feat(metrics): persist history in sqlite and add AMD memory validate tests

audit/internal/platform/sat.go

@@ -136,6 +136,54 @@ func (s *System) RunAMDAcceptancePack(ctx context.Context, baseDir string, logFu
 	}, logFunc)
 }
 
+// RunAMDMemIntegrityPack runs the official RVS MEM module as a validate-style memory integrity test.
+func (s *System) RunAMDMemIntegrityPack(ctx context.Context, baseDir string, logFunc func(string)) (string, error) {
+	if err := ensureAMDRuntimeReady(); err != nil {
+		return "", err
+	}
+	cfgFile := "/tmp/bee-amd-mem.conf"
+	cfg := `actions:
+- name: mem_integrity
+  device: all
+  module: mem
+  parallel: true
+  duration: 60000
+  copy_matrix: false
+  target_stress: 90
+  matrix_size: 8640
+`
+	_ = os.WriteFile(cfgFile, []byte(cfg), 0644)
+	return runAcceptancePackCtx(ctx, baseDir, "gpu-amd-mem", []satJob{
+		{name: "01-rocm-smi.log", cmd: []string{"rocm-smi"}},
+		{name: "02-rvs-mem.log", cmd: []string{"rvs", "-c", cfgFile}},
+		{name: "03-rocm-smi-after.log", cmd: []string{"rocm-smi", "--showtemp", "--showpower", "--showmemuse", "--csv"}},
+	}, logFunc)
+}
+
+// RunAMDMemBandwidthPack runs AMD's memory/interconnect bandwidth-oriented tools.
+func (s *System) RunAMDMemBandwidthPack(ctx context.Context, baseDir string, logFunc func(string)) (string, error) {
+	if err := ensureAMDRuntimeReady(); err != nil {
+		return "", err
+	}
+	cfgFile := "/tmp/bee-amd-babel.conf"
+	cfg := `actions:
+- name: babel_mem_bw
+  device: all
+  module: babel
+  parallel: true
+  copy_matrix: true
+  target_stress: 90
+  matrix_size: 134217728
+`
+	_ = os.WriteFile(cfgFile, []byte(cfg), 0644)
+	return runAcceptancePackCtx(ctx, baseDir, "gpu-amd-bandwidth", []satJob{
+		{name: "01-rocm-smi.log", cmd: []string{"rocm-smi"}},
+		{name: "02-rocm-bandwidth-test.log", cmd: []string{"rocm-bandwidth-test"}},
+		{name: "03-rvs-babel.log", cmd: []string{"rvs", "-c", cfgFile}},
+		{name: "04-rocm-smi-after.log", cmd: []string{"rocm-smi", "--showtemp", "--showpower", "--showmemuse", "--csv"}},
+	}, logFunc)
+}
+
 // RunAMDStressPack runs an AMD GPU burn-in pack.
 // Missing tools are reported as UNSUPPORTED, consistent with the existing SAT pattern.
 func (s *System) RunAMDStressPack(ctx context.Context, baseDir string, durationSec int, logFunc func(string)) (string, error) {
@@ -161,7 +209,7 @@ func amdStressRVSConfig(seconds int) string {
   module: gst
   parallel: true
   duration: %d
-  copy_matrix: true
+  copy_matrix: false
   target_stress: 90
   matrix_size_a: 8640
   matrix_size_b: 8640

audit/internal/platform/sat_test.go

@@ -39,15 +39,26 @@ func TestRunNvidiaAcceptancePackIncludesGPUStress(t *testing.T) {
 	}
 }
 
-func TestAMDStressConfigEnablesVRAMTraffic(t *testing.T) {
+func TestAMDStressConfigUsesSingleGSTAction(t *testing.T) {
 	t.Parallel()
 
 	cfg := amdStressRVSConfig(123)
-	if !strings.Contains(cfg, "copy_matrix: true") {
-		t.Fatalf("config missing VRAM copy path:\n%s", cfg)
+	if !strings.Contains(cfg, "module: gst") {
+		t.Fatalf("config missing gst module:\n%s", cfg)
 	}
-	if !strings.Contains(cfg, "duration: 123000") {
-		t.Fatalf("config missing millisecond duration:\n%s", cfg)
+	if strings.Contains(cfg, "module: mem") {
+		t.Fatalf("config should not include mem module:\n%s", cfg)
+	}
+	if !strings.Contains(cfg, "copy_matrix: false") {
+		t.Fatalf("config should use copy_matrix=false:\n%s", cfg)
+	}
+	if strings.Count(cfg, "duration: 123000") != 1 {
+		t.Fatalf("config should apply duration once:\n%s", cfg)
+	}
+	for _, field := range []string{"matrix_size_a: 8640", "matrix_size_b: 8640", "matrix_size_c: 8640"} {
+		if !strings.Contains(cfg, field) {
+			t.Fatalf("config missing %s:\n%s", field, cfg)
+		}
 	}
 }