Skip to content

Commit

Permalink
[hotfix] fix flags for run command, fix hotkeys flag for overview, an…
Browse files Browse the repository at this point in the history 
…d [feature] CUDA reg (#877)

* [feature] cpu register faster (#854)

* add update interval and num proc flag

* add better number output

* optimize multiproc cpu reg
keeping proc until solution

* fix test

* make sure to exit properly if registered during

* fix tests

* change import to use tests

* add optional type hints and None default

* change to count using allowed processes

* add documentation. Fix random start

* [hotfix] fix flags for multiproc register limit (#876)

* add dot get

* add to subtensor args and defaults

* remove dot get because in subtensor args

* typo

* fix test

* Fix/diff unpack bit shift (#878)

* fix incorrect bit shift

* move inner function out and add test for diff pack

* fix test

* fix call arg check in test

* add assert

* fix test for py37

* refactor the diff pack into two functions
move the test to a unit test

* fix test

* [Feature] [cubit] CUDA registration solver (#868)

* added cuda solver

* boost versions to fix pip error

* allow choosing device id

* fix solution check to use keccak

* adds params for cuda and dev_id to register

* list devices by name during selection

* add block number logging

* fix calculation of hashrate

* fix update interval default

* add --TPB arg to register

* add update_interval flag

* switch back to old looping/work structure

* change typing

* device count is a function

* stop early if wallet registered

* add update interval and num proc flag

* add better number output

* optimize multiproc cpu reg
keeping proc until solution

* fix test

* change import to cubit

* fix import and default

* up default
should have default in CLI call

* add comments about params

* fix config var access

* add cubit as extra

* handle stale pow differently
check registration after failure

* restrict number of processes for integration test

* fix stale check

* use wallet.is_registered instead

* attempt to fix test issue

* fix my test

* oops typo

* typo again ugh

* remove print out

* fix partly reg test

* fix if solution None

* fix test?

* fix patch

* add args for cuda to subtensor

* add cuda args to reregister call

* add to wallet register the cuda args

* fix refs and tests

* add for val test also

* fix tests with rereg

* fix patch for tests

* add mock_register to subtensor passed instead

* move register under the check for isregistered

* use patch obj instead

* fit patch object

* Fix/move overview args to cli (#867)

* move cli args to CLI and fix overview

* use dot get

* fix tests

* add hotkeys/all_hotkeys to (un)stake

* fix default

* fix default in unstake

* add unstake multiple

* add add stake multiple

* move all/hotkeys back to wallet args

* convert to balance first
add catch for unstake multi

* fix ref to wallet

* fix test patch for multi hotkeys

* try to fix tests

* fix tests patch

* fix mock wallet length

* don't use new?

* fix call args get

* typo

* fix typo

* Add/address CUDA reg changes (#879)

* added cuda solver

* boost versions to fix pip error

* allow choosing device id

* fix solution check to use keccak

* adds params for cuda and dev_id to register

* list devices by name during selection

* add block number logging

* fix calculation of hashrate

* fix update interval default

* add --TPB arg to register

* add update_interval flag

* switch back to old looping/work structure

* change typing

* device count is a function

* stop early if wallet registered

* add update interval and num proc flag

* add better number output

* optimize multiproc cpu reg
keeping proc until solution

* fix test

* change import to cubit

* fix import and default

* up default
should have default in CLI call

* add comments about params

* fix config var access

* add cubit as extra

* handle stale pow differently
check registration after failure

* restrict number of processes for integration test

* fix stale check

* use wallet.is_registered instead

* attempt to fix test issue

* fix my test

* oops typo

* typo again ugh

* remove print out

* fix partly reg test

* fix if solution None

* fix test?

* fix patch

* add args for cuda to subtensor

* add cuda args to reregister call

* add to wallet register the cuda args

* fix refs and tests

* add for val test also

* fix tests with rereg

* fix patch for tests

* add mock_register to subtensor passed instead

* move register under the check for isregistered

* use patch obj instead

* fit patch object

* fix prompt

* remove unneeded if

* modify POW submit to use rolling submit again

* add backoff to block get from network

* add test for backoff get block

* suppress the dev id flag if not set

* remove dest so it uses first arg

* fix pow submit loop

* move registration status with

* fix max attempts check

* remove status in subtensor.register

* add submit status

* change to neuron get instead

* fix count

* try to patch live display

* fix patch

* .

* separate test cases

* add POWNotStale and tests

* add more test cases for block get with retry

* fix return to None

* fix arg order

Co-authored-by: Eugene <etesting007@gmail.com>
  • Loading branch information
@Eugene-hu
Cameron Fairchild and Eugene-hu authored Aug 16, 2022
1 parent e2ca6a9 commit b54523e
Show file tree
Hide file tree
Showing 15 changed files with 1,087 additions and 242 deletions.
76 changes: 54 additions & 22 deletions bittensor/_cli/__init__.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -18,15 +18,18 @@
# OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
# DEALINGS IN THE SOFTWARE.

import argparse
import os
import sys
import argparse
from typing import List

import bittensor
from rich.prompt import Prompt
from rich.prompt import Confirm
import torch
from rich.prompt import Confirm, Prompt
from substrateinterface.utils.ss58 import ss58_decode, ss58_encode

from . import cli_impl

console = bittensor.__console__

class cli:
Expand Down Expand Up @@ -88,6 +91,27 @@ def config() -> 'bittensor.config':
help='''Set the output width of the overview. Defaults to automatic width from terminal.''',
default=None,
)
overview_parser.add_argument(
'--sort_by',
'--wallet.sort_by',
dest='sort_by',
required=False,
action='store',
default="",
type=str,
help='''Sort the hotkeys by the specified column title (e.g. name, uid, axon).'''
)
overview_parser.add_argument(
'--sort_order',
'--wallet.sort_order',
dest="sort_order",
required=False,
action='store',
default="ascending",
type=str,
help='''Sort the hotkeys in the specified ordering. (ascending/asc or descending/desc/reverse)'''
)

bittensor.wallet.add_args( overview_parser )
bittensor.subtensor.add_args( overview_parser )

Expand Down Expand Up @@ -250,6 +274,7 @@ def config() -> 'bittensor.config':
'register',
help='''Register a wallet to a network.'''
)

unstake_parser = cmd_parsers.add_parser(
'unstake',
help='''Unstake from hotkey accounts.'''
Expand Down Expand Up @@ -497,6 +522,7 @@ def config() -> 'bittensor.config':
help='''Set true to avoid prompting the user.''',
default=False,
)

bittensor.wallet.add_args( unstake_parser )
bittensor.subtensor.add_args( unstake_parser )

Expand Down Expand Up @@ -535,6 +561,7 @@ def config() -> 'bittensor.config':
help='''Set true to avoid prompting the user.''',
default=False,
)

bittensor.wallet.add_args( stake_parser )
bittensor.subtensor.add_args( stake_parser )

Expand Down Expand Up @@ -571,23 +598,6 @@ def config() -> 'bittensor.config':
help='''Set true to avoid prompting the user.''',
default=False,
)
register_parser.add_argument(
'--num_processes',
'--num',
'-n',
dest='num_processes',
help="Number of processors to use for registration",
type=int,
default=None,
)
register_parser.add_argument(
'--update_interval',
'-u',
dest='update_interval',
help="The number of nonces to process before checking for next block during registration",
type=int,
default=None,
)

bittensor.wallet.add_args( register_parser )
bittensor.subtensor.add_args( register_parser )
Expand Down Expand Up @@ -709,7 +719,7 @@ def check_unstake_config( config: 'bittensor.Config' ):
if config.wallet.get('all_hotkeys'):
hotkeys = "all hotkeys"
elif config.wallet.get('hotkeys'):
hotkeys = str(config.hotkeys).replace('[', '').replace(']', '')
hotkeys = str(config.wallet.hotkeys).replace('[', '').replace(']', '')
else:
hotkeys = str(config.wallet.hotkey)
if not Confirm.ask("Unstake all Tao from: [bold]'{}'[/bold]?".format(hotkeys)):
Expand Down Expand Up @@ -819,6 +829,28 @@ def check_register_config( config: 'bittensor.Config' ):
hotkey = Prompt.ask("Enter hotkey name", default = bittensor.defaults.wallet.hotkey)
config.wallet.hotkey = str(hotkey)

if not config.no_prompt and config.subtensor.register.cuda.use_cuda == bittensor.defaults.subtensor.register.cuda.use_cuda:
# Ask about cuda registration only if a CUDA device is available.
if torch.cuda.is_available():
cuda = Confirm.ask("Detected CUDA device, use CUDA for registration?\n")
config.subtensor.register.cuda.use_cuda = cuda
# Only ask about which CUDA device if the user has more than one CUDA device.
if cuda and config.subtensor.register.cuda.get('dev_id') is None and torch.cuda.device_count() > 0:
devices: List[str] = [str(x) for x in range(torch.cuda.device_count())]
device_names: List[str] = [torch.cuda.get_device_name(x) for x in range(torch.cuda.device_count())]
console.print("Available CUDA devices:")
choices_str: str = ""
for i, device in enumerate(devices):
choices_str += (" {}: {}\n".format(device, device_names[i]))
console.print(choices_str)
dev_id = Prompt.ask("Which GPU would you like to use?", choices=devices, default=str(bittensor.defaults.subtensor.register.cuda.dev_id))
try:
dev_id = int(dev_id)
except ValueError:
console.error(":cross_mark:[red]Invalid GPU device[/red] [bold white]{}[/bold white]\nAvailable CUDA devices:{}".format(dev_id, choices_str))
sys.exit(1)
config.subtensor.register.cuda.dev_id = dev_id

def check_new_coldkey_config( config: 'bittensor.Config' ):
if config.wallet.get('name') == bittensor.defaults.wallet.name and not config.no_prompt:
wallet_name = Prompt.ask("Enter wallet name", default = bittensor.defaults.wallet.name)
Expand Down Expand Up @@ -902,4 +934,4 @@ def check_help_config( config: 'bittensor.Config'):
def check_update_config( config: 'bittensor.Config'):
if not config.no_prompt:
answer = Prompt.ask('This will update the local bittensor package', choices = ['Y','N'], default = 'Y')
config.answer = answer
config.answer = answer
65 changes: 32 additions & 33 deletions bittensor/_cli/cli_impl.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@

import os
import sys
from typing import List, Union
from typing import List, Union, Optional

from cachetools import Cache

Expand Down Expand Up @@ -239,7 +239,15 @@ def register( self ):
"""
wallet = bittensor.wallet( config = self.config )
subtensor = bittensor.subtensor( config = self.config )
subtensor.register( wallet = wallet, prompt = not self.config.no_prompt, num_processes = self.config.num_processes, update_interval = self.config.update_interval )
subtensor.register(
wallet = wallet,
prompt = not self.config.no_prompt,
TPB = self.config.subtensor.register.cuda.get('TPB', None),
update_interval = self.config.subtensor.register.get('update_interval', None),
num_processes = self.config.subtensor.register.get('num_processes', None),
cuda = self.config.subtensor.register.cuda.get('use_cuda', None),
dev_id = self.config.subtensor.register.cuda.get('dev_id', None)
)

def transfer( self ):
r""" Transfer token of amount to destination.
Expand All @@ -251,7 +259,6 @@ def transfer( self ):
def unstake( self ):
r""" Unstake token of amount from hotkey(s).
"""
# TODO: Implement this without re-unlocking the coldkey.
config = self.config.copy()
config.hotkey = None
wallet = bittensor.wallet( config = self.config )
Expand All @@ -263,7 +270,7 @@ def unstake( self ):
all_hotkeys: List[bittensor.wallet] = self._get_hotkey_wallets_for_wallet( wallet = wallet )
# Exclude hotkeys that are specified.
wallets_to_unstake_from = [
wallet for wallet in all_hotkeys if wallet.hotkey_str not in self.config.wallet.get('hotkeys')
wallet for wallet in all_hotkeys if wallet.hotkey_str not in self.config.wallet.get('hotkeys', [])
]

elif self.config.wallet.get('hotkeys'):
Expand All @@ -276,9 +283,7 @@ def unstake( self ):
subtensor.unstake( wallet, amount = None if self.config.get('unstake_all') else self.config.get('amount'), wait_for_inclusion = True, prompt = not self.config.no_prompt )
return None

wallet_0: 'bittensor.wallet' = wallets_to_unstake_from[0]
# Decrypt coldkey for all wallet(s) to use
wallet_0.coldkey


final_wallets: List['bittensor.wallet'] = []
final_amounts: List[Union[float, Balance]] = []
Expand All @@ -287,9 +292,6 @@ def unstake( self ):
if not wallet.is_registered():
# Skip unregistered hotkeys.
continue
# Assign decrypted coldkey from wallet_0
# so we don't have to decrypt again
wallet._coldkey = wallet_0._coldkey

unstake_amount_tao: float = self.config.get('amount')
if self.config.get('max_stake'):
Expand All @@ -307,19 +309,17 @@ def unstake( self ):
if not self.config.no_prompt:
if not Confirm.ask("Do you want to unstake from the following keys:\n" + \
"".join([
f" [bold white]- {wallet.hotkey_str}: {amount}𝜏[/bold white]\n" for wallet, amount in zip(final_wallets, final_amounts)
f" [bold white]- {wallet.hotkey_str}: {amount.tao}𝜏[/bold white]\n" for wallet, amount in zip(final_wallets, final_amounts)
])
):
return None

for wallet, amount in zip(final_wallets, final_amounts):
subtensor.unstake( wallet, amount = None if self.config.get('unstake_all') else amount, wait_for_inclusion = True, prompt = False )

subtensor.unstake_multiple( wallets = final_wallets, amounts = None if self.config.get('unstake_all') else final_amounts, wait_for_inclusion = True, prompt = False )


def stake( self ):
r""" Stake token of amount to hotkey(s).
"""
# TODO: Implement this without re-unlocking the coldkey.
config = self.config.copy()
config.hotkey = None
wallet = bittensor.wallet( config = config )
Expand All @@ -331,7 +331,7 @@ def stake( self ):
all_hotkeys: List[bittensor.wallet] = self._get_hotkey_wallets_for_wallet( wallet = wallet )
# Exclude hotkeys that are specified.
wallets_to_stake_to = [
wallet for wallet in all_hotkeys if wallet.hotkey_str not in self.config.wallet.get('hotkeys')
wallet for wallet in all_hotkeys if wallet.hotkey_str not in self.config.wallet.get('hotkeys', [])
]

elif self.config.wallet.get('hotkeys'):
Expand Down Expand Up @@ -386,8 +386,7 @@ def stake( self ):
):
return None

for wallet, amount in zip(final_wallets, final_amounts):
subtensor.add_stake( wallet, amount = None if self.config.get('stake_all') else amount, wait_for_inclusion = True, prompt = False )
subtensor.add_stake_multiple( wallets = final_wallets, amounts = None if self.config.get('stake_all') else final_amounts, wait_for_inclusion = True, prompt = False )


def set_weights( self ):
Expand Down Expand Up @@ -596,24 +595,14 @@ def overview(self):

all_hotkeys = []
total_balance = bittensor.Balance(0)

# We are printing for every wallet.
# We are printing for every coldkey.
if self.config.all:
cold_wallets = CLI._get_coldkey_wallets_for_path(self.config.wallet.path)
for cold_wallet in tqdm(cold_wallets, desc="Pulling balances"):
if cold_wallet.coldkeypub_file.exists_on_device() and not cold_wallet.coldkeypub_file.is_encrypted():
total_balance = total_balance + subtensor.get_balance( cold_wallet.coldkeypub.ss58_address )
all_hotkeys = CLI._get_all_wallets_for_path( self.config.wallet.path )

# We are printing for a select number of hotkeys.
elif self.config.wallet.hotkeys:
# Only show hotkeys for wallets in the list
all_hotkeys = [hotkey for hotkey in all_hotkeys if hotkey.hotkey_str in self.config.wallet.hotkeys]
coldkey_wallet = bittensor.wallet( config = self.config )
if coldkey_wallet.coldkeypub_file.exists_on_device() and not coldkey_wallet.coldkeypub_file.is_encrypted():
total_balance = subtensor.get_balance( coldkey_wallet.coldkeypub.ss58_address )

# We are printing for all keys under the wallet.
else:
# We are only printing keys for a single coldkey
coldkey_wallet = bittensor.wallet( config = self.config )
Expand All @@ -624,6 +613,16 @@ def overview(self):
return
all_hotkeys = CLI._get_hotkey_wallets_for_wallet( coldkey_wallet )

# We are printing for a select number of hotkeys from all_hotkeys.

if self.config.wallet.get('hotkeys', []):
if not self.config.get('all_hotkeys', False):
# We are only showing hotkeys that are specified.
all_hotkeys = [hotkey for hotkey in all_hotkeys if hotkey.hotkey_str in self.config.wallet.hotkeys]
else:
# We are excluding the specified hotkeys from all_hotkeys.
all_hotkeys = [hotkey for hotkey in all_hotkeys if hotkey.hotkey_str not in self.config.wallet.hotkeys]

# Check we have keys to display.
if len(all_hotkeys) == 0:
console.print("[red]No wallets found.[/red]")
Expand Down Expand Up @@ -732,10 +731,10 @@ def overview(self):

console.clear()

sort_by: str = self.config.wallet.sort_by
sort_order: str = self.config.wallet.sort_order
sort_by: Optional[str] = self.config.get('sort_by', None)
sort_order: Optional[str] = self.config.get('sort_order', None)

if sort_by != "":
if sort_by is not None and sort_by != "":
column_to_sort_by: int = 0
highest_matching_ratio: int = 0
sort_descending: bool = False # Default sort_order to ascending
Expand Down
3 changes: 1 addition & 2 deletions bittensor/_dataset/dataset_impl.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -160,8 +160,7 @@ def __init__(

self.build_hash_table()

if not os.path.isdir(os.path.expanduser(data_dir)):
os.makedirs(os.path.expanduser(data_dir))
os.makedirs(os.path.expanduser(data_dir), exist_ok=True)

self.data_queue = ThreadQueue(
producer_target = self.reserve_multiple_data,
Expand Down
17 changes: 17 additions & 0 deletions bittensor/_subtensor/__init__.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -184,6 +184,14 @@ def add_args(cls, parser: argparse.ArgumentParser, prefix: str = None ):
help='''The subtensor endpoint flag. If set, overrides the --network flag.
''')
parser.add_argument('--' + prefix_str + 'subtensor._mock', action='store_true', help='To turn on subtensor mocking for testing purposes.', default=bittensor.defaults.subtensor._mock)

parser.add_argument('--' + prefix_str + 'subtensor.register.num_processes', '-n', dest='subtensor.register.num_processes', help="Number of processors to use for registration", type=int, default=bittensor.defaults.subtensor.register.num_processes)
parser.add_argument('--' + prefix_str + 'subtensor.register.update_interval', '--' + prefix_str + 'subtensor.register.cuda.update_interval', '--' + prefix_str + 'cuda.update_interval', '-u', help="The number of nonces to process before checking for next block during registration", type=int, default=bittensor.defaults.subtensor.register.update_interval)
# registration args. Used for register and re-register and anything that calls register.
parser.add_argument( '--' + prefix_str + 'subtensor.register.cuda.use_cuda', '--' + prefix_str + 'cuda', '--' + prefix_str + 'cuda.use_cuda', default=bittensor.defaults.subtensor.register.cuda.use_cuda, help='''Set true to use CUDA.''', action='store_true', required=False )
parser.add_argument( '--' + prefix_str + 'subtensor.register.cuda.dev_id', '--' + prefix_str + 'cuda.dev_id', type=int, default=argparse.SUPPRESS, help='''Set the CUDA device id. Goes by the order of speed. (i.e. 0 is the fastest).''', required=False )
parser.add_argument( '--' + prefix_str + 'subtensor.register.cuda.TPB', '--' + prefix_str + 'cuda.TPB', type=int, default=bittensor.defaults.subtensor.register.cuda.TPB, help='''Set the number of Threads Per Block for CUDA.''', required=False )

except argparse.ArgumentError:
# re-parsing arguments.
pass
Expand All @@ -197,6 +205,15 @@ def add_defaults(cls, defaults ):
defaults.subtensor.chain_endpoint = os.getenv('BT_SUBTENSOR_CHAIN_ENDPOINT') if os.getenv('BT_SUBTENSOR_CHAIN_ENDPOINT') != None else None
defaults.subtensor._mock = os.getenv('BT_SUBTENSOR_MOCK') if os.getenv('BT_SUBTENSOR_MOCK') != None else False

defaults.subtensor.register = bittensor.Config()
defaults.subtensor.register.num_processes = os.getenv('BT_SUBTENSOR_REGISTER_NUM_PROCESSES') if os.getenv('BT_SUBTENSOR_REGISTER_NUM_PROCESSES') != None else None # uses processor count by default within the function
defaults.subtensor.register.update_interval = os.getenv('BT_SUBTENSOR_REGISTER_UPDATE_INTERVAL') if os.getenv('BT_SUBTENSOR_REGISTER_UPDATE_INTERVAL') != None else 50_000

defaults.subtensor.register.cuda = bittensor.Config()
defaults.subtensor.register.cuda.dev_id = 0
defaults.subtensor.register.cuda.use_cuda = False
defaults.subtensor.register.cuda.TPB = 256

@staticmethod
def check_config( config: 'bittensor.Config' ):
assert config.subtensor
Expand Down
Loading

0 comments on commit b54523e

Please sign in to comment.