741 lines
33 KiB
Python
741 lines
33 KiB
Python
"""
|
|
DisTorch Safetensor Memory Management Module
|
|
Contains all safetensor related code for distributed memory management
|
|
Following the ethos: leverage ComfyUI core, monkey-patch minimally, don't rewrite
|
|
"""
|
|
|
|
from yaml import full_load
|
|
import torch
|
|
import logging
|
|
import hashlib
|
|
import copy
|
|
from collections import defaultdict
|
|
from . import current_device
|
|
import comfy.model_management as mm
|
|
import comfy.model_patcher
|
|
import comfy.float
|
|
import comfy.utils
|
|
|
|
logger = logging.getLogger("MultiGPU")
|
|
|
|
# Global store for safetensor model allocations
|
|
safetensor_allocation_store = {}
|
|
safetensor_settings_store = {}
|
|
|
|
|
|
def create_safetensor_model_hash(model, caller):
|
|
"""Create a unique hash for a safetensor model to track allocations - EXACTLY like GGUF"""
|
|
if hasattr(model, 'model'):
|
|
# For ModelPatcher objects
|
|
actual_model = model.model
|
|
model_type = type(actual_model).__name__
|
|
# Use ComfyUI's model_size if available
|
|
if hasattr(model, 'model_size'):
|
|
model_size = model.model_size()
|
|
else:
|
|
model_size = sum(p.numel() * p.element_size() for p in actual_model.parameters())
|
|
if hasattr(model, 'model_state_dict'):
|
|
first_layers = str(list(model.model_state_dict().keys())[:3])
|
|
else:
|
|
first_layers = str(list(actual_model.state_dict().keys())[:3])
|
|
else:
|
|
# Direct model
|
|
model_type = type(model).__name__
|
|
model_size = sum(p.numel() * p.element_size() for p in model.parameters())
|
|
first_layers = str(list(model.state_dict().keys())[:3])
|
|
|
|
identifier = f"{model_type}_{model_size}_{first_layers}"
|
|
final_hash = hashlib.sha256(identifier.encode()).hexdigest()
|
|
|
|
# DEBUG STATEMENT - ALWAYS LOG THE HASH
|
|
logger.debug(f"[MULTIGPU_DISTORCHV2_HASH] Created hash for {caller}: {final_hash[:8]}...")
|
|
return final_hash
|
|
|
|
|
|
def register_patched_safetensor_modelpatcher():
|
|
"""Register the PROPERLY IMPLEMENTED monkey-patch for ModelPatcher"""
|
|
from comfy.model_patcher import wipe_lowvram_weight, move_weight_functions, LowVramPatch, CallbacksMP
|
|
|
|
if not hasattr(comfy.model_patcher.ModelPatcher, '_distorch_patched'):
|
|
# Store original methods
|
|
original_partially_load = comfy.model_patcher.ModelPatcher.partially_load
|
|
original_load = comfy.model_patcher.ModelPatcher.load
|
|
|
|
def new_partially_load(self, device_to, extra_memory=0, force_patch_weights=False):
|
|
"""
|
|
Enhanced DisTorch2 partially_load that sets up block assignments
|
|
"""
|
|
global safetensor_allocation_store
|
|
|
|
if not hasattr(self.model, '_distorch_high_precision_loras'):
|
|
logger.debug(f"[DEBUG_NEW_LOAD] high_precision_loras flag not retrieved from model. DisTorchV2 Loader not used. Reverting to normal loading behavior")
|
|
result = original_partially_load(self, device_to, extra_memory, force_patch_weights)
|
|
|
|
# Clean up
|
|
if hasattr(self, '_distorch_block_assignments'):
|
|
del self._distorch_block_assignments
|
|
|
|
return result
|
|
|
|
# Check if we have allocations for this model
|
|
model_hash = create_safetensor_model_hash(self, "partial_load")
|
|
allocations = safetensor_allocation_store.get(model_hash)
|
|
|
|
# Call original
|
|
result = original_partially_load(self, device_to, extra_memory, force_patch_weights)
|
|
|
|
# Clean up
|
|
if hasattr(self, '_distorch_block_assignments'):
|
|
del self._distorch_block_assignments
|
|
|
|
return result
|
|
|
|
def new_load(self, device_to=None, lowvram_model_memory=0, force_patch_weights=False, full_load=False):
|
|
if hasattr(self.model, '_distorch_high_precision_loras'):
|
|
high_precision_loras = self.model._distorch_high_precision_loras
|
|
else:
|
|
logger.debug(f"[MultiGPU_DisTorch2] high_precision_loras flag not retrieved from model. DisTorchV2 Loader not used. Reverting to normal loading behavior")
|
|
return original_load(self, device_to, lowvram_model_memory, force_patch_weights, full_load)
|
|
|
|
with self.use_ejected():
|
|
self.unpatch_hooks()
|
|
mem_counter = 0
|
|
loading = self._load_list()
|
|
|
|
# Check if we have DisTorch assignments
|
|
has_distorch = hasattr(self, '_distorch_block_assignments')
|
|
model_original_dtype = comfy.utils.weight_dtype(self.model.state_dict())
|
|
|
|
if not has_distorch:
|
|
logger.info(f"[MultiGPU_DisTorch2] DisTorch block assignments not found. Reverting to normal loading behavior")
|
|
return original_load(self, device_to, lowvram_model_memory, force_patch_weights, full_load)
|
|
else:
|
|
block_assignments = self._distorch_block_assignments
|
|
|
|
loading.sort(reverse=True)
|
|
for module_size, module_name, module_object, params in loading:
|
|
# Step 1: Write block/tensor to compute device first
|
|
module_object.to(device_to)
|
|
|
|
# Step 2: Apply LoRa patches while on compute device
|
|
weight_key = "{}.weight".format(module_name)
|
|
bias_key = "{}.bias".format(module_name)
|
|
|
|
if weight_key in self.patches:
|
|
self.patch_weight_to_device(weight_key, device_to=device_to)
|
|
if weight_key in self.weight_wrapper_patches:
|
|
module_object.weight_function.extend(self.weight_wrapper_patches[weight_key])
|
|
|
|
if bias_key in self.patches:
|
|
self.patch_weight_to_device(bias_key, device_to=device_to)
|
|
if bias_key in self.weight_wrapper_patches:
|
|
module_object.bias_function.extend(self.weight_wrapper_patches[bias_key])
|
|
|
|
# Step 3: FP8 casting for CPU storage (if enabled)
|
|
block_target_device = block_assignments.get(module_name, device_to)
|
|
has_patches = weight_key in self.patches or bias_key in self.patches
|
|
|
|
logger.debug(f"[MultiGPU_DisTorch2] Processing {module_name} -> block_target_device={block_target_device}")
|
|
|
|
if not high_precision_loras and block_target_device == "cpu" and has_patches and model_original_dtype in [torch.float8_e4m3fn, torch.float8_e5m2]:
|
|
logger.info(f"[MultiGPU_DisTorch2] FP8 casting conditions met for {module_name}")
|
|
for param_name, param in module_object.named_parameters():
|
|
if param.dtype.is_floating_point:
|
|
cast_data = comfy.float.stochastic_rounding(param.data, torch.float8_e4m3fn)
|
|
new_param = torch.nn.Parameter(cast_data.to(torch.float8_e4m3fn))
|
|
new_param.requires_grad = param.requires_grad
|
|
setattr(module_object, param_name, new_param)
|
|
logger.debug(f"[MultiGPU_DisTorch2] Cast {module_name}.{param_name} to FP8 for CPU storage")
|
|
|
|
# Step 4: Move to ultimate destination based on DisTorch assignment
|
|
if block_target_device != device_to:
|
|
logger.debug(f"[MultiGPU_DisTorch2] Moving {module_name} from {device_to} to {block_target_device}")
|
|
module_object.to(block_target_device)
|
|
|
|
# Mark as patched and update memory counter
|
|
module_object.comfy_patched_weights = True
|
|
mem_counter += module_size
|
|
|
|
logger.info(f"[MultiGPU_DisTorch2] DisTorch loading completed. Total memory: {mem_counter / (1024 * 1024):.2f}MB")
|
|
|
|
self.model.model_lowvram = False
|
|
self.model.device = device_to
|
|
self.model.model_loaded_weight_memory = mem_counter
|
|
self.model.current_weight_patches_uuid = self.patches_uuid
|
|
|
|
for callback in self.get_all_callbacks(comfy.patcher_extension.CallbacksMP.ON_LOAD):
|
|
callback(self, device_to, lowvram_model_memory, force_patch_weights, full_load)
|
|
|
|
self.apply_hooks(self.forced_hooks, force_apply=True)
|
|
|
|
# Apply the monkey-patches
|
|
comfy.model_patcher.ModelPatcher.partially_load = new_partially_load
|
|
comfy.model_patcher.ModelPatcher.load = new_load
|
|
comfy.model_patcher.ModelPatcher._distorch_patched = True
|
|
|
|
|
|
def analyze_safetensor_loading(model_patcher, allocations_str):
|
|
"""
|
|
Analyze and distribute safetensor model blocks across devices
|
|
IDENTICAL LOGGING FORMAT TO analyze_ggml_loading
|
|
"""
|
|
DEVICE_RATIOS_DISTORCH = {}
|
|
device_table = {}
|
|
distorch_alloc = allocations_str
|
|
virtual_vram_gb = 0.0
|
|
|
|
# Parse allocation string EXACTLY like GGML
|
|
if '#' in allocations_str:
|
|
distorch_alloc, virtual_vram_str = allocations_str.split('#')
|
|
if not distorch_alloc:
|
|
distorch_alloc = calculate_safetensor_vvram_allocation(model_patcher, virtual_vram_str)
|
|
|
|
# EXACT SAME FORMATTING AS GGML
|
|
eq_line = "=" * 50
|
|
dash_line = "-" * 50
|
|
fmt_assign = "{:<18}{:>7}{:>14}{:>10}"
|
|
|
|
# Parse device allocations
|
|
for allocation in distorch_alloc.split(';'):
|
|
if ',' not in allocation:
|
|
continue
|
|
dev_name, fraction = allocation.split(',')
|
|
fraction = float(fraction)
|
|
total_mem_bytes = mm.get_total_memory(torch.device(dev_name))
|
|
alloc_gb = (total_mem_bytes * fraction) / (1024**3)
|
|
DEVICE_RATIOS_DISTORCH[dev_name] = alloc_gb
|
|
device_table[dev_name] = {
|
|
"fraction": fraction,
|
|
"total_gb": total_mem_bytes / (1024**3),
|
|
"alloc_gb": alloc_gb
|
|
}
|
|
|
|
# IDENTICAL LOGGING TO DISTORCH
|
|
logger.info(eq_line)
|
|
logger.info(" DisTorch2 Model Device Allocations")
|
|
logger.info(eq_line)
|
|
logger.info(fmt_assign.format("Device", "Alloc %", "Total (GB)", " Alloc (GB)"))
|
|
logger.info(dash_line)
|
|
|
|
sorted_devices = sorted(device_table.keys(), key=lambda d: (d == "cpu", d))
|
|
|
|
for dev in sorted_devices:
|
|
frac = device_table[dev]["fraction"]
|
|
tot_gb = device_table[dev]["total_gb"]
|
|
alloc_gb = device_table[dev]["alloc_gb"]
|
|
logger.info(fmt_assign.format(dev,f"{int(frac * 100)}%",f"{tot_gb:.2f}",f"{alloc_gb:.2f}"))
|
|
|
|
logger.info(dash_line)
|
|
|
|
# Get the model blocks using ComfyUI's method
|
|
block_list = model_patcher._load_list()
|
|
block_list.sort(reverse=True)
|
|
|
|
# Log layer distribution
|
|
total_memory = sum(b[0] for b in block_list)
|
|
memory_by_type = defaultdict(int)
|
|
block_summary = defaultdict(int)
|
|
for module_size, module_name, module_object, params in block_list:
|
|
block_type = module_object.__class__.__name__
|
|
block_summary[block_type] += 1
|
|
memory_by_type[block_type] += module_size
|
|
|
|
# Log layer distribution - IDENTICAL FORMAT TO GGML
|
|
logger.info(" DisTorch2 Model Layer Distribution")
|
|
logger.info(dash_line)
|
|
fmt_layer = "{:<18}{:>7}{:>14}{:>10}"
|
|
logger.info(fmt_layer.format("Layer Type", "Layers", "Memory (MB)", "% Total"))
|
|
logger.info(dash_line)
|
|
|
|
for layer_type, count in block_summary.items():
|
|
mem_mb = memory_by_type[layer_type] / (1024 * 1024)
|
|
mem_percent = (memory_by_type[layer_type] / total_memory) * 100 if total_memory > 0 else 0
|
|
logger.info(fmt_layer.format(layer_type[:18], str(count), f"{mem_mb:.2f}", f"{mem_percent:.1f}%"))
|
|
|
|
logger.info(dash_line)
|
|
|
|
# Distribute blocks sequentially
|
|
device_assignments = {device: [] for device in DEVICE_RATIOS_DISTORCH.keys()}
|
|
block_assignments = {}
|
|
|
|
compute_device = str(current_device)
|
|
# Calculate total memory to be offloaded to donor devices
|
|
total_offload_gb = sum(DEVICE_RATIOS_DISTORCH.get(d, 0) for d in sorted_devices if d != compute_device)
|
|
total_offload_bytes = total_offload_gb * (1024**3)
|
|
|
|
offloaded_bytes = 0
|
|
|
|
# Iterate through the sorted list (largest blocks first)
|
|
for module_size, module_name, module_object, params in block_list:
|
|
# Assign to donor device until target is met
|
|
if offloaded_bytes < total_offload_bytes:
|
|
# For now, simple offload to CPU, will expand for multi-donor
|
|
donor_device = "cpu"
|
|
for dev in sorted_devices:
|
|
if dev != compute_device:
|
|
donor_device = dev
|
|
break # Use first available donor
|
|
|
|
block_assignments[module_name] = donor_device
|
|
setattr(module_object, 'distorch2_cpu_offload', True) # Attach the attribute here
|
|
offloaded_bytes += module_size
|
|
else:
|
|
# Assign remaining blocks to the primary compute device
|
|
block_assignments[module_name] = compute_device
|
|
|
|
# Populate device_assignments from the final block_assignments
|
|
for module_size, module_name, module_object, params in block_list:
|
|
device = block_assignments[module_name]
|
|
if device not in device_assignments:
|
|
device_assignments[device] = []
|
|
device_assignments[device].append((module_name, module_object, module_object.__class__.__name__, module_size))
|
|
|
|
# Log final assignments - IDENTICAL FORMAT TO GGML
|
|
logger.info("DisTorch2 Model Final Device/Layer Assignments")
|
|
logger.info(dash_line)
|
|
logger.info(fmt_assign.format("Device", "Layers", "Memory (MB)", "% Total"))
|
|
logger.info(dash_line)
|
|
|
|
# Log distributed blocks
|
|
total_assigned_memory = 0
|
|
device_memories = {}
|
|
|
|
for device, blocks in device_assignments.items():
|
|
device_memory = sum(b[3] for b in blocks)
|
|
device_memories[device] = device_memory
|
|
total_assigned_memory += device_memory
|
|
|
|
sorted_assignments = sorted(device_memories.keys(), key=lambda d: (d == "cpu", d))
|
|
|
|
for dev in sorted_assignments:
|
|
if dev not in device_memories:
|
|
continue
|
|
mem_mb = device_memories[dev] / (1024 * 1024)
|
|
mem_percent = (device_memories[dev] / total_memory) * 100 if total_memory > 0 else 0
|
|
logger.info(fmt_assign.format(dev, str(len(device_assignments[dev])), f"{mem_mb:.2f}", f"{mem_percent:.1f}%"))
|
|
|
|
logger.info(dash_line)
|
|
|
|
return {
|
|
"device_assignments": device_assignments,
|
|
"block_assignments": block_assignments,
|
|
"lowvram_model_memory": total_assigned_memory,
|
|
}
|
|
|
|
|
|
def analyze_safetensor_loading_main(model_patcher, allocations_str):
|
|
"""
|
|
Analyze and distribute safetensor model blocks across devices
|
|
IDENTICAL LOGGING FORMAT TO analyze_ggml_loading
|
|
"""
|
|
DEVICE_RATIOS_DISTORCH = {}
|
|
device_table = {}
|
|
distorch_alloc = allocations_str
|
|
virtual_vram_gb = 0.0
|
|
|
|
# Clear existing allocations
|
|
global safetensor_allocation_store
|
|
safetensor_allocation_store.clear()
|
|
|
|
# Parse allocation string EXACTLY like GGML
|
|
if '#' in allocations_str:
|
|
distorch_alloc, virtual_vram_str = allocations_str.split('#')
|
|
if not distorch_alloc:
|
|
distorch_alloc = calculate_safetensor_vvram_allocation(model_patcher, virtual_vram_str)
|
|
|
|
# EXACT SAME FORMATTING AS GGML
|
|
eq_line = "=" * 50
|
|
dash_line = "-" * 50
|
|
fmt_assign = "{:<18}{:>7}{:>14}{:>10}"
|
|
|
|
# Parse device allocations
|
|
for allocation in distorch_alloc.split(';'):
|
|
if ',' not in allocation:
|
|
continue
|
|
dev_name, fraction = allocation.split(',')
|
|
fraction = float(fraction)
|
|
total_mem_bytes = mm.get_total_memory(torch.device(dev_name))
|
|
alloc_gb = (total_mem_bytes * fraction) / (1024**3)
|
|
DEVICE_RATIOS_DISTORCH[dev_name] = alloc_gb
|
|
device_table[dev_name] = {
|
|
"fraction": fraction,
|
|
"total_gb": total_mem_bytes / (1024**3),
|
|
"alloc_gb": alloc_gb
|
|
}
|
|
|
|
# IDENTICAL LOGGING TO DISTORCH
|
|
logger.info(eq_line)
|
|
logger.info(" DisTorch2 Model Device Allocations")
|
|
logger.info(eq_line)
|
|
logger.info(fmt_assign.format("Device", "Alloc %", "Total (GB)", " Alloc (GB)"))
|
|
logger.info(dash_line)
|
|
|
|
sorted_devices = sorted(device_table.keys(), key=lambda d: (d == "cpu", d))
|
|
|
|
for dev in sorted_devices:
|
|
frac = device_table[dev]["fraction"]
|
|
tot_gb = device_table[dev]["total_gb"]
|
|
alloc_gb = device_table[dev]["alloc_gb"]
|
|
logger.info(fmt_assign.format(dev,f"{int(frac * 100)}%",f"{tot_gb:.2f}",f"{alloc_gb:.2f}"))
|
|
|
|
logger.info(dash_line)
|
|
|
|
# Analyze model blocks using ComfyUI's structure
|
|
block_summary = {}
|
|
block_list = []
|
|
memory_by_type = defaultdict(int)
|
|
total_memory = 0
|
|
|
|
# Get the actual model from the patcher
|
|
model = model_patcher.model if hasattr(model_patcher, 'model') else model_patcher
|
|
|
|
# First pass: calculate total memory to establish threshold
|
|
total_memory = 0
|
|
for name, module in model.named_modules():
|
|
if hasattr(module, "weight") or hasattr(module, "comfy_cast_weights"):
|
|
try:
|
|
block_memory = mm.module_size(module)
|
|
except:
|
|
block_memory = 0
|
|
if hasattr(module, 'weight') and module.weight is not None:
|
|
block_memory += module.weight.numel() * module.weight.element_size()
|
|
if hasattr(module, 'bias') and module.bias is not None:
|
|
block_memory += module.bias.numel() * module.bias.element_size()
|
|
total_memory += block_memory
|
|
|
|
# Set the minimum block size threshold (0.01% of total model memory)
|
|
MIN_BLOCK_THRESHOLD = total_memory * 0.0001
|
|
logger.debug(f"[MultiGPU_DisTorch2] Total model memory: {total_memory} bytes")
|
|
logger.debug(f"[MultiGPU_DisTorch2] Tiny block threshold (0.01%): {MIN_BLOCK_THRESHOLD} bytes")
|
|
|
|
# Second pass: analyze and collect all blocks, then filter
|
|
all_blocks = []
|
|
for name, module in model.named_modules():
|
|
if hasattr(module, "weight") or hasattr(module, "comfy_cast_weights"):
|
|
block_type = type(module).__name__
|
|
|
|
try:
|
|
block_memory = mm.module_size(module)
|
|
except:
|
|
block_memory = 0
|
|
if hasattr(module, 'weight') and module.weight is not None:
|
|
block_memory += module.weight.numel() * module.weight.element_size()
|
|
if hasattr(module, 'bias') and module.bias is not None:
|
|
block_memory += module.bias.numel() * module.bias.element_size()
|
|
|
|
# Populate summary dictionaries with ALL blocks for accurate reporting
|
|
block_summary[block_type] = block_summary.get(block_type, 0) + 1
|
|
memory_by_type[block_type] += block_memory
|
|
all_blocks.append((name, module, block_type, block_memory))
|
|
|
|
# Filter out tiny blocks from the distribution list
|
|
block_list = [b for b in all_blocks if b[3] >= MIN_BLOCK_THRESHOLD]
|
|
tiny_block_list = [b for b in all_blocks if b[3] < MIN_BLOCK_THRESHOLD]
|
|
|
|
logger.debug(f"[MultiGPU_DisTorch2] Total blocks: {len(all_blocks)}")
|
|
logger.debug(f"[MultiGPU_DisTorch2] Distributable blocks: {len(block_list)}")
|
|
logger.debug(f"[MultiGPU_DisTorch2] Tiny blocks (<0.01%): {len(tiny_block_list)}")
|
|
|
|
# Log layer distribution - IDENTICAL FORMAT TO GGML
|
|
logger.info(" DisTorch2 Model Layer Distribution")
|
|
logger.info(dash_line)
|
|
fmt_layer = "{:<18}{:>7}{:>14}{:>10}"
|
|
logger.info(fmt_layer.format("Layer Type", "Layers", "Memory (MB)", "% Total"))
|
|
logger.info(dash_line)
|
|
|
|
for layer_type, count in block_summary.items():
|
|
mem_mb = memory_by_type[layer_type] / (1024 * 1024)
|
|
mem_percent = (memory_by_type[layer_type] / total_memory) * 100 if total_memory > 0 else 0
|
|
logger.info(fmt_layer.format(layer_type[:18], str(count), f"{mem_mb:.2f}", f"{mem_percent:.1f}%"))
|
|
|
|
logger.info(dash_line)
|
|
|
|
# Distribute blocks sequentially from the tail of the model
|
|
device_assignments = {device: [] for device in DEVICE_RATIOS_DISTORCH.keys()}
|
|
block_assignments = {}
|
|
|
|
# Determine the primary compute device (first non-cpu device)
|
|
compute_device = "cuda:0" # Fallback
|
|
for dev in sorted_devices:
|
|
if dev != "cpu":
|
|
compute_device = dev
|
|
break
|
|
|
|
# Calculate total memory to be offloaded to donor devices
|
|
total_offload_gb = sum(DEVICE_RATIOS_DISTORCH.get(d, 0) for d in sorted_devices if d != compute_device)
|
|
total_offload_bytes = total_offload_gb * (1024**3)
|
|
|
|
offloaded_bytes = 0
|
|
|
|
# Iterate from the TAIL of the model
|
|
for block_name, module, block_type, block_memory in reversed(block_list):
|
|
try:
|
|
# block_memory is already calculated
|
|
pass
|
|
except:
|
|
block_memory = 0
|
|
if hasattr(module, 'weight') and module.weight is not None:
|
|
block_memory += module.weight.numel() * module.weight.element_size()
|
|
if hasattr(module, 'bias') and module.bias is not None:
|
|
block_memory += module.bias.numel() * module.bias.element_size()
|
|
|
|
# Assign to donor device (currently assumes one donor 'cpu') until target is met
|
|
if offloaded_bytes < total_offload_bytes:
|
|
# For now, simple offload to CPU, will expand for multi-donor
|
|
donor_device = "cpu"
|
|
for dev in sorted_devices:
|
|
if dev != compute_device:
|
|
donor_device = dev
|
|
break # Use first available donor
|
|
|
|
block_assignments[block_name] = donor_device
|
|
logger.info(f"[MultiGPU_DisTorch2] Assigning block to donor device: {block_name} -> {donor_device}")
|
|
offloaded_bytes += block_memory
|
|
else:
|
|
# Assign remaining blocks to the primary compute device
|
|
block_assignments[block_name] = compute_device
|
|
|
|
# Explicitly assign tiny blocks to the compute device
|
|
if tiny_block_list:
|
|
for block_name, module, block_type, block_memory in tiny_block_list:
|
|
block_assignments[block_name] = compute_device
|
|
|
|
# Populate device_assignments from the final block_assignments
|
|
for block_name, device in block_assignments.items():
|
|
# Find the block in the original list to get all its info
|
|
for b_name, b_module, b_type, b_mem in all_blocks:
|
|
if b_name == block_name:
|
|
device_assignments[device].append((b_name, b_module, b_type, b_mem))
|
|
break
|
|
|
|
# Log final assignments - IDENTICAL FORMAT TO GGML
|
|
logger.info("DisTorch2 Model Final Device/Layer Assignments")
|
|
logger.info(dash_line)
|
|
logger.info(fmt_assign.format("Device", "Layers", "Memory (MB)", "% Total"))
|
|
logger.info(dash_line)
|
|
|
|
# Calculate and log tiny blocks separately
|
|
if tiny_block_list:
|
|
tiny_block_memory = sum(b[3] for b in tiny_block_list)
|
|
tiny_mem_mb = tiny_block_memory / (1024 * 1024)
|
|
tiny_mem_percent = (tiny_block_memory / total_memory) * 100 if total_memory > 0 else 0
|
|
device_label = f"{compute_device} (<0.01%)"
|
|
logger.info(fmt_assign.format(device_label, str(len(tiny_block_list)), f"{tiny_mem_mb:.2f}", f"{tiny_mem_percent:.1f}%"))
|
|
logger.debug(f"[MultiGPU_DisTorch2] Tiny block memory breakdown: {tiny_block_memory} bytes ({tiny_mem_mb:.2f} MB), which is {tiny_mem_percent:.4f}% of total model memory.")
|
|
|
|
# Log distributed blocks
|
|
total_assigned_memory = 0
|
|
device_memories = {}
|
|
|
|
for device, blocks in device_assignments.items():
|
|
# Exclude tiny blocks from this calculation
|
|
dist_blocks = [b for b in blocks if b[3] >= MIN_BLOCK_THRESHOLD]
|
|
if not dist_blocks:
|
|
continue
|
|
|
|
device_memory = sum(b[3] for b in dist_blocks)
|
|
device_memories[device] = device_memory
|
|
total_assigned_memory += device_memory
|
|
|
|
sorted_assignments = sorted(device_memories.keys(), key=lambda d: (d == "cpu", d))
|
|
|
|
for dev in sorted_assignments:
|
|
# Get only the distributed blocks for the count
|
|
dist_blocks = [b for b in device_assignments[dev] if b[3] >= MIN_BLOCK_THRESHOLD]
|
|
if not dist_blocks:
|
|
continue
|
|
|
|
mem_mb = device_memories[dev] / (1024 * 1024)
|
|
mem_percent = (device_memories[dev] / total_memory) * 100 if total_memory > 0 else 0
|
|
logger.info(fmt_assign.format(dev, str(len(dist_blocks)), f"{mem_mb:.2f}", f"{mem_percent:.1f}%"))
|
|
|
|
logger.info(dash_line)
|
|
|
|
return {
|
|
"device_assignments": device_assignments,
|
|
"block_assignments": block_assignments
|
|
}
|
|
|
|
|
|
def calculate_safetensor_vvram_allocation(model_patcher, virtual_vram_str):
|
|
"""Calculate virtual VRAM allocation string for distributed safetensor loading"""
|
|
recipient_device, vram_amount, donors = virtual_vram_str.split(';')
|
|
virtual_vram_gb = float(vram_amount)
|
|
|
|
# EXACT SAME FORMATTING AS GGML
|
|
eq_line = "=" * 47
|
|
dash_line = "-" * 47
|
|
fmt_assign = "{:<8} {:<6} {:>11} {:>9} {:>9}"
|
|
|
|
logger.info(eq_line)
|
|
logger.info(" DisTorch2 Model Virtual VRAM Analysis")
|
|
logger.info(eq_line)
|
|
logger.info(fmt_assign.format("Object", "Role", "Original(GB)", "Total(GB)", "Virt(GB)"))
|
|
logger.info(dash_line)
|
|
|
|
# Calculate recipient VRAM
|
|
recipient_vram = mm.get_total_memory(torch.device(recipient_device)) / (1024**3)
|
|
recipient_virtual = recipient_vram + virtual_vram_gb
|
|
|
|
logger.info(fmt_assign.format(recipient_device, 'recip', f"{recipient_vram:.2f}GB",f"{recipient_virtual:.2f}GB", f"+{virtual_vram_gb:.2f}GB"))
|
|
|
|
# Handle donor devices
|
|
ram_donors = [d for d in donors.split(',')]
|
|
remaining_vram_needed = virtual_vram_gb
|
|
|
|
donor_device_info = {}
|
|
donor_allocations = {}
|
|
|
|
for donor in ram_donors:
|
|
donor_vram = mm.get_total_memory(torch.device(donor)) / (1024**3)
|
|
max_donor_capacity = donor_vram
|
|
|
|
donation = min(remaining_vram_needed, max_donor_capacity)
|
|
donor_virtual = donor_vram - donation
|
|
remaining_vram_needed -= donation
|
|
donor_allocations[donor] = donation
|
|
|
|
donor_device_info[donor] = (donor_vram, donor_virtual)
|
|
logger.info(fmt_assign.format(donor, 'donor', f"{donor_vram:.2f}GB", f"{donor_virtual:.2f}GB", f"-{donation:.2f}GB"))
|
|
|
|
|
|
logger.info(dash_line)
|
|
|
|
# Calculate model size
|
|
model = model_patcher.model if hasattr(model_patcher, 'model') else model_patcher
|
|
total_memory = 0
|
|
|
|
for name, module in model.named_modules():
|
|
if hasattr(module, "weight"):
|
|
if module.weight is not None:
|
|
total_memory += module.weight.numel() * module.weight.element_size()
|
|
if hasattr(module, "bias") and module.bias is not None:
|
|
total_memory += module.bias.numel() * module.bias.element_size()
|
|
|
|
model_size_gb = total_memory / (1024**3)
|
|
new_model_size_gb = max(0, model_size_gb - virtual_vram_gb)
|
|
|
|
logger.info(fmt_assign.format('model', 'model', f"{model_size_gb:.2f}GB",f"{new_model_size_gb:.2f}GB", f"-{virtual_vram_gb:.2f}GB"))
|
|
|
|
# Warning if model too large
|
|
if model_size_gb > (recipient_vram * 0.9):
|
|
required_offload_gb = model_size_gb - (recipient_vram * 0.9)
|
|
logger.warning(f"[MultiGPU] WARNING: Model size ({model_size_gb:.2f}GB) is larger than 90% of available VRAM on {recipient_device} ({recipient_vram * 0.9:.2f}GB).")
|
|
logger.warning(f"[MultiGPU] To prevent an OOM error, set 'virtual_vram_gb' to at least {required_offload_gb:.2f}.")
|
|
|
|
new_on_recipient = max(0, model_size_gb - virtual_vram_gb)
|
|
|
|
# Build allocation string
|
|
allocation_parts = []
|
|
recipient_percent = new_on_recipient / recipient_vram
|
|
allocation_parts.append(f"{recipient_device},{recipient_percent:.4f}")
|
|
|
|
for donor in ram_donors:
|
|
donor_vram = donor_device_info[donor][0]
|
|
donor_percent = donor_allocations[donor] / donor_vram
|
|
allocation_parts.append(f"{donor},{donor_percent:.4f}")
|
|
|
|
allocation_string = ";".join(allocation_parts)
|
|
|
|
fmt_mem = "{:<20}{:>20}"
|
|
logger.info(fmt_mem.format("\n v2 Expert String", allocation_string))
|
|
|
|
return allocation_string
|
|
|
|
|
|
def override_class_with_distorch_safetensor_v2(cls):
|
|
"""DisTorch 2.0 wrapper for safetensor models - EXACTLY like GGUF wrapper"""
|
|
from .nodes import get_device_list
|
|
from . import current_device
|
|
|
|
class NodeOverrideDisTorchSafetensorV2(cls):
|
|
@classmethod
|
|
def INPUT_TYPES(s):
|
|
inputs = copy.deepcopy(cls.INPUT_TYPES())
|
|
devices = get_device_list()
|
|
compute_device = devices[1] if len(devices) > 1 else devices[0]
|
|
|
|
inputs["optional"] = inputs.get("optional", {})
|
|
inputs["optional"]["compute_device"] = (devices, {"default": compute_device})
|
|
inputs["optional"]["virtual_vram_gb"] = ("FLOAT", {"default": 4.0, "min": 0.0, "max": 128.0, "step": 0.1})
|
|
inputs["optional"]["donor_device"] = (devices, {"default": "cpu"})
|
|
inputs["optional"]["expert_mode_allocations"] = ("STRING", {"multiline": False, "default": ""})
|
|
inputs["optional"]["high_precision_loras"] = ("BOOLEAN", {"default": True})
|
|
return inputs
|
|
|
|
CATEGORY = "multigpu/distorch_2"
|
|
FUNCTION = "override"
|
|
TITLE = f"{cls.TITLE if hasattr(cls, 'TITLE') else cls.__name__} (DisTorch2)"
|
|
|
|
@classmethod
|
|
def IS_CHANGED(s, *args, compute_device=None, virtual_vram_gb=4.0,
|
|
donor_device="cpu", expert_mode_allocations="", high_precision_loras=True, **kwargs):
|
|
# Create a hash of our specific settings
|
|
settings_str = f"{compute_device}{virtual_vram_gb}{donor_device}{expert_mode_allocations}{high_precision_loras}"
|
|
return hashlib.sha256(settings_str.encode()).hexdigest()
|
|
|
|
def override(self, *args, compute_device=None, virtual_vram_gb=4.0,
|
|
donor_device="cpu", expert_mode_allocations="", high_precision_loras=True, **kwargs):
|
|
|
|
from . import set_current_device
|
|
if compute_device is not None:
|
|
set_current_device(compute_device)
|
|
|
|
# Register our patched ModelPatcher
|
|
register_patched_safetensor_modelpatcher()
|
|
|
|
# Call original function
|
|
fn = getattr(super(), cls.FUNCTION)
|
|
|
|
# --- Check if we need to unload the model due to settings change ---
|
|
# This logic is a bit redundant with IS_CHANGED, but provides clear logging
|
|
settings_str = f"{compute_device}{virtual_vram_gb}{donor_device}{expert_mode_allocations}"
|
|
settings_hash = hashlib.sha256(settings_str.encode()).hexdigest()
|
|
|
|
# Temporarily load to get hash without applying our patch
|
|
temp_out = fn(*args, **kwargs)
|
|
model_to_check = None
|
|
if hasattr(temp_out[0], 'model'):
|
|
model_to_check = temp_out[0]
|
|
elif hasattr(temp_out[0], 'patcher') and hasattr(temp_out[0].patcher, 'model'):
|
|
model_to_check = temp_out[0].patcher
|
|
|
|
if model_to_check:
|
|
model_hash = create_safetensor_model_hash(model_to_check, "override_check")
|
|
last_settings_hash = safetensor_settings_store.get(model_hash)
|
|
|
|
if last_settings_hash != settings_hash:
|
|
logger.info(f"[MultiGPU_DisTorch2] Settings changed for model {model_hash[:8]}. Previous settings hash: {last_settings_hash}, New settings hash: {settings_hash}. Forcing reload.")
|
|
# The IS_CHANGED mechanism should handle the reload, this is for logger.
|
|
else:
|
|
logger.info(f"[MultiGPU_DisTorch2] Settings unchanged for model {model_hash[:8]}. Using cached model.")
|
|
|
|
out = fn(*args, **kwargs)
|
|
|
|
# Store high_precision_loras in the model for later retrieval
|
|
if hasattr(out[0], 'model'):
|
|
out[0].model._distorch_high_precision_loras = high_precision_loras
|
|
elif hasattr(out[0], 'patcher') and hasattr(out[0].patcher, 'model'):
|
|
out[0].patcher.model._distorch_high_precision_loras = high_precision_loras
|
|
|
|
vram_string = ""
|
|
if virtual_vram_gb > 0:
|
|
vram_string = f"{compute_device};{virtual_vram_gb};{donor_device}"
|
|
|
|
full_allocation = f"{expert_mode_allocations}#{vram_string}" if expert_mode_allocations or vram_string else ""
|
|
|
|
logger.info(f"[MultiGPU_DisTorch2] Full allocation string: {full_allocation}")
|
|
|
|
if hasattr(out[0], 'model'):
|
|
model_hash = create_safetensor_model_hash(out[0], "override")
|
|
safetensor_allocation_store[model_hash] = full_allocation
|
|
safetensor_settings_store[model_hash] = settings_hash
|
|
elif hasattr(out[0], 'patcher') and hasattr(out[0].patcher, 'model'):
|
|
model_hash = create_safetensor_model_hash(out[0].patcher, "override")
|
|
safetensor_allocation_store[model_hash] = full_allocation
|
|
safetensor_settings_store[model_hash] = settings_hash
|
|
|
|
return out
|
|
|
|
return NodeOverrideDisTorchSafetensorV2
|