mirror of
https://github.com/CNugteren/CLBlast.git
synced 2024-07-07 12:23:46 +02:00
Updated tuning database for reduction/dot kernels based on the new tuner; partially repopulated the database
This commit is contained in:
parent
b9317d7d0c
commit
c94b628318
|
@ -18,54 +18,36 @@ const Database::DatabaseEntry Database::XdotSingle = {
|
|||
"Xdot", Precision::kSingle, {
|
||||
{ // AMD GPUs
|
||||
kDeviceTypeGPU, "AMD", {
|
||||
{ "AMD Radeon R9 M370X Compute Engine", { {"VW",1}, {"WGS1",64}, {"WGS2",128} } },
|
||||
{ "Tahiti", { {"VW",1}, {"WGS1",256}, {"WGS2",256} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",128} } },
|
||||
}
|
||||
},
|
||||
{ // ARM GPUs
|
||||
kDeviceTypeGPU, "ARM", {
|
||||
{ "Mali-T628", { {"VW",1}, {"WGS1",128}, {"WGS2",256} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",128}, {"WGS2",256} } },
|
||||
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",128}, {"WGS2",32} } },
|
||||
{ "Tahiti", { {"WGS1",128}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",128}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // Intel CPUs
|
||||
kDeviceTypeCPU, "Intel", {
|
||||
{ "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"VW",1}, {"WGS1",512}, {"WGS2",1024} } },
|
||||
{ "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz", { {"VW",1}, {"WGS1",512}, {"WGS2",1024} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"WGS1",1024}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",1024}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // Intel GPUs
|
||||
kDeviceTypeGPU, "Intel", {
|
||||
{ "Iris", { {"VW",1}, {"WGS1",512}, {"WGS2",32} } },
|
||||
{ "Iris Pro", { {"VW",1}, {"WGS1",128}, {"WGS2",512} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",128}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // Intel accelerators
|
||||
kDeviceTypeAccelerator, "Intel", {
|
||||
{ "Intel(R) Many Integrated Core Acceleration Card", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "Iris Pro", { {"WGS1",512}, {"WGS2",64} } },
|
||||
{ "default", { {"WGS1",512}, {"WGS2",64} } },
|
||||
}
|
||||
},
|
||||
{ // NVIDIA GPUs
|
||||
kDeviceTypeGPU, "NVIDIA", {
|
||||
{ "GeForce GTX 480", { {"VW",1}, {"WGS1",256}, {"WGS2",128} } },
|
||||
{ "GeForce GTX 680", { {"VW",1}, {"WGS1",128}, {"WGS2",128} } },
|
||||
{ "GeForce GTX 750 Ti", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "GeForce GTX 980", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "GeForce GTX TITAN", { {"VW",1}, {"WGS1",128}, {"WGS2",1024} } },
|
||||
{ "GeForce GTX TITAN X", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "Tesla K20m", { {"VW",1}, {"WGS1",128}, {"WGS2",1024} } },
|
||||
{ "Tesla K40m", { {"VW",1}, {"WGS1",128}, {"WGS2",1024} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",128} } },
|
||||
{ "GeForce GTX 480", { {"WGS1",512}, {"WGS2",32} } },
|
||||
{ "GeForce GTX 680", { {"WGS1",128}, {"WGS2",128} } },
|
||||
{ "GeForce GTX 980", { {"WGS1",256}, {"WGS2",32} } },
|
||||
{ "GeForce GTX TITAN X", { {"WGS1",256}, {"WGS2",32} } },
|
||||
{ "Tesla K20m", { {"WGS1",1024}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",128}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // Default
|
||||
kDeviceTypeAll, "default", {
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",128}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
}
|
||||
|
@ -77,54 +59,36 @@ const Database::DatabaseEntry Database::XdotComplexSingle = {
|
|||
"Xdot", Precision::kComplexSingle, {
|
||||
{ // AMD GPUs
|
||||
kDeviceTypeGPU, "AMD", {
|
||||
{ "AMD Radeon R9 M370X Compute Engine", { {"VW",1}, {"WGS1",64}, {"WGS2",32} } },
|
||||
{ "Tahiti", { {"VW",1}, {"WGS1",64}, {"WGS2",256} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // ARM GPUs
|
||||
kDeviceTypeGPU, "ARM", {
|
||||
{ "Mali-T628", { {"VW",1}, {"WGS1",128}, {"WGS2",512} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",128}, {"WGS2",512} } },
|
||||
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WGS2",32} } },
|
||||
{ "Tahiti", { {"WGS1",64}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",64}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // Intel CPUs
|
||||
kDeviceTypeCPU, "Intel", {
|
||||
{ "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"VW",1}, {"WGS1",256}, {"WGS2",1024} } },
|
||||
{ "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz", { {"VW",1}, {"WGS1",512}, {"WGS2",1024} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"WGS1",1024}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",1024}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // Intel GPUs
|
||||
kDeviceTypeGPU, "Intel", {
|
||||
{ "Iris", { {"VW",1}, {"WGS1",512}, {"WGS2",1024} } },
|
||||
{ "Iris Pro", { {"VW",1}, {"WGS1",256}, {"WGS2",1024} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",256}, {"WGS2",1024} } },
|
||||
}
|
||||
},
|
||||
{ // Intel accelerators
|
||||
kDeviceTypeAccelerator, "Intel", {
|
||||
{ "Intel(R) Many Integrated Core Acceleration Card", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "Iris Pro", { {"WGS1",32}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",32}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // NVIDIA GPUs
|
||||
kDeviceTypeGPU, "NVIDIA", {
|
||||
{ "GeForce GTX 480", { {"VW",1}, {"WGS1",512}, {"WGS2",512} } },
|
||||
{ "GeForce GTX 680", { {"VW",1}, {"WGS1",256}, {"WGS2",32} } },
|
||||
{ "GeForce GTX 750 Ti", { {"VW",1}, {"WGS1",128}, {"WGS2",32} } },
|
||||
{ "GeForce GTX 980", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "GeForce GTX TITAN", { {"VW",1}, {"WGS1",128}, {"WGS2",1024} } },
|
||||
{ "GeForce GTX TITAN X", { {"VW",1}, {"WGS1",128}, {"WGS2",1024} } },
|
||||
{ "Tesla K20m", { {"VW",1}, {"WGS1",256}, {"WGS2",512} } },
|
||||
{ "Tesla K40m", { {"VW",1}, {"WGS1",128}, {"WGS2",1024} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",32} } },
|
||||
{ "GeForce GTX 480", { {"WGS1",512}, {"WGS2",32} } },
|
||||
{ "GeForce GTX 680", { {"WGS1",128}, {"WGS2",64} } },
|
||||
{ "GeForce GTX 980", { {"WGS1",256}, {"WGS2",64} } },
|
||||
{ "GeForce GTX TITAN X", { {"WGS1",256}, {"WGS2",32} } },
|
||||
{ "Tesla K20m", { {"WGS1",512}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",128}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // Default
|
||||
kDeviceTypeAll, "default", {
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",32}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
}
|
||||
|
@ -136,47 +100,30 @@ const Database::DatabaseEntry Database::XdotDouble = {
|
|||
"Xdot", Precision::kDouble, {
|
||||
{ // AMD GPUs
|
||||
kDeviceTypeGPU, "AMD", {
|
||||
{ "AMD Radeon R9 M370X Compute Engine", { {"VW",1}, {"WGS1",64}, {"WGS2",128} } },
|
||||
{ "Tahiti", { {"VW",1}, {"WGS1",64}, {"WGS2",256} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",128} } },
|
||||
}
|
||||
},
|
||||
{ // ARM GPUs
|
||||
kDeviceTypeGPU, "ARM", {
|
||||
{ "Mali-T628", { {"VW",1}, {"WGS1",64}, {"WGS2",512} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",512} } },
|
||||
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WGS2",128} } },
|
||||
{ "Tahiti", { {"WGS1",256}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",64}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // Intel CPUs
|
||||
kDeviceTypeCPU, "Intel", {
|
||||
{ "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"VW",1}, {"WGS1",512}, {"WGS2",512} } },
|
||||
{ "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz", { {"VW",1}, {"WGS1",1024}, {"WGS2",512} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",512} } },
|
||||
}
|
||||
},
|
||||
{ // Intel accelerators
|
||||
kDeviceTypeAccelerator, "Intel", {
|
||||
{ "Intel(R) Many Integrated Core Acceleration Card", { {"VW",1}, {"WGS1",512}, {"WGS2",1024} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",512}, {"WGS2",1024} } },
|
||||
{ "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"WGS1",512}, {"WGS2",64} } },
|
||||
{ "default", { {"WGS1",512}, {"WGS2",64} } },
|
||||
}
|
||||
},
|
||||
{ // NVIDIA GPUs
|
||||
kDeviceTypeGPU, "NVIDIA", {
|
||||
{ "GeForce GTX 480", { {"VW",1}, {"WGS1",512}, {"WGS2",1024} } },
|
||||
{ "GeForce GTX 680", { {"VW",1}, {"WGS1",64}, {"WGS2",128} } },
|
||||
{ "GeForce GTX 750 Ti", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "GeForce GTX 980", { {"VW",1}, {"WGS1",32}, {"WGS2",512} } },
|
||||
{ "GeForce GTX TITAN", { {"VW",1}, {"WGS1",128}, {"WGS2",1024} } },
|
||||
{ "GeForce GTX TITAN X", { {"VW",1}, {"WGS1",128}, {"WGS2",128} } },
|
||||
{ "Tesla K20m", { {"VW",1}, {"WGS1",128}, {"WGS2",1024} } },
|
||||
{ "Tesla K40m", { {"VW",1}, {"WGS1",256}, {"WGS2",1024} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",32}, {"WGS2",128} } },
|
||||
{ "GeForce GTX 480", { {"WGS1",512}, {"WGS2",32} } },
|
||||
{ "GeForce GTX 680", { {"WGS1",128}, {"WGS2",64} } },
|
||||
{ "GeForce GTX 980", { {"WGS1",128}, {"WGS2",32} } },
|
||||
{ "GeForce GTX TITAN X", { {"WGS1",256}, {"WGS2",32} } },
|
||||
{ "Tesla K20m", { {"WGS1",512}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",128}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // Default
|
||||
kDeviceTypeAll, "default", {
|
||||
{ "default", { {"VW",1}, {"WGS1",32}, {"WGS2",128} } },
|
||||
{ "default", { {"WGS1",64}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
}
|
||||
|
@ -188,47 +135,30 @@ const Database::DatabaseEntry Database::XdotComplexDouble = {
|
|||
"Xdot", Precision::kComplexDouble, {
|
||||
{ // AMD GPUs
|
||||
kDeviceTypeGPU, "AMD", {
|
||||
{ "AMD Radeon R9 M370X Compute Engine", { {"VW",1}, {"WGS1",64}, {"WGS2",32} } },
|
||||
{ "Tahiti", { {"VW",1}, {"WGS1",64}, {"WGS2",256} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // ARM GPUs
|
||||
kDeviceTypeGPU, "ARM", {
|
||||
{ "Mali-T628", { {"VW",1}, {"WGS1",32}, {"WGS2",64} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",32}, {"WGS2",64} } },
|
||||
{ "AMD Radeon R9 M370X Compute Engine", { {"WGS1",64}, {"WGS2",32} } },
|
||||
{ "Tahiti", { {"WGS1",256}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",64}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // Intel CPUs
|
||||
kDeviceTypeCPU, "Intel", {
|
||||
{ "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"VW",1}, {"WGS1",256}, {"WGS2",1024} } },
|
||||
{ "Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
{ "Intel(R) Core(TM) i7-5930K CPU @ 3.50GHz", { {"VW",1}, {"WGS1",512}, {"WGS2",1024} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",64}, {"WGS2",1024} } },
|
||||
}
|
||||
},
|
||||
{ // Intel accelerators
|
||||
kDeviceTypeAccelerator, "Intel", {
|
||||
{ "Intel(R) Many Integrated Core Acceleration Card", { {"VW",1}, {"WGS1",32}, {"WGS2",1024} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",32}, {"WGS2",1024} } },
|
||||
{ "Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz", { {"WGS1",1024}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",1024}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // NVIDIA GPUs
|
||||
kDeviceTypeGPU, "NVIDIA", {
|
||||
{ "GeForce GTX 480", { {"VW",1}, {"WGS1",512}, {"WGS2",512} } },
|
||||
{ "GeForce GTX 680", { {"VW",1}, {"WGS1",256}, {"WGS2",64} } },
|
||||
{ "GeForce GTX 750 Ti", { {"VW",1}, {"WGS1",32}, {"WGS2",64} } },
|
||||
{ "GeForce GTX 980", { {"VW",1}, {"WGS1",32}, {"WGS2",128} } },
|
||||
{ "GeForce GTX TITAN", { {"VW",1}, {"WGS1",128}, {"WGS2",512} } },
|
||||
{ "GeForce GTX TITAN X", { {"VW",1}, {"WGS1",128}, {"WGS2",128} } },
|
||||
{ "Tesla K20m", { {"VW",1}, {"WGS1",128}, {"WGS2",1024} } },
|
||||
{ "Tesla K40m", { {"VW",1}, {"WGS1",128}, {"WGS2",1024} } },
|
||||
{ "default", { {"VW",1}, {"WGS1",32}, {"WGS2",64} } },
|
||||
{ "GeForce GTX 480", { {"WGS1",512}, {"WGS2",32} } },
|
||||
{ "GeForce GTX 680", { {"WGS1",256}, {"WGS2",64} } },
|
||||
{ "GeForce GTX 980", { {"WGS1",64}, {"WGS2",32} } },
|
||||
{ "GeForce GTX TITAN X", { {"WGS1",128}, {"WGS2",32} } },
|
||||
{ "Tesla K20m", { {"WGS1",128}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",64}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
{ // Default
|
||||
kDeviceTypeAll, "default", {
|
||||
{ "default", { {"VW",1}, {"WGS1",32}, {"WGS2",32} } },
|
||||
{ "default", { {"WGS1",64}, {"WGS2",32} } },
|
||||
}
|
||||
},
|
||||
}
|
||||
|
|
|
@ -95,6 +95,9 @@ def RemoveDuplicates(df):
|
|||
def RemoveEntriesByDevice(df, devicename):
|
||||
return df[df["device"] != devicename]
|
||||
|
||||
def RemoveEntriesByKernelFamily(df, familyname):
|
||||
return df[df["kernel_family"] != familyname]
|
||||
|
||||
def GetEntriesByField(df, field, value):
|
||||
return df[df[field] == value]
|
||||
|
||||
|
|
Loading…
Reference in a new issue