ggml: fix review comments

* drop vxe feature
* add nnpa feature

Signed-off-by: Aaron Teo <[email protected]>

Author: taronaeo

ggml/src/CMakeLists.txt

@@ -308,6 +308,10 @@ function(ggml_add_cpu_backend_variant tag_name)
             set(GGML_INTERNAL_${feat} ON)
         endforeach()
     elseif (GGML_SYSTEM_ARCH STREQUAL "s390x")
+        foreach (feat VXE2 NNPA)
+            set(GGML_INTERNAL_${feat} OFF)
+        endforeach()
+
         foreach (feat ${ARGN})
             set(GGML_INTERNAL_${feat} ON)
         endforeach()
@@ -377,8 +381,8 @@ if (GGML_CPU_ALL_VARIANTS)
         endif()
     elseif (GGML_SYSTEM_ARCH STREQUAL "s390x")
         if (CMAKE_SYSTEM_NAME MATCHES "Linux")
-            ggml_add_cpu_backend_variant(z15    Z15 VXE)
-            ggml_add_cpu_backend_variant(z16    Z16 VXE)
+            ggml_add_cpu_backend_variant(z15    Z15 VXE2)
+            ggml_add_cpu_backend_variant(z16    Z16 VXE2 NNPA)
         else()
             message(FATAL_ERROR "Unsupported s390x target OS: ${CMAKE_SYSTEM_NAME}")
         endif()

ggml/src/ggml-cpu/CMakeLists.txt

@@ -504,10 +504,15 @@ function(ggml_add_cpu_backend_variant_impl tag_name)
             endforeach()
         endif()
 
-        if (GGML_VXE OR GGML_INTERNAL_VXE)
-            message(STATUS "VX/VXE/VXE2 enabled")
+        if (GGML_VXE OR GGML_INTERNAL_VXE2)
+            message(STATUS "VXE2 enabled")
             list(APPEND ARCH_FLAGS -mvx -mzvector)
-            list(APPEND ARCH_DEFINITIONS GGML_VXE)
+            list(APPEND ARCH_DEFINITIONS GGML_USE_VXE2)
+        endif()
+
+        if (GGML_INTERNAL_NNPA)
+            message(STATUS "NNPA enabled")
+            list(APPEND ARCH_DEFINITIONS GGML_USE_NNPA)
         endif()
 
         ggml_add_cpu_backend_features(${GGML_CPU_NAME} s390 ${ARCH_DEFINITIONS})

ggml/src/ggml-cpu/arch/s390/cpu-feats.cpp

@@ -4,10 +4,6 @@
 #include <sys/auxv.h>
 
 // find hwcap bits in asm/elf.h
-#ifndef HWCAP_VXRS_EXT
-#define HWCAP_VXRS_EXT (1 << 13)
-#endif
-
 #ifndef HWCAP_VXRS_EXT2
 #define HWCAP_VXRS_EXT2 (1 << 15)
 #endif
@@ -17,7 +13,6 @@
 #endif
 
 struct s390x_features {
-    bool has_vxe  = false;
     bool has_vxe2 = false;
     bool has_nnpa = false;
 
@@ -26,7 +21,6 @@ struct s390x_features {
         // NOTE: use hwcap2 with DFLT for z17 and later
         // uint32_t hwcap2 = getauxval(AT_HWCAP2);
 
-        has_vxe  = !!(hwcap & HWCAP_VXRS_EXT);
         has_vxe2 = !!(hwcap & HWCAP_VXRS_EXT2);
         has_nnpa = !!(hwcap & HWCAP_NNPA);
     }
@@ -36,17 +30,16 @@ static int ggml_backend_cpu_s390x_score() {
     int score = 1;
     s390x_features sf;
 
-#ifdef GGML_USE_VXE
-    if (!sf.has_vxe) { return 0; }
-    score += 1 << 1;
-#endif
+// IBM z15 / LinuxONE 3
 #ifdef GGML_USE_VXE2
     if (!sf.has_vxe2) { return 0; }
-    score += 1 << 2;
+    score += 1 << 1;
 #endif
+
+// IBM z16 / LinuxONE 4 and z17 / LinuxONE 5
 #ifdef GGML_USE_NNPA
     if (!sf.has_nnpa) { return 0; }
-    score += 1 << 3;
+    score += 1 << 2;
 #endif
 
     return score;