# tensorplay.distributed.elastic.multiprocessing API

Source: https://www.tensorplay.cn/docs/api/tensorplay.distributed.elastic.multiprocessing.html

## Functions 2

[#](#api-tensorplay.distributed.elastic.multiprocessing.start_processes)

### start_processes

function[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.multiprocessing.start_processes.html)

```python
tensorplay.distributed.elastic.multiprocessing.start_processes(*args, **kwargs)
```

[#](#api-tensorplay.distributed.elastic.multiprocessing.to_map)

### to_map

function[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.multiprocessing.to_map.html)

```python
tensorplay.distributed.elastic.multiprocessing.to_map(val_or_map: Std | dict[int, Std], local_world_size: int) → dict[int, Std]
```

Expand a per-rank redirection spec into one entry per local rank.

## Classes 5

[#](#api-tensorplay.distributed.elastic.multiprocessing.PContext)

### PContext

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.multiprocessing.PContext.html)

```python
class tensorplay.distributed.elastic.multiprocessing.PContext(name: str, entrypoint: ~collections.abc.Callable | str, args: tuple, envs: dict[int, dict[str, str]], logs_specs: ~tensorplay.distributed.elastic.multiprocessing.api.LogsSpecs | None = None, log_dir: str | None = None, redirects: ~tensorplay.distributed.elastic.multiprocessing.redirects.Std | dict[int, ~tensorplay.distributed.elastic.multiprocessing.redirects.Std] = <Std.NONE: 0>, tee: ~tensorplay.distributed.elastic.multiprocessing.redirects.Std | dict[int, ~tensorplay.distributed.elastic.multiprocessing.redirects.Std] = <Std.NONE: 0>, log_line_prefixes: dict[int, str] | None = None, duplicate_stdout_filters: list[str] | None = None, duplicate_stderr_filters: list[str] | None = None)
```

Base class owning a homogeneous group of worker processes.

```python
close(death_sig: Signals | None = None, timeout: int = 30) &#x2192; None
```

Terminate all workers with death_sig, escalating to kill.

```python
poll() &#x2192; RunProcsResult | None
```

Return the terminal result, or None while workers are running.

```python
start() &#x2192; None
```

Launch all workers.

```python
wait(timeout: float = -1, period: float = 1) &#x2192; RunProcsResult | None
```

Block until completion (or timeout seconds); returns the result.

[#](#api-tensorplay.distributed.elastic.multiprocessing.ProcessFailure)

### ProcessFailure

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.multiprocessing.ProcessFailure.html)

```python
class tensorplay.distributed.elastic.multiprocessing.ProcessFailure(local_rank: int, pid: int, exitcode: int, error_file: str | None = None, error_file_data: dict[str, Any] | None = None, message: str = '', timestamp: int = 0)
```

Structured failure of one worker process.

[#](#api-tensorplay.distributed.elastic.multiprocessing.Redirects)

### Redirects

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.multiprocessing.Redirects.html)

```python
class tensorplay.distributed.elastic.multiprocessing.Redirects(stdouts: Std | dict[int, ~tensorplay.distributed.elastic.multiprocessing.redirects.Std]=<Std.NONE: 0>, stderrs: Std | dict[int, ~tensorplay.distributed.elastic.multiprocessing.redirects.Std]=<Std.NONE: 0>)
```

Per-stream redirection modes for a worker group.

[#](#api-tensorplay.distributed.elastic.multiprocessing.RunProcsResult)

### RunProcsResult

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.multiprocessing.RunProcsResult.html)

```python
class tensorplay.distributed.elastic.multiprocessing.RunProcsResult(state: str = 'UNKNOWN', return_values: dict[int, ~typing.Any]=<factory>, failures: dict[int, ~tensorplay.distributed.elastic.multiprocessing.errors.ProcessFailure]=<factory>, stdouts: dict[int, str]=<factory>, stderrs: dict[int, str]=<factory>)
```

Outcome of monitoring a worker group to completion.

```python
is_failed() &#x2192; bool
```

Whether any worker failed.

[#](#api-tensorplay.distributed.elastic.multiprocessing.Std)

### Std

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.multiprocessing.Std.html)

```python
class tensorplay.distributed.elastic.multiprocessing.Std(*values)
```

Which standard streams a worker’s output should go to.

```python
as_integer_ratio()
```

Return a pair of integers, whose ratio is equal to the original int.

The ratio is in lowest terms and has a positive denominator.

```
>>> (10).as_integer_ratio()
(10, 1)
>>> (-10).as_integer_ratio()
(-10, 1)
>>> (0).as_integer_ratio()
(0, 1)
```

```python
bit_count()
```

Number of ones in the binary representation of the absolute value of self.

Also known as the population count.

```
>>> bin(13)
'0b1101'
>>> (13).bit_count()
3
```

```python
bit_length()
```

Number of bits necessary to represent self in binary.

```
>>> bin(37)
'0b100101'
>>> (37).bit_length()
6
```

```python
conjugate()
```

Returns self, the complex conjugate of any int.

```python
denominator
```

the denominator of a rational number in lowest terms

```python
classmethod from_bytes(bytes, byteorder='big', *, signed=False)
```

```python
classmethod from_str(vm: str) &#x2192; Std | dict[int, Std]
```

Parse to_map-style strings: 0/1/2/3 or i:j:k.

The per-rank form maps rank i (or default) to a mode; both
0:1 (shorthand for 0:1:1) and full triples are accepted.

```python
imag
```

the imaginary part of a complex number

```python
is_integer()
```

Returns True. Exists for duck type compatibility with float.is_integer.

```python
numerator
```

the numerator of a rational number in lowest terms

```python
real
```

the real part of a complex number

```python
to_bytes(length=1, byteorder='big', *, signed=False)
```

## Exceptions 2

[#](#api-tensorplay.distributed.elastic.multiprocessing.ChildFailedError)

### ChildFailedError

exception[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.multiprocessing.ChildFailedError.html)

```python
exception tensorplay.distributed.elastic.multiprocessing.ChildFailedError(name_or_failures: str | list[tuple[str, ProcessFailure]] | None = None, failures: dict[int, ProcessFailure] | None = None)
```

Raised by the launcher when one or more workers failed.

failures is a list of (role_name, ProcessFailure) pairs so the
caller can report which role failed and why.

[#](#api-tensorplay.distributed.elastic.multiprocessing.SignalException)

### SignalException

exception[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.multiprocessing.SignalException.html)

```python
exception tensorplay.distributed.elastic.multiprocessing.SignalException(msg: str, sigval: Signals)
```