set DDR size in planner to 0 (#3052)

Summary:
Pull Request resolved: #3052

DDR memory for KV ZCH (Virtual Tables) is stilll WiP.
We are setting it to 0 temporarily while we determine the right way to estimate it

Differential Revision: D76146650

Author: kausv

torchrec/distributed/planner/shard_estimators.py

@@ -1245,7 +1245,8 @@ def calculate_shard_storages(
             for _ in hbm_specific_sizes
         ]
         ddr_specific_sizes = [
-            (key_value_params.l2_cache_size or 0) * 1024 * 1024 * 1024
+            # TODO: revisit the logic for SSD virtual table
+            0
             for _ in ddr_specific_sizes
         ]
 

torchrec/distributed/planner/tests/test_planners.py

@@ -618,7 +618,7 @@ def test_planner_with_virtual_table(self) -> None:
         # L1 cache size is 64GB per shard and L2 cache size is 128MB per shard per table
         self.assertTrue(
             any(
-                "dram_virtual_table: HBM: 0.501 GB, DDR: 256.0 GB" in line
+                "dram_virtual_table: HBM: 0.501 GB, DDR: 0.0 GB" in line
                 for line in stats
             )
         )
@@ -748,7 +748,7 @@ def test_planner_with_virtual_table(self) -> None:
         # L2 cache size is 128MB per shard per table
         self.assertTrue(
             any(
-                "dram_virtual_table: HBM: 0.002 GB, DDR: 256.0 GB" in line
+                "dram_virtual_table: HBM: 0.002 GB, DDR: 0.0 GB" in line
                 for line in stats
             )
         )
@@ -800,7 +800,7 @@ def test_planner_with_virtual_table(self) -> None:
         # L2 cache size is 128MB per shard per table
         self.assertTrue(
             any(
-                "dram_virtual_table: HBM: 0.005 GB, DDR: 256.0 GB" in line
+                "dram_virtual_table: HBM: 0.005 GB, DDR: 0.0 GB" in line
                 for line in stats
             )
         )