[pull] master from Comfy-Org:master

comfy/cli_args.py

@@ -115,6 +115,7 @@ def from_string(cls, value: str):
 cache_group.add_argument("--cache-classic", action="store_true", help="Use the old style (aggressive) caching.")
 cache_group.add_argument("--cache-lru", type=int, default=0, help="Use LRU caching with a maximum of N node results cached. May use more RAM/VRAM.")
 cache_group.add_argument("--cache-none", action="store_true", help="Reduced RAM/VRAM usage at the expense of executing every node for each run.")
+cache_group.add_argument("--high-ram", action="store_true", help="Can improve performance slightly on high RAM or on systems where pagefile use is preferred over model loading.")
 
 attn_group = parser.add_mutually_exclusive_group()
 attn_group.add_argument("--use-split-cross-attention", action="store_true", help="Use the split cross attention optimization. Ignored when xformers is used.")
@@ -249,6 +250,9 @@ def is_valid_directory(path: str) -> str:
 if args.cache_ram is not None and len(args.cache_ram) > 2:
     parser.error("--cache-ram accepts at most two values: active GB and inactive GB")
 
+if args.high_ram:
+    args.cache_classic = True
+
 if args.windows_standalone_build:
     args.auto_launch = True
 

comfy/ldm/ideogram4/model.py

@@ -106,11 +106,11 @@ def __init__(self, dim, input_range=(0.0, 1.0), dtype=None, device=None, operati
         self.mlp_in = operations.Linear(dim, dim, bias=True, dtype=dtype, device=device)
         self.mlp_out = operations.Linear(dim, dim, bias=True, dtype=dtype, device=device)
 
-    def forward(self, x):
+    def forward(self, x, dtype):
         x = x.to(torch.float32)
         scaled = 1e4 * (x - self.range_min) / (self.range_max - self.range_min)
         emb = _sinusoidal_embedding(scaled, self.dim)
-        emb = emb.to(self.mlp_in.weight.dtype)
+        emb = emb.to(dtype)
         emb = F.silu(self.mlp_in(emb))
         return self.mlp_out(emb)
 
@@ -161,7 +161,7 @@ def _backbone(self, llm_features, x, t, position_ids, attn_mask, indicator, tran
         x = x * output_image_mask
         h = self.input_proj(x) * output_image_mask
 
-        t_cond = self.t_embedding(t)
+        t_cond = self.t_embedding(t, dtype=x.dtype)
         if t.dim() == 1:
             t_cond = t_cond.unsqueeze(1)
         adaln_input = F.silu(self.adaln_proj(t_cond))

comfy/model_management.py

@@ -643,6 +643,8 @@ def free_pins(size, evict_active=False):
     return freed_total
 
 def ensure_pin_budget(size, evict_active=False):
+    if args.high_ram:
+        return True
     if args.fast_disk:
         shortfall = TOTAL_PINNED_MEMORY + size - MAX_PINNED_MEMORY
     else:
@@ -1496,6 +1498,8 @@ def cast_to_device(tensor, device, dtype, copy=False):
 PINNING_ALLOWED_TYPES = set(["Tensor", "Parameter", "QuantizedTensor"])
 
 def pinned_hostbuf_size(size):
+    if args.high_ram:
+        return max(0, int(size * 2))
     return max(0, int(min(size, MAX_PINNED_MEMORY) * 2))
 
 def discard_cuda_async_error():

comfy/ops.py

@@ -180,7 +180,7 @@ def handle_pin(m, pin, source, dest, subset="weights", size=None):
             if pin is not None:
                 cast_maybe_lowvram_patch([pin], dest, offload_stream)
                 return
-            if signature is None:
+            if signature is None or args.high_ram:
                 comfy.pinned_memory.pin_memory(m, subset=subset, size=size)
                 pin = comfy.pinned_memory.get_pin(m, subset=subset)
             cast_maybe_lowvram_patch(source, pin, offload_stream, xfer_dest2=dest)