gpu-compute: fix GPU TLB outstandingReqs vs. associativity
The GPU TLB maxOutstandingReqs field gets limited by the associativity. In the current setup, this means that the max outstanding requests is 32 even though the setup is for 64 entries. Update the associativity to all 64 entries. Change-Id: I2104e4647d97bf4d1cf5ac447e38ad6ac6a1a0d8
This commit is contained in:
@@ -44,8 +44,8 @@ def tlb_options(parser):
|
||||
# L1 TLB Options (D-TLB, I-TLB, Dispatcher-TLB)
|
||||
# ===================================================================
|
||||
|
||||
parser.add_argument("--L1TLBentries", type=int, default="32")
|
||||
parser.add_argument("--L1TLBassoc", type=int, default="32")
|
||||
parser.add_argument("--L1TLBentries", type=int, default="64")
|
||||
parser.add_argument("--L1TLBassoc", type=int, default="64")
|
||||
parser.add_argument(
|
||||
"--L1AccessLatency",
|
||||
type=int,
|
||||
@@ -68,7 +68,7 @@ def tlb_options(parser):
|
||||
# ===================================================================
|
||||
|
||||
parser.add_argument("--L2TLBentries", type=int, default="4096")
|
||||
parser.add_argument("--L2TLBassoc", type=int, default="32")
|
||||
parser.add_argument("--L2TLBassoc", type=int, default="64")
|
||||
parser.add_argument(
|
||||
"--L2AccessLatency",
|
||||
type=int,
|
||||
@@ -90,7 +90,7 @@ def tlb_options(parser):
|
||||
# ===================================================================
|
||||
|
||||
parser.add_argument("--L3TLBentries", type=int, default="8192")
|
||||
parser.add_argument("--L3TLBassoc", type=int, default="32")
|
||||
parser.add_argument("--L3TLBassoc", type=int, default="64")
|
||||
parser.add_argument(
|
||||
"--L3AccessLatency",
|
||||
type=int,
|
||||
|
||||
Reference in New Issue
Block a user