LOADDUP variant of c2 microkernel

- instead of 1 LD1 and 4 DUP, use 1 LD4R

PiperOrigin-RevId: 410282504
diff --git a/CMakeLists.txt b/CMakeLists.txt
index 2ab805a..809f82e 100755
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -1508,24 +1508,28 @@
   src/qc8-dwconv/gen/up32x9-minmax-fp32-neon-mul16.c
   src/qc8-dwconv/gen/up32x25-minmax-fp32-neon-mul16.c
   src/qc8-gemm/gen/1x8c2-minmax-fp32-neon-mlal-dup.c
+  src/qc8-gemm/gen/1x8c2-minmax-fp32-neon-mlal-ld4r.c
   src/qc8-gemm/gen/1x8c2s4-minmax-fp32-neon-mlal.c
   src/qc8-gemm/gen/1x8c4-minmax-fp32-neon-mlal-dup.c
   src/qc8-gemm/gen/1x8c4s2-minmax-fp32-neon-mlal.c
   src/qc8-gemm/gen/1x8c8-minmax-fp32-neon-mlal.c
   src/qc8-gemm/gen/1x16-minmax-fp32-neon-mlal-lane.c
   src/qc8-gemm/gen/2x8c2-minmax-fp32-neon-mlal-dup.c
+  src/qc8-gemm/gen/2x8c2-minmax-fp32-neon-mlal-ld4r.c
   src/qc8-gemm/gen/2x8c2s4-minmax-fp32-neon-mlal.c
   src/qc8-gemm/gen/2x8c4-minmax-fp32-neon-mlal-dup.c
   src/qc8-gemm/gen/2x8c4s2-minmax-fp32-neon-mlal.c
   src/qc8-gemm/gen/2x8c8-minmax-fp32-neon-mlal.c
   src/qc8-gemm/gen/4x16-minmax-fp32-neon-mlal-lane.c
   src/qc8-igemm/gen/1x8c2-minmax-fp32-neon-mlal-dup.c
+  src/qc8-igemm/gen/1x8c2-minmax-fp32-neon-mlal-ld4r.c
   src/qc8-igemm/gen/1x8c2s4-minmax-fp32-neon-mlal.c
   src/qc8-igemm/gen/1x8c4-minmax-fp32-neon-mlal-dup.c
   src/qc8-igemm/gen/1x8c4s2-minmax-fp32-neon-mlal.c
   src/qc8-igemm/gen/1x8c8-minmax-fp32-neon-mlal.c
   src/qc8-igemm/gen/1x16-minmax-fp32-neon-mlal-lane.c
   src/qc8-igemm/gen/2x8c2-minmax-fp32-neon-mlal-dup.c
+  src/qc8-igemm/gen/2x8c2-minmax-fp32-neon-mlal-ld4r.c
   src/qc8-igemm/gen/2x8c2s4-minmax-fp32-neon-mlal.c
   src/qc8-igemm/gen/2x8c4-minmax-fp32-neon-mlal-dup.c
   src/qc8-igemm/gen/2x8c4s2-minmax-fp32-neon-mlal.c
@@ -1579,9 +1583,12 @@
   src/qs8-gemm/gen/1x8-minmax-gemmlowp-neon-mull-addw-dup.c
   src/qs8-gemm/gen/1x8-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-gemm/gen/1x8c2-minmax-fp32-neon-mlal-dup.c
+  src/qs8-gemm/gen/1x8c2-minmax-fp32-neon-mlal-ld4r.c
   src/qs8-gemm/gen/1x8c2-minmax-gemmlowp-neon-mlal-dup.c
   src/qs8-gemm/gen/1x8c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-gemm/gen/1x8c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-gemm/gen/1x8c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-gemm/gen/1x8c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-gemm/gen/1x8c2s4-minmax-fp32-neon-mlal.c
   src/qs8-gemm/gen/1x8c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-gemm/gen/1x8c2s4-minmax-rndnu-neon-mull.c
@@ -1602,7 +1609,9 @@
   src/qs8-gemm/gen/1x16-minmax-rndnu-neon-mlal-lane.c
   src/qs8-gemm/gen/1x16-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-gemm/gen/1x16c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-gemm/gen/1x16c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-gemm/gen/1x16c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-gemm/gen/1x16c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-gemm/gen/1x16c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-gemm/gen/1x16c2s4-minmax-rndnu-neon-mull.c
   src/qs8-gemm/gen/1x16c4-minmax-rndnu-neon-mlal-dup.c
@@ -1615,9 +1624,12 @@
   src/qs8-gemm/gen/2x8-minmax-gemmlowp-neon-mlal-lane.c
   src/qs8-gemm/gen/2x8-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-gemm/gen/2x8c2-minmax-fp32-neon-mlal-dup.c
+  src/qs8-gemm/gen/2x8c2-minmax-fp32-neon-mlal-ld4r.c
   src/qs8-gemm/gen/2x8c2-minmax-gemmlowp-neon-mlal-dup.c
   src/qs8-gemm/gen/2x8c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-gemm/gen/2x8c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-gemm/gen/2x8c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-gemm/gen/2x8c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-gemm/gen/2x8c2s4-minmax-fp32-neon-mlal.c
   src/qs8-gemm/gen/2x8c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-gemm/gen/2x8c2s4-minmax-rndnu-neon-mull.c
@@ -1635,7 +1647,9 @@
   src/qs8-gemm/gen/2x16-minmax-gemmlowp-neon-mlal-lane.c
   src/qs8-gemm/gen/2x16-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-gemm/gen/2x16c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-gemm/gen/2x16c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-gemm/gen/2x16c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-gemm/gen/2x16c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-gemm/gen/2x16c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-gemm/gen/2x16c2s4-minmax-rndnu-neon-mull.c
   src/qs8-gemm/gen/2x16c4-minmax-rndnu-neon-mlal-dup.c
@@ -1648,7 +1662,9 @@
   src/qs8-gemm/gen/3x8-minmax-gemmlowp-neon-mlal-lane.c
   src/qs8-gemm/gen/3x8-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-gemm/gen/3x8c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-gemm/gen/3x8c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-gemm/gen/3x8c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-gemm/gen/3x8c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-gemm/gen/3x8c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-gemm/gen/3x8c2s4-minmax-rndnu-neon-mull.c
   src/qs8-gemm/gen/3x8c4-minmax-rndnu-neon-mlal-dup.c
@@ -1661,7 +1677,9 @@
   src/qs8-gemm/gen/3x16-minmax-gemmlowp-neon-mlal-lane.c
   src/qs8-gemm/gen/3x16-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-gemm/gen/3x16c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-gemm/gen/3x16c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-gemm/gen/3x16c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-gemm/gen/3x16c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-gemm/gen/3x16c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-gemm/gen/3x16c2s4-minmax-rndnu-neon-mull.c
   src/qs8-gemm/gen/3x16c4-minmax-rndnu-neon-mlal-dup.c
@@ -1674,7 +1692,9 @@
   src/qs8-gemm/gen/4x8-minmax-gemmlowp-neon-mlal-lane.c
   src/qs8-gemm/gen/4x8-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-gemm/gen/4x8c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-gemm/gen/4x8c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-gemm/gen/4x8c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-gemm/gen/4x8c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-gemm/gen/4x8c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-gemm/gen/4x8c2s4-minmax-rndnu-neon-mull.c
   src/qs8-gemm/gen/4x8c4-minmax-rndnu-neon-mlal-dup.c
@@ -1690,7 +1710,9 @@
   src/qs8-gemm/gen/4x16-minmax-rndnu-neon-mlal-lane.c
   src/qs8-gemm/gen/4x16-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-gemm/gen/4x16c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-gemm/gen/4x16c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-gemm/gen/4x16c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-gemm/gen/4x16c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-gemm/gen/4x16c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-gemm/gen/4x16c2s4-minmax-rndnu-neon-mull.c
   src/qs8-gemm/gen/4x16c4-minmax-rndnu-neon-mlal-dup.c
@@ -1706,9 +1728,12 @@
   src/qs8-igemm/gen/1x8-minmax-gemmlowp-neon-mull-addw-dup.c
   src/qs8-igemm/gen/1x8-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-igemm/gen/1x8c2-minmax-fp32-neon-mlal-dup.c
+  src/qs8-igemm/gen/1x8c2-minmax-fp32-neon-mlal-ld4r.c
   src/qs8-igemm/gen/1x8c2-minmax-gemmlowp-neon-mlal-dup.c
   src/qs8-igemm/gen/1x8c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-igemm/gen/1x8c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-igemm/gen/1x8c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-igemm/gen/1x8c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-igemm/gen/1x8c2s4-minmax-fp32-neon-mlal.c
   src/qs8-igemm/gen/1x8c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-igemm/gen/1x8c2s4-minmax-rndnu-neon-mull.c
@@ -1729,7 +1754,9 @@
   src/qs8-igemm/gen/1x16-minmax-rndnu-neon-mlal-lane.c
   src/qs8-igemm/gen/1x16-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-igemm/gen/1x16c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-igemm/gen/1x16c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-igemm/gen/1x16c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-igemm/gen/1x16c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-igemm/gen/1x16c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-igemm/gen/1x16c2s4-minmax-rndnu-neon-mull.c
   src/qs8-igemm/gen/1x16c4-minmax-rndnu-neon-mlal-dup.c
@@ -1742,9 +1769,12 @@
   src/qs8-igemm/gen/2x8-minmax-gemmlowp-neon-mlal-lane.c
   src/qs8-igemm/gen/2x8-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-igemm/gen/2x8c2-minmax-fp32-neon-mlal-dup.c
+  src/qs8-igemm/gen/2x8c2-minmax-fp32-neon-mlal-ld4r.c
   src/qs8-igemm/gen/2x8c2-minmax-gemmlowp-neon-mlal-dup.c
   src/qs8-igemm/gen/2x8c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-igemm/gen/2x8c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-igemm/gen/2x8c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-igemm/gen/2x8c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-igemm/gen/2x8c2s4-minmax-fp32-neon-mlal.c
   src/qs8-igemm/gen/2x8c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-igemm/gen/2x8c2s4-minmax-rndnu-neon-mull.c
@@ -1762,7 +1792,9 @@
   src/qs8-igemm/gen/2x16-minmax-gemmlowp-neon-mlal-lane.c
   src/qs8-igemm/gen/2x16-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-igemm/gen/2x16c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-igemm/gen/2x16c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-igemm/gen/2x16c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-igemm/gen/2x16c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-igemm/gen/2x16c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-igemm/gen/2x16c2s4-minmax-rndnu-neon-mull.c
   src/qs8-igemm/gen/2x16c4-minmax-rndnu-neon-mlal-dup.c
@@ -1775,7 +1807,9 @@
   src/qs8-igemm/gen/3x8-minmax-gemmlowp-neon-mlal-lane.c
   src/qs8-igemm/gen/3x8-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-igemm/gen/3x8c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-igemm/gen/3x8c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-igemm/gen/3x8c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-igemm/gen/3x8c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-igemm/gen/3x8c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-igemm/gen/3x8c2s4-minmax-rndnu-neon-mull.c
   src/qs8-igemm/gen/3x8c4-minmax-rndnu-neon-mlal-dup.c
@@ -1788,7 +1822,9 @@
   src/qs8-igemm/gen/3x16-minmax-gemmlowp-neon-mlal-lane.c
   src/qs8-igemm/gen/3x16-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-igemm/gen/3x16c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-igemm/gen/3x16c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-igemm/gen/3x16c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-igemm/gen/3x16c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-igemm/gen/3x16c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-igemm/gen/3x16c2s4-minmax-rndnu-neon-mull.c
   src/qs8-igemm/gen/3x16c4-minmax-rndnu-neon-mlal-dup.c
@@ -1801,7 +1837,9 @@
   src/qs8-igemm/gen/4x8-minmax-gemmlowp-neon-mlal-lane.c
   src/qs8-igemm/gen/4x8-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-igemm/gen/4x8c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-igemm/gen/4x8c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-igemm/gen/4x8c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-igemm/gen/4x8c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-igemm/gen/4x8c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-igemm/gen/4x8c2s4-minmax-rndnu-neon-mull.c
   src/qs8-igemm/gen/4x8c4-minmax-rndnu-neon-mlal-dup.c
@@ -1817,7 +1855,9 @@
   src/qs8-igemm/gen/4x16-minmax-rndnu-neon-mlal-lane.c
   src/qs8-igemm/gen/4x16-minmax-rndnu-neon-mull-addw-dup.c
   src/qs8-igemm/gen/4x16c2-minmax-rndnu-neon-mlal-dup.c
+  src/qs8-igemm/gen/4x16c2-minmax-rndnu-neon-mlal-ld4r.c
   src/qs8-igemm/gen/4x16c2-minmax-rndnu-neon-mull-dup.c
+  src/qs8-igemm/gen/4x16c2-minmax-rndnu-neon-mull-ld4r.c
   src/qs8-igemm/gen/4x16c2s4-minmax-rndnu-neon-mlal.c
   src/qs8-igemm/gen/4x16c2s4-minmax-rndnu-neon-mull.c
   src/qs8-igemm/gen/4x16c4-minmax-rndnu-neon-mlal-dup.c
@@ -2376,24 +2416,28 @@
   src/qc8-dwconv/gen/up32x9-minmax-fp32-neonv8-mul16.c
   src/qc8-dwconv/gen/up32x25-minmax-fp32-neonv8-mul16.c
   src/qc8-gemm/gen/1x8c2-minmax-fp32-neonv8-mlal-dup.c
+  src/qc8-gemm/gen/1x8c2-minmax-fp32-neonv8-mlal-ld4r.c
   src/qc8-gemm/gen/1x8c2s4-minmax-fp32-neonv8-mlal.c
   src/qc8-gemm/gen/1x8c4-minmax-fp32-neonv8-mlal-dup.c
   src/qc8-gemm/gen/1x8c4s2-minmax-fp32-neonv8-mlal.c
   src/qc8-gemm/gen/1x8c8-minmax-fp32-neonv8-mlal.c
   src/qc8-gemm/gen/1x16-minmax-fp32-neonv8-mlal-lane.c
   src/qc8-gemm/gen/2x8c2-minmax-fp32-neonv8-mlal-dup.c
+  src/qc8-gemm/gen/2x8c2-minmax-fp32-neonv8-mlal-ld4r.c
   src/qc8-gemm/gen/2x8c2s4-minmax-fp32-neonv8-mlal.c
   src/qc8-gemm/gen/2x8c4-minmax-fp32-neonv8-mlal-dup.c
   src/qc8-gemm/gen/2x8c4s2-minmax-fp32-neonv8-mlal.c
   src/qc8-gemm/gen/2x8c8-minmax-fp32-neonv8-mlal.c
   src/qc8-gemm/gen/4x16-minmax-fp32-neonv8-mlal-lane.c
   src/qc8-igemm/gen/1x8c2-minmax-fp32-neonv8-mlal-dup.c
+  src/qc8-igemm/gen/1x8c2-minmax-fp32-neonv8-mlal-ld4r.c
   src/qc8-igemm/gen/1x8c2s4-minmax-fp32-neonv8-mlal.c
   src/qc8-igemm/gen/1x8c4-minmax-fp32-neonv8-mlal-dup.c
   src/qc8-igemm/gen/1x8c4s2-minmax-fp32-neonv8-mlal.c
   src/qc8-igemm/gen/1x8c8-minmax-fp32-neonv8-mlal.c
   src/qc8-igemm/gen/1x16-minmax-fp32-neonv8-mlal-lane.c
   src/qc8-igemm/gen/2x8c2-minmax-fp32-neonv8-mlal-dup.c
+  src/qc8-igemm/gen/2x8c2-minmax-fp32-neonv8-mlal-ld4r.c
   src/qc8-igemm/gen/2x8c2s4-minmax-fp32-neonv8-mlal.c
   src/qc8-igemm/gen/2x8c4-minmax-fp32-neonv8-mlal-dup.c
   src/qc8-igemm/gen/2x8c4s2-minmax-fp32-neonv8-mlal.c
@@ -2408,24 +2452,28 @@
   src/qs8-dwconv/gen/up32x9-minmax-fp32-neonv8-mul16.c
   src/qs8-dwconv/gen/up32x25-minmax-fp32-neonv8-mul16.c
   src/qs8-gemm/gen/1x8c2-minmax-fp32-neonv8-mlal-dup.c
+  src/qs8-gemm/gen/1x8c2-minmax-fp32-neonv8-mlal-ld4r.c
   src/qs8-gemm/gen/1x8c2s4-minmax-fp32-neonv8-mlal.c
   src/qs8-gemm/gen/1x8c4-minmax-fp32-neonv8-mlal-dup.c
   src/qs8-gemm/gen/1x8c4s2-minmax-fp32-neonv8-mlal.c
   src/qs8-gemm/gen/1x8c8-minmax-fp32-neonv8-mlal.c
   src/qs8-gemm/gen/1x16-minmax-fp32-neonv8-mlal-lane.c
   src/qs8-gemm/gen/2x8c2-minmax-fp32-neonv8-mlal-dup.c
+  src/qs8-gemm/gen/2x8c2-minmax-fp32-neonv8-mlal-ld4r.c
   src/qs8-gemm/gen/2x8c2s4-minmax-fp32-neonv8-mlal.c
   src/qs8-gemm/gen/2x8c4-minmax-fp32-neonv8-mlal-dup.c
   src/qs8-gemm/gen/2x8c4s2-minmax-fp32-neonv8-mlal.c
   src/qs8-gemm/gen/2x8c8-minmax-fp32-neonv8-mlal.c
   src/qs8-gemm/gen/4x16-minmax-fp32-neonv8-mlal-lane.c
   src/qs8-igemm/gen/1x8c2-minmax-fp32-neonv8-mlal-dup.c
+  src/qs8-igemm/gen/1x8c2-minmax-fp32-neonv8-mlal-ld4r.c
   src/qs8-igemm/gen/1x8c2s4-minmax-fp32-neonv8-mlal.c
   src/qs8-igemm/gen/1x8c4-minmax-fp32-neonv8-mlal-dup.c
   src/qs8-igemm/gen/1x8c4s2-minmax-fp32-neonv8-mlal.c
   src/qs8-igemm/gen/1x8c8-minmax-fp32-neonv8-mlal.c
   src/qs8-igemm/gen/1x16-minmax-fp32-neonv8-mlal-lane.c
   src/qs8-igemm/gen/2x8c2-minmax-fp32-neonv8-mlal-dup.c
+  src/qs8-igemm/gen/2x8c2-minmax-fp32-neonv8-mlal-ld4r.c
   src/qs8-igemm/gen/2x8c2s4-minmax-fp32-neonv8-mlal.c
   src/qs8-igemm/gen/2x8c4-minmax-fp32-neonv8-mlal-dup.c
   src/qs8-igemm/gen/2x8c4s2-minmax-fp32-neonv8-mlal.c