Extract amdgcn_cc from kfd

Author: zerefwayne

init/eessi_archdetect.sh

@@ -212,24 +212,70 @@ accelpath() {
     # fi
 
     # check for AMD GPUs via amd-smi command
-    amd_smi=$(command -v amd-smi)
-    if [[ $? -eq 0 ]]; then
-        log "DEBUG" "accelpath: amd-smi command found @ ${amd_smi}"
-        amd_smi_out=$(mktemp -p /tmp amd_smi_out.XXXXX)
-        amd-smi static --asic | grep TARGET_GRAPHICS_VERSION 2>&1 > $amd_smi_out
-        if [[ $? -eq 0 ]]; then
-            amd_smi_info=$(head -n 1 $amd_smi_out)
-            amdgcn_cc=$(echo $amd_smi_info | sed 's/.*: //')
-            log "DEBUG" "accelpath: AMDGCN compute capability '${amdgcn_cc}' derived from amd-smi output '${amd_smi_info}'"
+    # amd_smi=$(command -v amd-smi)
+    # if [[ $? -eq 0 ]]; then
+    #     log "DEBUG" "accelpath: amd-smi command found @ ${amd_smi}"
+    #     amd_smi_out=$(mktemp -p /tmp amd_smi_out.XXXXX)
+    #     amd-smi static --asic | grep TARGET_GRAPHICS_VERSION 2>&1 > $amd_smi_out
+    #     if [[ $? -eq 0 ]]; then
+    #         amd_smi_info=$(head -n 1 $amd_smi_out)
+    #         amdgcn_cc=$(echo $amd_smi_info | sed 's/.*: //')
+    #         log "DEBUG" "accelpath: AMDGCN compute capability '${amdgcn_cc}' derived from amd-smi output '${amd_smi_info}'"
+    #         res="accel/amd/${amdgcn_cc}"
+    #         echo $res
+    #         rm -f $amd_smi_out
+    #     else
+    #         log "DEBUG" "accelpath: ami-smi command failed, see output in $amd_smi_out"
+    #         exit 3
+    #     fi
+    # else
+    #     log "DEBUG" "accelpath: amd-smi command not found"
+    #     exit 2
+    # fi
+
+    # logic ported from https://github.com/llvm/llvm-project/blob/6e738e187055bbd33b6c3d203b6b55904dfcb624/clang/tools/offload-arch/AMDGPUArchByKFD.cpp
+    # check for AMD GPUs via KFD sysfs interface (No amd-smi or Python required)
+    kfd_nodes="/sys/devices/virtual/kfd/kfd/topology/nodes"
+
+    if [[ -d "$kfd_nodes" ]]; then
+        log "DEBUG" "accelpath: KFD sysfs path found @ ${kfd_nodes}"
+        amdgcn_cc=""
+
+        # ls -1v ensures numeric/version sorting (nodes/0, nodes/1, ..., nodes/10)
+        # just like LLVM's llvm::sort by node ID.
+        # Assuming homogeneous nodes for EESSI, grab the first valid GPU and break
+        for node in $(ls -1v "$kfd_nodes" 2>/dev/null); do
+            prop_file="$kfd_nodes/$node/properties"
+            
+            if [[ -f "$prop_file" ]]; then
+                # Extract the integer value. 2>/dev/null suppresses read errors.
+                gfx_ver=$(grep "^gfx_target_version" "$prop_file" 2>/dev/null | awk '{print $2}')
+                
+                # If gfx_ver is non-empty and greater than 0 (0 means it's a CPU node)
+                if [[ -n "$gfx_ver" && "$gfx_ver" -gt 0 ]]; then
+                    # Perform the exact math from AMDGPUArchByKFD.cpp
+                    major=$(( (gfx_ver / 10000) % 100 ))
+                    minor=$(( (gfx_ver / 100) % 100 ))
+                    step=$(( gfx_ver % 100 ))
+                    
+                    # Format as gfx<major><minor><hex_step> (e.g., 9 0 a -> gfx90a)
+                    amdgcn_cc=$(printf "gfx%d%d%x" $major $minor $step)
+                    
+                    log "DEBUG" "accelpath: AMDGCN compute capability '${amdgcn_cc}' derived from KFD node ${node}"
+                    break 
+                fi
+            fi
+        done
+
+        if [[ -n "$amdgcn_cc" ]]; then
             res="accel/amd/${amdgcn_cc}"
-            echo $res
-            rm -f $amd_smi_out
+            echo "$res"
         else
-            log "DEBUG" "accelpath: ami-smi command failed, see output in $amd_smi_out"
+            log "DEBUG" "accelpath: KFD topology found, but no AMD GPUs detected (only CPUs)"
             exit 3
         fi
     else
-        log "DEBUG" "accelpath: amd-smi command not found"
+        log "DEBUG" "accelpath: KFD sysfs path not found. AMD GPU driver not loaded?"
         exit 2
     fi
 }