M3Drop 0.4.33__py3-none-any.whl → 0.4.34__py3-none-any.whl

m3Drop/normalizationGPU.py

@@ -19,14 +19,15 @@ def NBumiPearsonResidualsGPU(
     output_filename: str
 ):
     """
-    Calculates Pearson residuals. Safe Mode: Multiplier increased to 6.0.
+    Calculates Pearson residuals. Safe Mode: Multiplier increased to 10.0.
     """
     start_time = time.perf_counter()
     print(f"FUNCTION: NBumiPearsonResiduals() | FILE: {cleaned_filename}")
 
     # --- SAFETY UPDATE ---
-    # Multiplier 6.0 (Was 4.0): Forces smaller chunks to avoid fragmentation crashes.
-    chunk_size = get_optimal_chunk_size(cleaned_filename, multiplier=6.0, is_dense=True)
+    # Multiplier 10.0 (Was 6.0): Accounts for Float64 precision (8 bytes) vs Governor default (4 bytes).
+    # 4 matrices * 8 bytes = 32 bytes/cell. Governor 10 * 4 = 40 bytes. Safe buffer established.
+    chunk_size = get_optimal_chunk_size(cleaned_filename, multiplier=10.0, is_dense=True)
 
     # --- Phase 1: Initialization ---
     print("Phase [1/2]: Initializing parameters and preparing output file...")
@@ -98,7 +99,8 @@ def NBumiPearsonResidualsGPU(
                 pearson_chunk_gpu = (counts_chunk_dense_gpu - mus_chunk_gpu) / denominator_gpu
                 
                 # Write to Disk
-                out_x[i:end_row, :] = pearson_chunk_gpu.get()
+                # [OPTIMIZATION] Cast to float32 on GPU to halve PCIe transfer time
+                out_x[i:end_row, :] = pearson_chunk_gpu.astype(cupy.float32).get()
                 
                 # Cleanup
                 del counts_chunk_dense_gpu, counts_chunk_sparse_gpu, mus_chunk_gpu, pearson_chunk_gpu, denominator_gpu
@@ -126,8 +128,8 @@ def NBumiPearsonResidualsApproxGPU(
     print(f"FUNCTION: NBumiPearsonResidualsApprox() | FILE: {cleaned_filename}")
 
     # --- HANDSHAKE ---
-    # Multiplier 6.0: Same safety logic as Full residuals.
-    chunk_size = get_optimal_chunk_size(cleaned_filename, multiplier=6.0, is_dense=True)
+    # Multiplier 10.0: Same safety logic as Full residuals.
+    chunk_size = get_optimal_chunk_size(cleaned_filename, multiplier=10.0, is_dense=True)
 
     # --- Phase 1: Initialization ---
     print("Phase [1/2]: Initializing parameters and preparing output file...")
@@ -191,7 +193,8 @@ def NBumiPearsonResidualsApproxGPU(
                 
                 pearson_chunk_gpu = (counts_chunk_dense_gpu - mus_chunk_gpu) / denominator_gpu
                 
-                out_x[i:end_row, :] = pearson_chunk_gpu.get()
+                # [OPTIMIZATION] Cast to float32 on GPU to halve PCIe transfer time
+                out_x[i:end_row, :] = pearson_chunk_gpu.astype(cupy.float32).get()
                 
                 del counts_chunk_dense_gpu, counts_chunk_sparse_gpu, mus_chunk_gpu, pearson_chunk_gpu, denominator_gpu
                 cupy.get_default_memory_pool().free_all_blocks()

{m3drop-0.4.33.dist-info → m3drop-0.4.34.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: M3Drop
-Version: 0.4.33
+Version: 0.4.34
 Summary: A Python implementation of the M3Drop single-cell RNA-seq analysis tool.
 Home-page: https://github.com/PragalvhaSharma/m3DropNew
 Author: Tallulah Andrews

{m3drop-0.4.33.dist-info → m3drop-0.4.34.dist-info}/RECORD

@@ -4,9 +4,9 @@ m3Drop/coreGPU.py,sha256=k7A06VNgfJ59J8g1VpfKxhTIKrEbW7Bj8pTbQqHaQL8,24571
 m3Drop/diagnosticsCPU.py,sha256=BecOKTz2GDjzjs9ycXYsyrSHi2UVgsM58RBuNE62vmU,14273
 m3Drop/diagnosticsGPU.py,sha256=SMJKsCttbI5feOIQC7w7eCe3kpxw1RzHBsWy-OtQh2M,19996
 m3Drop/normalizationCPU.py,sha256=4ulCrDZZjxVFh2y0i4ayPkNCsZYaOP-Xq2Dnzu9WXtg,5697
-m3Drop/normalizationGPU.py,sha256=0tJhmKKdKFXsw9rhnixGH2OuS6UYmQIjndYJAcj9Bvg,8263
-m3drop-0.4.33.dist-info/licenses/LICENSE,sha256=44Iqpp8Fc10Xzd5T7cT9UhO31Qftk3gBiCjtpwilP_k,1074
-m3drop-0.4.33.dist-info/METADATA,sha256=wAXS5oS_8FFRZY-myi-JSBOVcCksFemff4E9xsP4fRM,5161
-m3drop-0.4.33.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-m3drop-0.4.33.dist-info/top_level.txt,sha256=AEULFEFIgFtAwS-KBlIFoYXrqczX_rwqrEcdK46GIrA,7
-m3drop-0.4.33.dist-info/RECORD,,
+m3Drop/normalizationGPU.py,sha256=mHu_Or4ma6qzujGQQQ0oN3D-yoEngLAN4UTknkArRAY,8596
+m3drop-0.4.34.dist-info/licenses/LICENSE,sha256=44Iqpp8Fc10Xzd5T7cT9UhO31Qftk3gBiCjtpwilP_k,1074
+m3drop-0.4.34.dist-info/METADATA,sha256=gAefE3nHHVo4UzTL-0hbMiJTuOcFOxs75kQZ6-d7bJ8,5161
+m3drop-0.4.34.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+m3drop-0.4.34.dist-info/top_level.txt,sha256=AEULFEFIgFtAwS-KBlIFoYXrqczX_rwqrEcdK46GIrA,7
+m3drop-0.4.34.dist-info/RECORD,,