feat(nccl): add nccl-tests all_reduce_perf for GPU bandwidth testing

- Dockerfile: install cuda-nvcc-13-0 from NVIDIA repo for compilation
- build-nccl-tests.sh: downloads libnccl-dev for nccl.h, builds all_reduce_perf
- build.sh: runs nccl-tests build, injects binary into /usr/local/bin/
- platform: RunNCCLTests() auto-detects GPU count, runs all_reduce_perf
- TUI: NCCL bandwidth test entry in Burn-in Tests screen [N] hotkey

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

audit/internal/tui/forms.go

@@ -1,6 +1,7 @@
 package tui
 
 import (
+	"context"
 	"time"
 
 	"bee/audit/internal/platform"
@@ -140,6 +141,15 @@ func (m model) updateConfirm(msg tea.KeyMsg) (tea.Model, tea.Cmd) {
 			)
 		case actionRunFanStress:
 			return m.startGPUStressTest()
+		case actionRunNCCLTests:
+			m.busy = true
+			m.busyTitle = "NCCL bandwidth test"
+			ctx, cancel := context.WithCancel(context.Background())
+			m.ncclCancel = cancel
+			return m, func() tea.Msg {
+				result, err := m.app.RunNCCLTestsResult(ctx)
+				return resultMsg{title: result.Title, body: result.Body, err: err, back: screenBurnInTests}
+			}
 		}
 	case "ctrl+c":
 		return m, tea.Quit
@@ -153,7 +163,7 @@ func (m model) confirmCancelTarget() screen {
 		return screenExportTargets
 	case actionRunAll, actionRunMemorySAT, actionRunStorageSAT, actionRunCPUSAT, actionRunAMDGPUSAT:
 		return screenHealthCheck
-	case actionRunFanStress:
+	case actionRunFanStress, actionRunNCCLTests:
 		return screenBurnInTests
 	default:
 		return screenMain