Surface VRAM shortfall to the user and let them unload models

When the geometry in view needs more GPU memory than the cache can
hold, the viewer keeps the largest on-screen chunks resident and streams
the rest as the camera moves. That is the right degradation, but it was
invisible: nothing told the user the scene did not fit, and the only
lever was removing or hiding models, neither of which is "keep it in the
federation but stop spending GPU memory on it".

Viewer core:
- ModelGpuData::unloaded, with drawable() = !hidden && !unloaded now the
  test every cull / draw / pick / streaming pass uses. unloadModel evicts
  every chunk and releases the model's own buffers; loadModel recreates
  them from the CPU mirrors (no disk read) and lets chunks stream back.
  Recompose keeps the CPU instances current while a model is unloaded so
  a reload sees up-to-date transforms. The MeshGpu/InstanceGpu record
  builders are factored out so load and reload share them.
- FrameStats reports the camera's working set: chunks wanted, how many
  of those are not resident, and their bytes.
- modelVramBytes / isModelUnloaded accessors, forwarded by ViewportWindow.

BonsaiViewer:
- Models tree gains a memory column (name | MB | eye) refreshed once a
  second and on load-state changes; unloaded models read "unloaded" in
  italics. The viewport stays the single authority for the state;
  SessionState only carries the modelLoadStateChanged notification.
- Context menu: "Unload Model" / "Load Model", distinct from hide and
  remove, reporting the MB freed in the status bar.
- Status bar notice, independent of the perf-stats toggle, once the
  shortfall has persisted for 3 s (a moment of missing chunks after any
  camera move is normal): "GPU memory full: N of M visible chunks (X MB)
  not loaded", with a tooltip pointing at Unload. The perf label also
  shows "N/M chunks waiting".

Verified on the GPU: unloading a 497 MB model frees it immediately with
the others still rendering; reloading streams all 180 chunks back.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

src/ifcviewer/FrameStats.h

@@ -44,6 +44,15 @@ struct FrameStats {
     std::uint64_t vram_used_bytes;
     std::uint64_t vram_capacity_bytes;
     std::uint64_t vram_budget_bytes;
+    // The camera's working set: chunks the streaming driver wants resident
+    // (in frustum and large enough on screen) and how many of those are
+    // not — i.e. geometry the user should be seeing but is not yet, or
+    // cannot be because it does not fit the cache. Transiently non-zero
+    // after any camera move; persistently non-zero means the scene does
+    // not fit in VRAM.
+    std::uint32_t chunks_wanted;
+    std::uint32_t chunks_wanted_missing;
+    std::uint64_t wanted_missing_bytes;   // raw vertex + index bytes of the missing chunks
     // Whole-device VRAM from the driver (NVML / sysfs, see GpuMemory.h).
     // Desktop only; zero on web or when no backend could answer, so
     // consumers must treat 0 as "unknown" rather than as empty.

src/ifcviewer/ModelGpuData.h

@@ -447,6 +447,15 @@ struct ModelGpuData {
     // is gone; cull iterates m.chunks instead.
 
     bool hidden = false;
+    // Unloaded by the user: every chunk evicted and the model's own GPU
+    // buffers released, while the CPU mirrors (meshes, instances, chunk
+    // plan, element metadata) stay so the entry remains in the scene and
+    // loadModel can bring it back without touching the disk. Distinct
+    // from hidden (a viewing state; the geometry may stay resident) and
+    // from removal (the model leaves the scene).
+    bool unloaded = false;
+    // Whether cull / draw / pick / streaming should consider this model.
+    bool drawable() const { return !hidden && !unloaded; }
 
     // Per-model federation matrices in metres. Default identity → no
     // per-model contribution to the composed transform. See bonsai's

src/ifcviewer/ViewportCore.cpp

@@ -109,6 +109,39 @@ void releaseWgpuModelGpuData(ModelGpuData& m, BufferPool& pool) {
 
 // ---- Scene mutators -------------------------------------------------------
 
+namespace {
+
+// The GPU-side records the shaders read, built from the CPU mirrors.
+std::vector<MeshGpu> meshGpuRecords(const std::vector<MeshInfo>& meshes) {
+    std::vector<MeshGpu> gpu;
+    gpu.reserve(meshes.size());
+    for (const auto& mesh_info : meshes) {
+        MeshGpu rec = {};
+        for (int a = 0; a < 3; ++a) {
+            rec.aabb_min[a] = mesh_info.local_aabb_min[a];
+            rec.aabb_max[a] = mesh_info.local_aabb_max[a];
+        }
+        gpu.push_back(rec);
+    }
+    return gpu;
+}
+
+std::vector<InstanceGpu> instanceGpuRecords(const std::vector<InstanceInfo>& instances) {
+    std::vector<InstanceGpu> gpu(instances.size());
+    for (size_t i = 0; i < instances.size(); ++i) {
+        const InstanceInfo& inst = instances[i];
+        InstanceGpu& dst = gpu[i];
+        std::memcpy(dst.transform, inst.transform, sizeof(dst.transform));
+        dst.object_id            = inst.object_id;
+        dst.color_override_rgba8 = inst.color_override_rgba8;
+        dst.mesh_id              = inst.mesh_id;
+        dst._pad1                = 0;
+    }
+    return gpu;
+}
+
+}  // namespace
+
 void ViewportCore::removeModel(uint32_t session_model_id) {
     auto it = models_gpu_.find(session_model_id);
     if (it == models_gpu_.end()) return;
@@ -142,6 +175,46 @@ void ViewportCore::showModel(uint32_t session_model_id) {
     host_->requestFrame();
 }
 
+void ViewportCore::unloadModel(uint32_t session_model_id) {
+    auto it = models_gpu_.find(session_model_id);
+    if (it == models_gpu_.end() || it->second.unloaded) return;
+    ModelGpuData& m = it->second;
+    for (std::size_t ci = 0; ci < m.chunks.size(); ++ci) unloadChunk(m, ci);
+    releaseModelBuffers(m);
+    m.unloaded = true;
+    host_->requestFrame();
+}
+
+bool ViewportCore::loadModel(uint32_t session_model_id) {
+    auto it = models_gpu_.find(session_model_id);
+    if (it == models_gpu_.end()) return false;
+    ModelGpuData& m = it->second;
+    if (!m.unloaded) return true;
+    // Chunks stream back in on demand once the model is drawable again;
+    // only the model's own buffers have to be recreated here.
+    if (!createModelBuffers(session_model_id, m,
+                            meshGpuRecords(m.meshes), instanceGpuRecords(m.instances))) {
+        Log::warn() << "[wgpu] model " << session_model_id
+                    << " not reloaded: the device cannot fit its metadata buffers";
+        return false;
+    }
+    m.unloaded = false;
+    host_->requestFrame();
+    return true;
+}
+
+bool ViewportCore::isModelUnloaded(uint32_t session_model_id) const {
+    auto it = models_gpu_.find(session_model_id);
+    return it != models_gpu_.end() && it->second.unloaded;
+}
+
+std::uint64_t ViewportCore::modelVramBytes(uint32_t session_model_id) const {
+    auto it = models_gpu_.find(session_model_id);
+    if (it == models_gpu_.end()) return 0;
+    const ModelGpuData& m = it->second;
+    return m.vram_bytes_vbo + m.vram_bytes_ebo + m.vram_bytes_ssbo;
+}
+
 void ViewportCore::setFederatedFalseOrigin(const Eigen::Matrix4d& matrix_meters) {
     if (federated_false_origin_meters_ == matrix_meters) return;
     federated_false_origin_meters_ = matrix_meters;
@@ -264,17 +337,7 @@ float ViewportCore::chunkScreenAreaPx(const ModelGpuData::Chunk& c,
 
 void ViewportCore::uploadInstanceRecords(ModelGpuData& m) {
     if (!wgpu_initialized_ || m.instances.empty() || m.instance_storage == nullptr) return;
-
-    std::vector<InstanceGpu> gpu(m.instances.size());
-    for (size_t i = 0; i < m.instances.size(); ++i) {
-        const InstanceInfo& inst = m.instances[i];
-        InstanceGpu& dst = gpu[i];
-        std::memcpy(dst.transform, inst.transform, sizeof(dst.transform));
-        dst.object_id            = inst.object_id;
-        dst.color_override_rgba8 = inst.color_override_rgba8;
-        dst.mesh_id              = inst.mesh_id;
-        dst._pad1                = 0;
-    }
+    const std::vector<InstanceGpu> gpu = instanceGpuRecords(m.instances);
     wgpuQueueWriteBuffer(queue_, m.instance_storage, 0,
                          gpu.data(), gpu.size() * sizeof(InstanceGpu));
 }
@@ -284,8 +347,11 @@ void ViewportCore::recomposeAndUploadModel(uint32_t session_model_id) {
     auto it = models_gpu_.find(session_model_id);
     if (it == models_gpu_.end()) return;
     ModelGpuData& m = it->second;
-    if (m.instances.empty() || m.instance_storage == nullptr) return;
+    if (m.instances.empty()) return;
 
+    // The CPU mirrors are recomposed even while the model is unloaded (no
+    // instance_storage): cull and loadModel read them, and the upload
+    // below is skipped on its own.
     for (auto& inst : m.instances) composeInstanceFromPlacement(inst, m);
     uploadInstanceRecords(m);
 
@@ -2451,7 +2517,7 @@ void ViewportCore::driveStreamingLoads() {
     // last ~30 frames.
     constexpr float HISTORY_ALPHA = 1.0f / 30.0f;
     for (auto& [session_model_id, m] : models_gpu_) {
-        if (m.hidden) continue;
+        if (!m.drawable()) continue;
         for (auto& c : m.chunks) {
             if (c.is_resident && c.frustum_visible_count > 0) {
                 c.last_visible_frame_idx = streaming_frame_idx_;
@@ -2656,7 +2722,7 @@ void ViewportCore::driveStreamingLoads() {
     std::vector<Candidate> candidates;
     candidates.reserve(64);
     for (auto& [session_model_id, m] : models_gpu_) {
-        if (m.streaming_file_path.empty() || m.hidden) continue;
+        if (m.streaming_file_path.empty() || !m.drawable()) continue;
         for (std::size_t ci = 0; ci < m.chunks.size(); ++ci) {
             auto& c = m.chunks[ci];
             if (c.is_resident)                       continue;
@@ -2872,7 +2938,7 @@ void ViewportCore::driveStreamingLoads() {
     const bool growth_may_land = pool_.growth_pending() || pool_.can_grow();
     bool visible_pending = false;
     for (const auto& [session_model_id, m] : models_gpu_) {
-        if (m.streaming_file_path.empty() || m.hidden) continue;
+        if (m.streaming_file_path.empty() || !m.drawable()) continue;
         for (const auto& c : m.chunks) {
             if (c.is_resident) continue;
             if (c.is_loading) { visible_pending = true; break; }
@@ -3557,38 +3623,18 @@ void ViewportCore::applyCachedModel(std::uint32_t session_model_id,
     // Index section is NOT loaded upfront. Each chunk's index slice is
     // range-read alongside its vertex bytes in loadChunkBytesAndUploadGpu.
 
-    // MeshGpu storage (per-mesh quant basis).
-    std::vector<MeshGpu> mesh_gpu;
-    mesh_gpu.reserve(metadata.meta.meshes.size());
-    for (const auto& mesh_info : metadata.meta.meshes) {
-        MeshGpu mesh_gpu_record = {};
-        mesh_gpu_record.aabb_min[0] = mesh_info.local_aabb_min[0];
-        mesh_gpu_record.aabb_min[1] = mesh_info.local_aabb_min[1];
-        mesh_gpu_record.aabb_min[2] = mesh_info.local_aabb_min[2];
-        mesh_gpu_record.aabb_max[0] = mesh_info.local_aabb_max[0];
-        mesh_gpu_record.aabb_max[1] = mesh_info.local_aabb_max[1];
-        mesh_gpu_record.aabb_max[2] = mesh_info.local_aabb_max[2];
-        mesh_gpu.push_back(mesh_gpu_record);
-    }
-
-    // InstanceGpu storage. Rebase object_ids globally.
+    // Rebase object_ids globally (element metadata records rebase to match).
     const std::uint32_t object_id_base = next_object_id_;
     std::uint32_t max_local_id = 0;
-    std::vector<InstanceGpu> inst_gpu;
-    inst_gpu.reserve(metadata.meta.instances.size());
     for (auto& instance_cpu : metadata.meta.instances) {
         if (instance_cpu.object_id > max_local_id) max_local_id = instance_cpu.object_id;
         instance_cpu.object_id = object_id_base + instance_cpu.object_id;
-        InstanceGpu instance_gpu = {};
-        std::memcpy(instance_gpu.transform, instance_cpu.transform, sizeof(instance_gpu.transform));
-        instance_gpu.object_id            = instance_cpu.object_id;
-        instance_gpu.color_override_rgba8 = instance_cpu.color_override_rgba8;
-        instance_gpu.mesh_id              = instance_cpu.mesh_id;
-        inst_gpu.push_back(instance_gpu);
     }
     next_object_id_ = object_id_base + max_local_id + 1;
-    model_gpu_data.object_id_base = object_id_base;  // element metadata records rebase to match
+    model_gpu_data.object_id_base = object_id_base;
 
+    const std::vector<MeshGpu>     mesh_gpu = meshGpuRecords(metadata.meta.meshes);
+    const std::vector<InstanceGpu> inst_gpu = instanceGpuRecords(metadata.meta.instances);
     if (!createModelBuffers(session_model_id, model_gpu_data, mesh_gpu, inst_gpu)) {
         Log::warn() << "[wgpu] model " << session_model_id
                     << " not loaded: the device cannot fit its metadata buffers";
@@ -3980,7 +4026,7 @@ void ViewportCore::pumpWebChunkLoads() {
         auto it = models_gpu_.find(next.session_model_id);
         if (it == models_gpu_.end()) continue;
         ModelGpuData& m = it->second;
-        if (m.hidden || next.ci >= m.chunks.size()) continue;
+        if (!m.drawable() || next.ci >= m.chunks.size()) continue;
         auto& c = m.chunks[next.ci];
         if (c.is_resident || c.is_loading)     continue;
         if (c.contribution_visible_count == 0) continue;   // no longer worth drawing
@@ -4328,6 +4374,7 @@ void ViewportCore::streamingProgress(int& resident_chunks, int& total_chunks) co
     resident_chunks = 0;
     total_chunks    = 0;
     for (const auto& [session_model_id, m] : models_gpu_) {
+        if (m.unloaded) continue;
         for (const auto& c : m.chunks) {
             ++total_chunks;
             if (c.is_resident) ++resident_chunks;
@@ -4359,7 +4406,7 @@ void ViewportCore::streamingModelProgress(int idx, int& resident_chunks,
     total_chunks    = 0;
     if (idx < 0 || idx >= int(models_gpu_.size())) return;
     auto it = models_gpu_.find(modelIdsInLoadOrder()[std::size_t(idx)]);
-    if (it == models_gpu_.end()) return;
+    if (it == models_gpu_.end() || it->second.unloaded) return;
     for (const auto& c : it->second.chunks) {
         ++total_chunks;
         if (c.is_resident) ++resident_chunks;
@@ -4428,7 +4475,7 @@ void ViewportCore::streamingByteProgress(std::uint64_t& total_bytes,
     // whole model this view even requires.
     total_bytes = needed_bytes = loaded_bytes = 0;
     for (const auto& [session_model_id, m] : models_gpu_) {
-        if (m.hidden) continue;
+        if (!m.drawable()) continue;
         for (const auto& c : m.chunks) {
             // Report COMPRESSED bytes — what actually crosses the network. Fall
             // back to raw for direct (in-memory) loads that have no blobs.
@@ -5682,7 +5729,7 @@ void ViewportCore::encodeSelectionMaskPass(WGPUCommandEncoder enc) {
     // Same draw stream as the main pass, opaque and transparent together —
     // a selected element that happens to be translucent still gets a halo.
     for (const auto& [session_model_id, m] : models_gpu_) {
-        if (m.hidden) continue;
+        if (!m.drawable()) continue;
         for (const auto& c : m.chunks) {
             if (!c.bind_group || c.total_visible_vertices == 0) continue;
             wgpuRenderPassEncoderSetBindGroup(pass, 1, c.bind_group, 0, nullptr);
@@ -6144,7 +6191,7 @@ void ViewportCore::encodePickReadbackToStaging(int x_pixels, int y_pixels,
     wgpuRenderPassEncoderSetPipeline(pass, pick_pipeline_);
     wgpuRenderPassEncoderSetBindGroup(pass, 0, frame_bind_group_, 0, nullptr);
     for (const auto& [session_model_id, m] : models_gpu_) {
-        if (m.hidden) continue;
+        if (!m.drawable()) continue;
         for (const auto& c : m.chunks) {
             if (!c.bind_group || c.total_visible_vertices == 0) continue;
             wgpuRenderPassEncoderSetBindGroup(pass, 1, c.bind_group, 0, nullptr);
@@ -6343,7 +6390,7 @@ void ViewportCore::isolateSelected() {
     // object_id 0 (unpickable) is skipped.
     const auto& sel_ids = selection_.selectionIds();
     for (const auto& [session_model_id, m] : models_gpu_) {
-        if (m.hidden) continue;
+        if (!m.drawable()) continue;
         for (const InstanceInfo& inst : m.instances) {
             if (inst.object_id == 0) continue;
             if (sel_ids.find(inst.object_id) == sel_ids.end())
@@ -6366,7 +6413,7 @@ void ViewportCore::hideAll() {
     // showAll, and isolateSelected with an empty selection. Model-hidden
     // models are already gone from the cull, so they contribute nothing.
     for (const auto& [session_model_id, m] : models_gpu_) {
-        if (m.hidden) continue;
+        if (!m.drawable()) continue;
         for (const InstanceInfo& inst : m.instances) visibility_.hide(inst.object_id);
     }
     Log::info().noquote().nospace() << "[wgpu] hid all (" << visibility_.hiddenCount() << ")";
@@ -6546,7 +6593,7 @@ bool ViewportCore::encodeBoxPickToStaging(int& x, int& y, int& w, int& h,
     wgpuRenderPassEncoderSetPipeline(pass, pick_pipeline_);
     wgpuRenderPassEncoderSetBindGroup(pass, 0, frame_bind_group_, 0, nullptr);
     for (const auto& [session_model_id, m] : models_gpu_) {
-        if (m.hidden) continue;
+        if (!m.drawable()) continue;
         for (const auto& c : m.chunks) {
             if (!c.bind_group || c.total_visible_vertices == 0) continue;
             wgpuRenderPassEncoderSetBindGroup(pass, 1, c.bind_group, 0, nullptr);
@@ -6666,7 +6713,7 @@ bool ViewportCore::encodeXrayBoxPickToStaging(int& x, int& y, int& w, int& h,
     wgpuRenderPassEncoderSetPipeline(pass, box_pick_pipeline_);
     wgpuRenderPassEncoderSetBindGroup(pass, 0, frame_bind_group_, 0, nullptr);
     for (const auto& [session_model_id, m] : models_gpu_) {
-        if (m.hidden) continue;
+        if (!m.drawable()) continue;
         for (const auto& c : m.chunks) {
             if (!c.bind_group || c.total_visible_vertices == 0) continue;
             wgpuRenderPassEncoderSetBindGroup(pass, 1, c.bind_group, 0, nullptr);
@@ -6856,7 +6903,7 @@ bool ViewportCore::raycastSurfaceForObject(std::uint32_t object_id, int x_pixels
     float           best_radius = 0.0f;
     bool found = false;
     for (const auto& [session_model_id, m] : models_gpu_) {
-        if (m.hidden) continue;
+        if (!m.drawable()) continue;
         for (const auto& inst : m.instances) {
             if (inst.object_id != object_id) continue;
             float t = 0.0f;
@@ -7171,7 +7218,7 @@ bool ViewportCore::raycast(const float origin[3], const float dir[3],
     float         best_normal[3] = {0, 0, 0};
 
     for (const auto& [session_model_id, m] : models_gpu_) {
-        if (m.hidden) continue;
+        if (!m.drawable()) continue;
         for (std::uint32_t inst_idx = 0; inst_idx < std::uint32_t(m.instances.size()); ++inst_idx) {
             const InstanceInfo& inst = m.instances[inst_idx];
             if (!rayAabbSlab(origin, inv_d, inst.world_aabb_min, inst.world_aabb_max)) {
@@ -7732,7 +7779,7 @@ void ViewportCore::render() {
             std::vector<std::pair<std::uint32_t, std::future<std::uint32_t>>> futures;
             futures.reserve(models_gpu_.size());
             for (auto& [session_model_id, m] : models_gpu_) {
-                if (m.hidden) continue;
+                if (!m.drawable()) continue;
                 auto& m_ref = m;
                 futures.emplace_back(session_model_id, std::async(std::launch::async,
                     [this, &m_ref, &planes, &eye_a, &fwd_a, &right_a, &up_a,
@@ -7749,7 +7796,7 @@ void ViewportCore::render() {
             }
         } else {
             for (auto& [session_model_id, m] : models_gpu_) {
-                if (m.hidden) continue;
+                if (!m.drawable()) continue;
                 hiz_reject_count_ += cullModelCpuCompute(
                     m, planes, eye_a, fwd_a, right_a, up_a, focal_px,
                     effective_min_px, lod1_pixel_threshold_,
@@ -7761,7 +7808,7 @@ void ViewportCore::render() {
         Stopwatch upload_timer;
         upload_timer.start();
         for (auto& [session_model_id, m] : models_gpu_) {
-            if (m.hidden) continue;
+            if (!m.drawable()) continue;
             cullModelCpuUpload(m);
             for (const auto& c : m.chunks) {
                 last_visible_objects_   += c.total_visible_draws;
@@ -7842,7 +7889,7 @@ void ViewportCore::render() {
         wgpuRenderPassEncoderSetBindGroup(pass, 0, frame_bind_group_, 0, nullptr);
 
         for (const auto& [session_model_id, m] : models_gpu_) {
-            if (m.hidden) continue;
+            if (!m.drawable()) continue;
             for (const auto& c : m.chunks) {
                 if (!c.bind_group || c.opaque_visible_vertices == 0) continue;
                 wgpuRenderPassEncoderSetBindGroup(pass, 1, c.bind_group, 0, nullptr);
@@ -7853,7 +7900,7 @@ void ViewportCore::render() {
 
         wgpuRenderPassEncoderSetPipeline(pass, main_pipeline_transparent_);
         for (const auto& [session_model_id, m] : models_gpu_) {
-            if (m.hidden) continue;
+            if (!m.drawable()) continue;
             for (const auto& c : m.chunks) {
                 if (!c.bind_group) continue;
                 const std::uint32_t transparent_verts =
@@ -7975,13 +8022,29 @@ void ViewportCore::render() {
         stats.unique_meshes     = total_meshes;
         std::uint32_t draw_calls = 0;
         for (const auto& [session_model_id, mm] : models_gpu_) {
-            if (mm.hidden) continue;
+            if (!mm.drawable()) continue;
             for (const auto& c : mm.chunks) {
                 if (c.is_resident && c.total_visible_draws > 0) ++draw_calls;
             }
         }
         stats.gl_draw_calls      = draw_calls;
         stats.indirect_sub_draws = last_sub_draws_;
+        std::uint32_t wanted = 0, wanted_missing = 0;
+        std::uint64_t wanted_missing_bytes = 0;
+        for (const auto& [session_model_id, mm] : models_gpu_) {
+            if (!mm.drawable()) continue;
+            for (const auto& c : mm.chunks) {
+                if (c.contribution_visible_count == 0) continue;
+                ++wanted;
+                if (c.is_resident) continue;
+                ++wanted_missing;
+                wanted_missing_bytes += c.vertex_byte_size
+                                      + c.index_count * sizeof(std::uint32_t);
+            }
+        }
+        stats.chunks_wanted         = wanted;
+        stats.chunks_wanted_missing = wanted_missing;
+        stats.wanted_missing_bytes  = wanted_missing_bytes;
         stats.vram_used_bytes     = pool_.total_used_bytes();
         stats.vram_capacity_bytes = pool_.total_capacity_bytes();
         stats.vram_budget_bytes   = budget_.bounded() ? budget_.cache_budget_bytes() : 0;

src/ifcviewer/ViewportCore.h

@@ -157,6 +157,17 @@ public:
     void resetScene();
     void hideModel(uint32_t session_model_id);
     void showModel(uint32_t session_model_id);
+    // Release a model's GPU memory (every chunk + its own buffers) while
+    // keeping it in the scene; loadModel recreates the buffers from the
+    // CPU mirrors and lets chunks stream back. Neither touches hidden.
+    // loadModel returns false when the device cannot fit the model's
+    // buffers even after the cache yielded (it stays unloaded).
+    void unloadModel(uint32_t session_model_id);
+    bool loadModel(uint32_t session_model_id);
+    bool isModelUnloaded(uint32_t session_model_id) const;
+    // Bytes this model currently holds on the GPU: resident chunk
+    // geometry plus its mesh/instance/cull buffers. 0 when unloaded.
+    std::uint64_t modelVramBytes(uint32_t session_model_id) const;
 
     // Federation matrix setters. Each writes to model state and posts
     // a recompose so per-instance world matrices stay consistent with

src/ifcviewer/ViewportWindow.cpp

@@ -594,6 +594,10 @@ void ViewportWindow::removeModel(uint32_t session_model_id)   { core_.removeMode
 void ViewportWindow::resetScene()                     { core_.resetScene(); }
 void ViewportWindow::hideModel(uint32_t session_model_id)     { core_.hideModel(session_model_id); }
 void ViewportWindow::showModel(uint32_t session_model_id)     { core_.showModel(session_model_id); }
+void ViewportWindow::unloadModel(uint32_t session_model_id)   { core_.unloadModel(session_model_id); }
+bool ViewportWindow::loadModel(uint32_t session_model_id)     { return core_.loadModel(session_model_id); }
+bool ViewportWindow::isModelUnloaded(uint32_t session_model_id) const { return core_.isModelUnloaded(session_model_id); }
+std::uint64_t ViewportWindow::modelVramBytes(uint32_t session_model_id) const { return core_.modelVramBytes(session_model_id); }
 
 void ViewportWindow::setFederatedFalseOrigin(const Eigen::Matrix4d& m) {
     core_.setFederatedFalseOrigin(m);

src/ifcviewer/ViewportWindow.h

@@ -142,6 +142,13 @@ public:
     // consults. requestUpdate() so the change is visible immediately.
     void hideModel(uint32_t session_model_id);
     void showModel(uint32_t session_model_id);
+    // GPU residency of a model, independent of visibility: unloadModel
+    // frees everything it holds on the device while it stays in the
+    // scene; loadModel brings it back (false if the device cannot fit it).
+    void unloadModel(uint32_t session_model_id);
+    bool loadModel(uint32_t session_model_id);
+    bool isModelUnloaded(uint32_t session_model_id) const;
+    std::uint64_t modelVramBytes(uint32_t session_model_id) const;
 
     // Federation pipeline: composed instance transform =
     //   FederatedFalseOrigin · ModelTransformation · CoordinateOperation