Temporary workaround for RMM initial pool size bug (NVIDIA#530)

Signed-off-by: Jason Lowe <[email protected]>
nartal1 · Aug 7, 2020 · 97f2e7b · 97f2e7b
1 parent 101b49b
commit 97f2e7b
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/docs/configs.md b/docs/configs.md
@@ -29,7 +29,7 @@ scala> spark.conf.set("spark.rapids.sql.incompatibleOps.enabled", true)
 
 Name | Description | Default Value
 -----|-------------|--------------
-<a name="memory.gpu.allocFraction"></a>spark.rapids.memory.gpu.allocFraction|The fraction of total GPU memory that should be initially allocated for pooled memory. Extra memory will be allocated as needed, but it may result in more fragmentation.|0.9
+<a name="memory.gpu.allocFraction"></a>spark.rapids.memory.gpu.allocFraction|The fraction of total GPU memory that should be initially allocated for pooled memory. Extra memory will be allocated as needed, but it may result in more fragmentation.|0.5
 <a name="memory.gpu.debug"></a>spark.rapids.memory.gpu.debug|Provides a log of GPU memory allocations and frees. If set to STDOUT or STDERR the logging will go there. Setting it to NONE disables logging. All other values are reserved for possible future expansion and in the mean time will disable logging.|NONE
 <a name="memory.gpu.pooling.enabled"></a>spark.rapids.memory.gpu.pooling.enabled|Should RMM act as a pooling allocator for GPU memory, or should it just pass through to CUDA memory allocation directly.|true
 <a name="memory.host.spillStorageSize"></a>spark.rapids.memory.host.spillStorageSize|Amount of off-heap host memory to use for buffering spilled GPU data before spilling to local disk|1073741824

diff --git a/sql-plugin/src/main/scala/com/nvidia/spark/rapids/RapidsConf.scala b/sql-plugin/src/main/scala/com/nvidia/spark/rapids/RapidsConf.scala
@@ -258,7 +258,7 @@ object RapidsConf {
       "result in more fragmentation.")
     .doubleConf
     .checkValue(v => v >= 0 && v <= 1, "The fraction value must be in [0, 1].")
-    .createWithDefault(0.9)
+    .createWithDefault(0.5)
 
   val HOST_SPILL_STORAGE_SIZE = conf("spark.rapids.memory.host.spillStorageSize")
     .doc("Amount of off-heap host memory to use for buffering spilled GPU data " +