class. This is the primary entry point for converting assembly lines into structured instruction objects.

### Method
POST

### Endpoint
/instruction/build

### Parameters
#### Request Body
- **c** (any) - Required - The target instruction class to parse as.
- **src** (string) - Required - The raw assembly string to parse.
- **mnemonic** (string) - Required - The mnemonic associated with the instruction.

### Request Example
{
  "c": "AArch64Instruction",
  "src": "add v0.4s, v1.4s, v2.4s",
  "mnemonic": "add"
}

### Response
#### Success Response (200)
- **instruction** (object) - The parsed instruction instance.

#### Response Example
{
  "status": "success",
  "instruction": { "type": "AArch64Instruction", "mnemonic": "add" }
}
```

--------------------------------

### Use slothy-cli for Assembly Optimization

Source: https://context7.com/slothy-optimizer/slothy/llms.txt

This section provides examples of using the `slothy-cli` command-line interface for optimizing assembly files. It covers basic optimization, specifying code sections, setting configuration options via flags, enabling LLVM MCA, and activating debug mode.

```bash
# Basic optimization
slothy-cli Arm_AArch64 Arm_Cortex_A55 input.s -o output.s

# Optimize a specific loop
slothy-cli Arm_AArch64 Arm_Cortex_A55 input.s -o output.s \
    --loop loop_start

# Optimize code between start/end labels
slothy-cli Arm_AArch64 Arm_Cortex_A55 input.s -o output.s \
    --start start_label --end end_label

# Set configuration options
slothy-cli Arm_AArch64 Arm_Cortex_A55 input.s -o output.s \
    -c sw_pipelining.enabled=True \
    -c variable_size=True \
    -c constraints.stalls_first_attempt=32 \
    -c inputs_are_outputs=True \
    -c "reserved_regs=[x0,x1,x2,sp,x30]"

# Enable LLVM MCA analysis
slothy-cli Arm_AArch64 Arm_Cortex_A55 input.s -o output.s \
    -c with_llvm_mca=True

# Debug mode with verbose output
slothy-cli Arm_AArch64 Arm_Cortex_A55 input.s -o output.s --debug

# List all configuration options
slothy-cli Arm_AArch64 Arm_Cortex_A55 input.s -c
```

--------------------------------

### Slothy Initialization and Logging

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.helper.md

Details on initializing the Slothy optimizer and its logging capabilities.

```APIDOC
## Initialization

Initializes the instance - basically setting the formatter to None and the filter list to empty.

### emit(record)

### forward(logger)

Send all captured records to the given logger.

### forward_to_file(log_label, filename, lvl=logging.DEBUG)

Store all captured records in a file.
```

--------------------------------

### ADD Instructions

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.arm_v7m.arch_v7m.md

Instructions for performing addition operations.

```APIDOC
## add

### Description
Adds two registers and stores the result in a destination register.

### Method
N/A (Class definition)

### Endpoint
N/A

### Parameters
#### Path Parameters
N/A

#### Query Parameters
N/A

#### Request Body
N/A

### Request Example
N/A

### Response
#### Success Response (200)
N/A

#### Response Example
N/A

#### Pattern
`add<width> <Rd>, <Ra>, <Rb>`

#### Inputs
- `Ra` (register)
- `Rb` (register)

#### Outputs
- `Rd` (register)
```

```APIDOC
## add_short

### Description
Adds a register to itself and stores the result in the same register.

### Method
N/A (Class definition)

### Endpoint
N/A

### Parameters
#### Path Parameters
N/A

#### Query Parameters
N/A

#### Request Body
N/A

### Request Example
N/A

### Response
#### Success Response (200)
N/A

#### Response Example
N/A

#### Pattern
`add<width> <Rd>, <Ra>`

#### Inputs
- `Ra` (register)

#### In-Outs
- `Rd` (register)
```

```APIDOC
## add_imm

### Description
Adds an immediate value to a register and stores the result in a destination register.

### Method
N/A (Class definition)

### Endpoint
N/A

### Parameters
#### Path Parameters
N/A

#### Query Parameters
N/A

#### Request Body
N/A

### Request Example
N/A

### Response
#### Success Response (200)
N/A

#### Response Example
N/A

#### Pattern
`add<width> <Rd>, <Ra>, <imm>`

#### Inputs
- `Ra` (register)

#### Outputs
- `Rd` (register)
```

```APIDOC
## add_imm_short

### Description
Adds an immediate value to a register, storing the result in the same register.

### Method
N/A (Class definition)

### Endpoint
N/A

### Parameters
#### Path Parameters
N/A

#### Query Parameters
N/A

#### Request Body
N/A

### Request Example
N/A

### Response
#### Success Response (200)
N/A

#### Response Example
N/A

#### Pattern
`add<width> <Rd>, <imm>`

#### In-Outs
- `Rd` (register)
```

```APIDOC
## add_shifted

### Description
Adds a register to a shifted register and stores the result in a destination register.

### Method
N/A (Class definition)

### Endpoint
N/A

### Parameters
#### Path Parameters
N/A

#### Query Parameters
N/A

#### Request Body
N/A

### Request Example
N/A

### Response
#### Success Response (200)
N/A

#### Response Example
N/A

#### Pattern
`add<width> <Rd>, <Ra>, <Rb>, <barrel><imm>`

#### Inputs
- `Ra` (register)
- `Rb` (register)

#### Outputs
- `Rd` (register)
```

--------------------------------

### Get SSA Form Helper (Python)

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.core.heuristics.md

A private helper function to get the Static Single Assignment (SSA) form of the code body. This is an internal utility function used within the Slothy optimizer for specific optimization passes.

```python
def _get_ssa_form(body, logger, conf):
    pass
```

--------------------------------

### GET /slothy/targets/arm_v7m/arch_v7m/RegisterType

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.arm_v7m.arch_v7m.md

Accesses the RegisterType enum definitions for ARMv7-M.

```APIDOC
## GET /slothy/targets/arm_v7m/arch_v7m/RegisterType

### Description
Provides access to the RegisterType enumeration, defining categories such as GPR, FPR, FLAGS, and HINT.

### Method
GET

### Endpoint
/slothy/targets/arm_v7m/arch_v7m/RegisterType

### Response
#### Success Response (200)
- **GPR** (integer) - 1
- **FPR** (integer) - 2
- **FLAGS** (integer) - 3
- **HINT** (integer) - 4

#### Response Example
{
  "GPR": 1,
  "FPR": 2,
  "FLAGS": 3,
  "HINT": 4
}
```

--------------------------------

### ARMv8.1-M Store Instructions

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.arm_v81m.arch_v81m.md

Provides documentation for various store instructions available in the ARMv8.1-M architecture.

```APIDOC
## Store Instructions (ARMv8.1-M)

### Description
This section details various store instructions for the ARMv8.1-M architecture, including double quadword and quadword stores, with options for writeback.

### Store Double Quadword (VST2)

#### VST20

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VST21

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VST20 with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VST21 with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

### Store Quadword (VST4)

#### VST40

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VST41

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VST42

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VST43

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VST40 with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VST41 with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VST42 with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VST43 with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A
```

--------------------------------

### Optimize Loop API

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.core.slothy.md

Optimizes a loop starting at a given label.

```APIDOC
## optimize_loop

### Description
Optimize the loop starting at a given label.

### Method
POST

### Endpoint
/slothy-optimizer/slothy/optimize_loop

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
- **loop_lbl** (str) - Required - Label of loop to optimize.
- **postamble_label** (str) - Optional - Marks end of loop kernel.
- **forced_loop_type** (any) - Optional - Forces the loop to be parsed as a certain type.

### Request Example
```json
{
  "loop_lbl": "main_loop",
  "postamble_label": "loop_end",
  "forced_loop_type": "type_B"
}
```

### Response
#### Success Response (200)
- **status** (string) - Indicates success of the optimization.

#### Response Example
```json
{
  "status": "Loop optimized successfully"
}
```
```

--------------------------------

### MOV Instructions (Immediate)

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.arm_v7m.arch_v7m.md

Instructions for moving immediate values into registers.

```APIDOC
## movw_imm

### Description
Moves a 16-bit immediate value into a destination register.

### Method
N/A (Class definition)

### Endpoint
N/A

### Parameters
#### Path Parameters
N/A

#### Query Parameters
N/A

#### Request Body
N/A

### Request Example
N/A

### Response
#### Success Response (200)
N/A

#### Response Example
N/A

#### Pattern
`movw <Rd>, <imm>`

#### Outputs
- `Rd` (register)
```

```APIDOC
## movt_imm

### Description
Moves a 16-bit immediate value into the upper half of a destination register.

### Method
N/A (Class definition)

### Endpoint
N/A

### Parameters
#### Path Parameters
N/A

#### Query Parameters
N/A

#### Request Body
N/A

### Request Example
N/A

### Response
#### Success Response (200)
N/A

#### Response Example
N/A

#### Pattern
`movt <Rd>, <imm>`

#### In-Outs
- `Rd` (register)
```

--------------------------------

### GET /graph/dependencies

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.core.dataflow.md

Returns an iterator over all dependencies in the data flow graph.

```APIDOC
## GET /graph/dependencies

### Description
Returns an iterator over all dependencies in the data flow graph. Each element represents a dependency between a consumer and a producer.

### Method
GET

### Endpoint
/graph/dependencies

### Response
#### Success Response (200)
- **dependencies** (array) - List of tuples (consumer, producer, ty, idx).

#### Response Example
{
  "dependencies": [["node_2", "node_1", "in", 0]]
}
```

--------------------------------

### Load Assembly Source

Source: https://context7.com/slothy-optimizer/slothy/llms.txt

Shows methods for loading assembly code into the optimizer, either from an external file or directly as a raw string.

```python
import slothy
import slothy.targets.aarch64.aarch64_neon as AArch64_Neon
import slothy.targets.aarch64.cortex_a55 as Target_CortexA55

s = slothy.Slothy(AArch64_Neon, Target_CortexA55)
s.load_source_from_file("examples/naive/aarch64/kyber/ntt_kyber_123_4567.s")

assembly_code = """
start:
    ldr q0, [x1, #0]
    ldr q1, [x2, #0]
    mul v2.8h, v0.8h, v1.h[0]
    str q2, [x3]
end:
"""
s.load_source_raw(assembly_code)
```

--------------------------------

### GET /slothy/targets/aarch64/cortex_a72_frontend/units

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.aarch64.cortex_a72_frontend.md

Retrieves the execution units associated with a specific instruction.

```APIDOC
## GET /slothy/targets/aarch64/cortex_a72_frontend/get_units

### Description
Returns the execution units required by a specific instruction source.

### Method
GET

### Endpoint
/slothy/targets/aarch64/cortex_a72_frontend/get_units

### Parameters
#### Query Parameters
- **src** (string) - Required - The instruction source identifier

### Request Example
GET /slothy/targets/aarch64/cortex_a72_frontend/get_units?src=LDR

### Response
#### Success Response (200)
- **units** (array) - List of execution units (e.g., LOAD0, LOAD1)

#### Response Example
{
  "units": ["LOAD0", "LOAD1"]
}
```

--------------------------------

### Get Loop Input/Output API

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.core.slothy.md

Finds all registers that a loop body depends on.

```APIDOC
## get_loop_input_output

### Description
Find all registers that a loop body depends on.

### Method
GET

### Endpoint
/slothy-optimizer/slothy/get_loop_input_output

### Parameters
#### Path Parameters
None

#### Query Parameters
- **loop_lbl** (str) - Required - Label of loop to process.
- **forced_loop_type** (any) - Optional - Forces the loop to be parsed as a certain type.

### Request Example
```
GET /slothy-optimizer/slothy/get_loop_input_output?loop_lbl=loop1&forced_loop_type=type_C
```

### Response
#### Success Response (200)
- **inputs** (list) - A list of all input registers of the loop.

#### Response Example
```json
{
  "inputs": ["reg1", "reg2", "reg3"]
}
```
```

--------------------------------

### ARM v8.1-M Vector Store Instructions

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.arm_v81m.arch_v81m.md

Definitions for various VST (Vector Store) instructions, including variants with and without writeback support.

```APIDOC
## POST /slothy/targets/arm_v81m/vst21

### Description
Defines the vst21 instruction for storing vector registers to memory.

### Method
POST

### Endpoint
/slothy/targets/arm_v81m/vst21

### Parameters
#### Request Body
- **pattern** (string) - Required - 'vst21.<dt> {<Qd0>, <Qd1>}, [<Rn>]'
- **inputs** (list) - Optional - ['Rn', 'Qd0', 'Qd1']

### Response
#### Success Response (200)
- **status** (string) - Instruction object created successfully.
```

```APIDOC
## POST /slothy/targets/arm_v81m/vst43_with_writeback

### Description
Defines the vst43 instruction with memory writeback support.

### Method
POST

### Endpoint
/slothy/targets/arm_v81m/vst43_with_writeback

### Parameters
#### Request Body
- **pattern** (string) - Required - 'vst43.<dt> {<Qd0>, <Qd1>, <Qd2>, <Qd3>}, [<Rn>]!'
- **inputs** (list) - Optional - ['Qd0', 'Qd1', 'Qd2', 'Qd3']
- **in_outs** (list) - Optional - ['Rn']

### Response
#### Success Response (200)
- **status** (string) - Instruction object created successfully.
```

--------------------------------

### ARMv7m Load/Store Instructions

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.arm_v7m.arch_v7m.md

Details on load and store operations for ARMv7m architecture.

```APIDOC
## Armv7mLoadInstruction

### Description
Represents a load instruction for ARMv7m.

### Method
N/A (Class definition)

### Endpoint
N/A

### Parameters
N/A

### Request Example
N/A

### Response
N/A
```

```APIDOC
## Armv7mStoreInstruction

### Description
Represents a store instruction for ARMv7m.

### Method
N/A (Class definition)

### Endpoint
N/A

### Parameters
N/A

### Request Example
N/A

### Response
N/A
```

--------------------------------

### GET /slothy/targets/arm_v81m/helium_experimental/get_units

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.arm_v81m.helium_experimental.md

Retrieves the execution units associated with a specific instruction source.

```APIDOC
## GET /slothy/targets/arm_v81m/helium_experimental/get_units

### Description
Returns the execution unit mapping for a given instruction source based on the Helium experimental target definition.

### Method
GET

### Endpoint
/slothy/targets/arm_v81m/helium_experimental/get_units

### Parameters
#### Query Parameters
- **src** (object) - Required - The instruction source object.

### Response
#### Success Response (200)
- **units** (list) - A list of applicable ExecutionUnit enums.

#### Response Example
{
  "units": ["SCALAR"]
}
```

--------------------------------

### Basic Slothy Optimization

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/tutorial/README.md

This snippet demonstrates the basic usage of the Slothy optimizer to load an assembly file, configure optimization settings, perform optimization, and write the optimized code to a new file.

```APIDOC
## Slothy Basic Optimization

### Description
This endpoint demonstrates the basic usage of the Slothy optimizer. It loads an assembly file, configures optimization settings, performs the optimization, and writes the optimized code to a specified output file.

### Method
N/A (Python API)

### Endpoint
N/A (Python API)

### Parameters
N/A

### Request Example
```python
import slothy

# Load the assembly source file
slothy.load_source_from_file("tests/naive/aarch64/aarch64_simple0.s")

# Configure Slothy settings
slothy.config.variable_size=True
slothy.config.constraints.stalls_first_attempt=32

# Perform the optimization
slothy.optimize()

# Write the optimized source code to a file
slothy.write_source_to_file("opt/aarch64_simple0_a55.s")
```

### Response
N/A (Python API)

### Response Example
N/A (Python API)
```

--------------------------------

### GET /slothy/targets/aarch64/cortex_a72_frontend/latency

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.aarch64.cortex_a72_frontend.md

Retrieves the latency for a specific instruction within the Cortex-A72 model.

```APIDOC
## GET /slothy/targets/aarch64/cortex_a72_frontend/get_latency

### Description
Returns the latency value for a given instruction source, output index, and destination.

### Method
GET

### Endpoint
/slothy/targets/aarch64/cortex_a72_frontend/get_latency

### Parameters
#### Query Parameters
- **src** (string) - Required - The instruction source identifier
- **out_idx** (integer) - Required - The output index of the instruction
- **dst** (string) - Required - The destination register or operand

### Request Example
GET /slothy/targets/aarch64/cortex_a72_frontend/get_latency?src=ADD&out_idx=0&dst=X0

### Response
#### Success Response (200)
- **latency** (integer) - The calculated latency for the specified instruction parameters

#### Response Example
{
  "latency": 1
}
```

--------------------------------

### ASimdCompare Class

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.aarch64.aarch64_neon.md

Documentation for the ASimdCompare class used for SIMD comparison instructions.

```APIDOC
## Class: ASimdCompare

### Description
Parent class for all ASIMD (Advanced SIMD) compare instructions in the AArch64 target.

### Methods
- cmge: Implements the compare greater than or equal instruction.

### Inheritance
- Inherits from base instruction classes within the `slothy.targets.aarch64.aarch64_neon` namespace.
```

--------------------------------

### GET /targets/aarch64/neoverse_n1_experimental/units

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.aarch64.neoverse_n1_experimental.md

Retrieves the execution units associated with a specific instruction source.

```APIDOC
## GET /targets/aarch64/neoverse_n1_experimental/units

### Description
Identifies which execution units are used by a specific instruction source.

### Method
GET

### Endpoint
/targets/aarch64/neoverse_n1_experimental/get_units

### Parameters
#### Query Parameters
- **src** (string) - Required - The source instruction identifier.

### Request Example
GET /targets/aarch64/neoverse_n1_experimental/get_units?src=FADD

### Response
#### Success Response (200)
- **units** (array) - List of execution units (e.g., VEC0, VEC1).

#### Response Example
{
  "units": ["VEC0", "VEC1"]
}
```

--------------------------------

### ARMv8.1-M Load Instructions

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.arm_v81m.arch_v81m.md

Provides documentation for various load instructions available in the ARMv8.1-M architecture.

```APIDOC
## Load Instructions (ARMv8.1-M)

### Description
This section details various load instructions for the ARMv8.1-M architecture, including byte, halfword, word, and doubleword loads, with options for writeback, post-indexing, and gather operations.

### Load Byte (LDRB)

#### LDRB with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### LDRB with Post-indexing

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

### Load Halfword (LDRH)

#### LDRH

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### LDRH without Immediate Offset

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### LDRH with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### LDRH with Post-indexing

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

### Load Word (LDRW)

#### LDRW

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### LDRW without Immediate Offset

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### LDRW with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### LDRW with Post-indexing

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### LDRW Gather

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### LDRW Gather with UXTW

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

### Load Byte Gather (LDRB_GATHER)

#### LDRB Gather

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### LDRB Gather with UXTW

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

### Load Halfword Gather (LDRH_GATHER)

#### LDRH Gather

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### LDRH Gather with UXTW

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

### Load Double Quadword (VLD2)

#### VLD20

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VLD21

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VLD20 with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VLD21 with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

### Load Quadword (VLD4)

#### VLD40

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VLD41

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VLD42

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VLD43

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VLD40 with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VLD41 with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VLD42 with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A

#### VLD43 with Writeback

### Method
Various (Specific instruction dependent)

### Endpoint
N/A (Instruction level)

### Parameters
N/A (Instruction level)

### Request Example
N/A

### Response
N/A
```

--------------------------------

### GET /slothy/targets/arm_v81m/helium_experimental/get_latency

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.arm_v81m.helium_experimental.md

Retrieves the latency for a specific instruction source, output index, and destination.

```APIDOC
## GET /slothy/targets/arm_v81m/helium_experimental/get_latency

### Description
Calculates the latency associated with a given instruction source, output index, and destination within the Helium architecture.

### Method
GET

### Endpoint
/slothy/targets/arm_v81m/helium_experimental/get_latency

### Parameters
#### Query Parameters
- **src** (object) - Required - The instruction source object.
- **out_idx** (integer) - Required - The output index of the instruction.
- **dst** (object) - Required - The destination operand.

### Response
#### Success Response (200)
- **latency** (integer) - The calculated latency value.

#### Response Example
{
  "latency": 1
}
```

--------------------------------

### Model Instruction Characteristics

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/tutorial/README.md

Shows how to map specific instructions to their execution units, inverse throughput, and latency using dictionary structures.

```python
execution_units = {
    ( vmull, vadd ): [[ExecutionUnit.VEC0, ExecutionUnit.VEC1]],
}

inverse_throughput = {
    ( vmull, vadd ) : 1,
}

default_latencies = {
    ( vmull ) : 4,
    ( vadd ) : 3,
}
```

--------------------------------

### GET /slothy/targets/arm_v7m/arch_v7m/unicorn_reg_by_name

Source: https://github.com/slothy-optimizer/slothy/blob/main/docs/source/apidocs/slothy/slothy.targets.arm_v7m.arch_v7m.md

Retrieves the numerical identifier for a given register name for use with the Unicorn engine.

```APIDOC
## GET /slothy/targets/arm_v7m/arch_v7m/unicorn_reg_by_name

### Description
Converts the string name of a register into the numerical identifier required by the Unicorn engine.

### Method
GET

### Endpoint
/slothy/targets/arm_v7m/arch_v7m/unicorn_reg_by_name

### Parameters
#### Query Parameters
- **reg** (string) - Required - The name of the register to look up.

### Request Example
{
  "reg": "r0"
}

### Response
#### Success Response (200)
- **id** (integer) - The numerical identifier for the register.

#### Response Example
{
  "id": 1
}
```