Add AMXTF32 detection (#154)

cpuid: Add AMXTF32 detection Signed-off-by: Oleg Zhurakivskyy <[email protected]>
klauspost · Feb 24, 2025 · 88f6499 · 88f6499
1 parent 68cbe93
commit 88f6499
Show file tree

Hide file tree

Showing 3 changed files with 223 additions and 219 deletions.
diff --git a/README.md b/README.md
@@ -284,6 +284,7 @@ Exit Code 1
 | AMXFP8             | Tile computational operations on FP8 numbers                                                                                                                                      |
 | AMXCOMPLEX         | Tile computational operations on complex numbers                                                                                                                                   |
 | AMXTILE            | Tile architecture                                                                                                                                                                  |
+| AMXTF32            | Matrix Multiplication of TF32 Tiles into Packed Single Precision Tile                                                                                                              |
 | APX_F              | Intel APX                                                                                                                                                                          |
 | AVX                | AVX functions                                                                                                                                                                      |
 | AVX10              | If set the Intel AVX10 Converged Vector ISA is supported                                                                                                                           |

diff --git a/cpuid.go b/cpuid.go
@@ -83,7 +83,8 @@ const (
 	AMXINT8                              // Tile computational operations on 8-bit integers
 	AMXFP8                               // Tile computational operations on FP8 numbers
 	AMXTILE                              // Tile architecture
-	AMXCOMPLEX                           // Tile computational operations on complex numbers
+	AMXTF32                              // Tile architecture
+	AMXCOMPLEX                           // Matrix Multiplication of TF32 Tiles into Packed Single Precision Tile
 	APX_F                                // Intel APX
 	AVX                                  // AVX functions
 	AVX10                                // If set the Intel AVX10 Converged Vector ISA is supported
@@ -1289,6 +1290,7 @@ func support() flagSet {
 		// CPUID.(EAX=7, ECX=1).EDX
 		fs.setIf(edx1&(1<<4) != 0, AVXVNNIINT8)
 		fs.setIf(edx1&(1<<5) != 0, AVXNECONVERT)
+		fs.setIf(edx1&(1<<7) != 0, AMXTF32)
 		fs.setIf(edx1&(1<<8) != 0, AMXCOMPLEX)
 		fs.setIf(edx1&(1<<10) != 0, AVXVNNIINT16)
 		fs.setIf(edx1&(1<<14) != 0, PREFETCHI)