Refine burn UI and NVIDIA stress flows

iso/builder/build.sh

@@ -302,6 +302,12 @@ memtest_fail() {
     return 0
 }
 
+nvidia_runtime_fail() {
+    msg="$1"
+    echo "ERROR: ${msg}" >&2
+    exit 1
+}
+
 iso_memtest_present() {
     iso_path="$1"
     iso_files="$(mktemp)"
@@ -439,6 +445,44 @@ validate_iso_memtest() {
     echo "=== memtest validation OK ==="
 }
 
+validate_iso_nvidia_runtime() {
+    iso_path="$1"
+    [ "$BEE_GPU_VENDOR" = "nvidia" ] || return 0
+
+    echo "=== validating NVIDIA runtime in ISO ==="
+
+    [ -f "$iso_path" ] || nvidia_runtime_fail "ISO not found for NVIDIA runtime validation: $iso_path"
+    require_iso_reader "$iso_path" >/dev/null 2>&1 || nvidia_runtime_fail "ISO reader unavailable for NVIDIA runtime validation"
+    command -v unsquashfs >/dev/null 2>&1 || nvidia_runtime_fail "unsquashfs is required for NVIDIA runtime validation"
+
+    squashfs_tmp="$(mktemp)"
+    squashfs_list="$(mktemp)"
+    iso_read_member "$iso_path" live/filesystem.squashfs "$squashfs_tmp" || {
+        rm -f "$squashfs_tmp" "$squashfs_list"
+        nvidia_runtime_fail "failed to extract live/filesystem.squashfs from ISO"
+    }
+    unsquashfs -ll "$squashfs_tmp" > "$squashfs_list" 2>/dev/null || {
+        rm -f "$squashfs_tmp" "$squashfs_list"
+        nvidia_runtime_fail "failed to inspect filesystem.squashfs from ISO"
+    }
+
+    grep -Eq 'usr/bin/dcgmi$' "$squashfs_list" || {
+        rm -f "$squashfs_tmp" "$squashfs_list"
+        nvidia_runtime_fail "dcgmi missing from final NVIDIA ISO"
+    }
+    grep -Eq 'usr/bin/nv-hostengine$' "$squashfs_list" || {
+        rm -f "$squashfs_tmp" "$squashfs_list"
+        nvidia_runtime_fail "nv-hostengine missing from final NVIDIA ISO"
+    }
+    grep -Eq 'usr/bin/dcgmproftester([0-9]+)?$' "$squashfs_list" || {
+        rm -f "$squashfs_tmp" "$squashfs_list"
+        nvidia_runtime_fail "dcgmproftester missing from final NVIDIA ISO"
+    }
+
+    rm -f "$squashfs_tmp" "$squashfs_list"
+    echo "=== NVIDIA runtime validation OK ==="
+}
+
 append_memtest_grub_entry() {
     grub_cfg="$1"
     [ -f "$grub_cfg" ] || return 1
@@ -1144,6 +1188,7 @@ if [ -f "$ISO_RAW" ]; then
         fi
     fi
     validate_iso_memtest "$ISO_RAW"
+    validate_iso_nvidia_runtime "$ISO_RAW"
     cp "$ISO_RAW" "$ISO_OUT"
     echo ""
     echo "=== done (${BEE_GPU_VENDOR}) ==="

iso/builder/config/package-lists/bee-nvidia.list.chroot

@@ -1,6 +1,10 @@
-# NVIDIA DCGM (Data Center GPU Manager) — dcgmi diag for acceptance testing.
-# DCGM 4 is packaged per CUDA major. The image ships NVIDIA driver 590 with CUDA 13 userspace,
-# so install the CUDA 13 build plus proprietary diagnostic components explicitly.
+# NVIDIA DCGM (Data Center GPU Manager).
+# Validate uses dcgmi diagnostics; Burn uses dcgmproftester as the official
+# NVIDIA max-compute recipe. The smoketest/runtime contract treats
+# dcgmproftester as required in the LiveCD.
+# DCGM 4 is packaged per CUDA major. The image ships NVIDIA driver 590 with
+# CUDA 13 userspace, so install the CUDA 13 build plus proprietary components
+# explicitly.
 datacenter-gpu-manager-4-cuda13=1:%%DCGM_VERSION%%
 datacenter-gpu-manager-4-proprietary=1:%%DCGM_VERSION%%
 datacenter-gpu-manager-4-proprietary-cuda13=1:%%DCGM_VERSION%%

iso/builder/smoketest.sh

@@ -52,6 +52,31 @@ else
     fail "nvidia-smi: NOT FOUND"
 fi
 
+if p=$(PATH="/usr/local/bin:$PATH" command -v dcgmi 2>/dev/null); then
+    ok "dcgmi found: $p"
+else
+    fail "dcgmi: NOT FOUND"
+fi
+
+if p=$(PATH="/usr/local/bin:$PATH" command -v nv-hostengine 2>/dev/null); then
+    ok "nv-hostengine found: $p"
+else
+    fail "nv-hostengine: NOT FOUND"
+fi
+
+DCGM_PROFTESTER=""
+for tool in dcgmproftester dcgmproftester13 dcgmproftester12 dcgmproftester11; do
+    if p=$(PATH="/usr/local/bin:$PATH" command -v "$tool" 2>/dev/null); then
+        DCGM_PROFTESTER="$p"
+        break
+    fi
+done
+if [ -n "$DCGM_PROFTESTER" ]; then
+    ok "dcgmproftester found: $DCGM_PROFTESTER"
+else
+    fail "dcgmproftester: NOT FOUND"
+fi
+
 for tool in bee-gpu-burn bee-john-gpu-stress bee-nccl-gpu-stress all_reduce_perf; do
     if p=$(PATH="/usr/local/bin:$PATH" command -v "$tool" 2>/dev/null); then
         ok "$tool found: $p"
@@ -60,6 +85,12 @@ for tool in bee-gpu-burn bee-john-gpu-stress bee-nccl-gpu-stress all_reduce_perf
     fi
 done
 
+if p=$(PATH="/usr/local/bin:$PATH" command -v nvbandwidth 2>/dev/null); then
+    ok "nvbandwidth found: $p"
+else
+    warn "nvbandwidth: NOT FOUND"
+fi
+
 echo ""
 echo "-- NVIDIA modules --"
 KO_DIR="/usr/local/lib/nvidia"