Simplify and make simd_util cross-platform
This new algorithm uses a Scalar->Vector->Scalar iteration loop which requires no masking off of any incomplete data chunks. Also, the width was reduced to 32 bytes instead of 64, as I found this to be about as fast as the previous 64-byte x86 version.
This commit is contained in:
5 files changed
+101
-151
No files matched your search
@@ -1,4 +1,3 @@
|
||||
//+build i386, amd64
|
||||
package benchmark_simd_util
|
||||
|
||||
import "core:fmt"
|
||||
|
||||
@@ -1,4 +1,3 @@
|
||||
//+build i386, amd64
|
||||
package test_core_simd_util
|
||||
|
||||
import simd_util "core:simd/util"
|
||||
|
||||
Reference in new issue
Block a user