LRC 14.2 Preview generation frequently using less than 50% CPU after several minutes

Report · Apr 05, 2025

I am running Lightroom Classic 14.2 on a PC with a Ryzen 9950X processor, Radeon 6800XT GPU, PCIe Gen5 NVMe, and 64GB RAM. I have noticed that during 1:1 preview generation, the percentage of CPU being used is primarily below 50% after several minutes of runtime (in the case I'm running now, this started after 5 minutes or so, and the generation is about 10 minutes in at this point. I am generating 1:1 previews for a batch of 4750 Canon R5 and R5II RAW files as a performance test. I previously deleted the Lightroom preview cache and optimized the catalog.

The CPU temperature seems within limits (around 80-85C, but sometimes even lower), and core clock is showing up as running at 100% maximum frequency in the Windows Resource Monitor. There is plenty of free disk space (over 2.2 TB free).

Is this slowdown in core utilization expected? I just moved from an older PC with a Ryzen 5600X that also exhibited a similar issue. I was hoping to see more of an improvement with the new PC. System info below.

Thanks!

------

Lightroom Classic version: 14.2 [ 202502071718-3869eef7 ]

License: Creative Cloud

Language setting: en

Operating system: Windows 11 - Home Premium Edition

Version: 11.0.26100

Application architecture: x64

System architecture: x64

Logical processor count: 32

Processor speed: 4.3GHz

SqLite Version: 3.36.0

CPU Utilisation: 41.0%

Power Source: Plugged In

Built-in memory: 63091.8 MB

Dedicated GPU memory used by Lightroom: 128.4MB / 16311.2MB (0%)

Real memory available to Lightroom: 63091.8 MB

Real memory used by Lightroom: 21836.1 MB (34.6%)

Virtual memory used by Lightroom: 25158.1 MB

GDI objects count: 705

USER objects count: 2440

Process handles count: 4884

Memory cache size: 33131.5MB

Internal Camera Raw version: 17.2 [ 2155 ]

Maximum thread count used by Camera Raw: 5

Camera Raw SIMD optimization: SSE2,AVX,AVX2

Camera Raw virtual memory: 14140MB / 31545MB (44%)

Camera Raw real memory: 14152MB / 63091MB (22%)

Cache1:

NT- RAM:0.0MB, VRAM:0.0MB, Combined:0.0MB

Cache2:

m:33131.5MB, n:0.0MB

U-main: 108.0MB

System DPI setting: 144 DPI (high DPI mode)

Desktop composition enabled: Yes

Standard Preview Size: 2344 pixels

Displays: 1) 2560x1440, 2) 3840x2160

Input types: Multitouch: No, Integrated touch: No, Integrated pen: No, External touch: No, External pen: No, Keyboard: No

Graphics Processor Info:

DirectX: AMD Radeon RX 6800 XT (32.0.13031.3015)

Init State: GPU for Export supported by default

User Preference: Auto

Enable HDR in Library: OFF

Application folder: C:\Program Files\Adobe\Adobe Lightroom Classic

Library Path: C:\Users\USERNAME\Pictures\Lightroom\Lightroom Catalog.lrcat

Settings Folder: C:\Users\USERNAME\AppData\Roaming\Adobe\Lightroom

Installed Plugins:

1) AdobeStock

2) Flickr

Config.lua flags:

Adapter #1: Vendor : 1002

Device : 73bf

Subsystem : e3a1002

Revision : c1

Video Memory : 16311

Adapter #2: Vendor : 1002

Device : 13c0

Subsystem : d0001458

Revision : c1

Video Memory : 7d1

Adapter #3: Vendor : 1414

Device : 8c

Subsystem : 0

Revision : 0

Video Memory : 0

AudioDeviceIOBlockSize: 1024

AudioDeviceName: System Default - 3 - ASUS MG279 (AMD High Definition Audio Device)

AudioDeviceNumberOfChannels: 2

AudioDeviceSampleRate: 48000

Build: LR5x26

Direct2DEnabled: false

GL_ACCUM_ALPHA_BITS: 8

GL_ACCUM_BLUE_BITS: 8

GL_ACCUM_GREEN_BITS: 8

GL_ACCUM_RED_BITS: 8

GL_ALPHA_BITS: 8

GL_BLUE_BITS: 8

GL_DEPTH_BITS: 32

GL_GREEN_BITS: 8

GL_MAX_3D_TEXTURE_SIZE: 8192

GL_MAX_TEXTURE_SIZE: 16384

GL_MAX_TEXTURE_UNITS: 8

GL_MAX_VIEWPORT_DIMS: 16384,16384

GL_RED_BITS: 8

GL_RENDERER: AMD Radeon RX 6800 XT

GL_SHADING_LANGUAGE_VERSION: 4.60

GL_STENCIL_BITS: 8

GL_VENDOR: ATI Technologies Inc.

GL_VERSION: 4.6.0 Compatibility Profile Context 25.3.1.250220

GPUDeviceEnabled: false

OGLEnabled: true

GL_EXTENSIONS: GL_EXT_abgr GL_EXT_blend_color GL_EXT_blend_minmax GL_EXT_blend_subtract GL_EXT_copy_texture GL_EXT_texture GL_EXT_texture_object GL_EXT_vertex_array GL_EXT_compiled_vertex_array GL_EXT_subtexture GL_EXT_texture3D GL_EXT_bgra GL_EXT_draw_range_elements GL_EXT_packed_pixels GL_EXT_point_parameters GL_EXT_rescale_normal GL_EXT_separate_specular_color GL_EXT_texture_edge_clamp GL_SGIS_generate_mipmap GL_SGIS_texture_edge_clamp GL_SGIS_texture_lod GL_ARB_multitexture GL_IBM_texture_mirrored_repeat GL_AMD_debug_output GL_ARB_multisample GL_ARB_texture_cube_map GL_ARB_texture_env_add GL_ARB_transpose_matrix GL_EXT_blend_func_separate GL_EXT_fog_coord GL_EXT_multi_draw_arrays GL_EXT_secondary_color GL_EXT_texture_cube_map GL_EXT_texture_env_add GL_EXT_texture_filter_anisotropic GL_EXT_texture_lod GL_EXT_texture_lod_bias GL_NV_blend_square GL_NV_texgen_reflection GL_SUN_multi_draw_arrays WGL_EXT_swap_control GL_ARB_point_parameters GL_ARB_texture_border_clamp GL_ARB_texture_compression GL_ARB_texture_env_crossbar GL_EXT_texture_compression_s3tc GL_EXT_texture_env_combine GL_EXT_texture_env_dot3 GL_ARB_shadow GL_ARB_shadow_ambient GL_ARB_texture_env_combine GL_ARB_texture_env_dot3 GL_ARB_texture_mirrored_repeat GL_NV_half_float GL_ARB_depth_texture GL_ARB_fragment_program GL_ARB_vertex_program GL_ARB_window_pos GL_ATI_draw_buffers GL_ATI_texture_env_combine3 GL_ATI_texture_float GL_EXT_shadow_funcs GL_EXT_stencil_wrap GL_EXT_texture_rectangle GL_NV_primitive_restart GL_S3_s3tc GL_ARB_fragment_program_shadow GL_ARB_fragment_shader GL_ARB_occlusion_query GL_ARB_point_sprite GL_ARB_shader_objects GL_ARB_shading_language_100 GL_ARB_texture_non_power_of_two GL_ARB_vertex_buffer_object GL_ARB_vertex_shader GL_EXT_blend_equation_separate GL_EXT_depth_bounds_test GL_ARB_color_buffer_float GL_ARB_draw_buffers GL_ARB_half_float_pixel GL_ARB_pixel_buffer_object GL_ARB_texture_float GL_ARB_texture_rectangle GL_EXT_framebuffer_object GL_EXT_pixel_buffer_object GL_EXT_texture_mirror_clamp GL_NV_float_buffer GL_EXT_framebuffer_blit GL_EXT_framebuffer_multisample GL_EXT_packed_depth_stencil GL_ATI_separate_stencil GL_ATI_shader_texture_lod GL_ATI_texture_mirror_once GL_EXT_bindable_uniform GL_EXT_draw_buffers2 GL_EXT_draw_instanced GL_EXT_framebuffer_sRGB GL_EXT_geometry_shader4 GL_EXT_gpu_program_parameters GL_EXT_gpu_shader4 GL_EXT_packed_float GL_EXT_texture_array GL_EXT_texture_buffer_object GL_EXT_texture_compression_latc GL_EXT_texture_compression_rgtc GL_EXT_texture_integer GL_EXT_transform_feedback GL_NV_depth_buffer_float GL_OES_EGL_image GL_EXT_provoking_vertex GL_EXT_texture_sRGB GL_EXT_texture_shared_exponent GL_EXT_vertex_array_bgra GL_NV_conditional_render GL_ARB_depth_buffer_float GL_ARB_draw_instanced GL_ARB_framebuffer_object GL_ARB_framebuffer_sRGB GL_ARB_geometry_shader4 GL_ARB_half_float_vertex GL_ARB_instanced_arrays GL_ARB_map_buffer_range GL_ARB_texture_buffer_object GL_ARB_texture_compression_rgtc GL_ARB_texture_rg GL_ARB_transform_feedback GL_ARB_vertex_array_object GL_EXT_direct_state_access GL_EXT_texture_snorm GL_EXT_texture_swizzle GL_AMD_conservative_depth GL_AMD_seamless_cubemap_per_texture GL_AMD_texture_cube_map_array GL_ARB_compatibility GL_ARB_copy_buffer GL_ARB_depth_clamp GL_ARB_draw_buffers_blend GL_ARB_draw_elements_base_vertex GL_ARB_fragment_coord_conventions GL_ARB_provoking_vertex GL_ARB_sample_shading GL_ARB_seamless_cube_map GL_ARB_shader_texture_lod GL_ARB_sync GL_ARB_texture_buffer_object_rgb32 GL_ARB_texture_cube_map_array GL_ARB_texture_gather GL_ARB_texture_multisample GL_ARB_texture_multisample_no_array GL_ARB_texture_query_lod GL_ARB_uniform_buffer_object GL_ARB_vertex_array_bgra GL_ATI_meminfo GL_EXT_texture_storage GL_NV_copy_image GL_NV_texture_barrier GL_AMD_blend_minmax_factor GL_AMD_depth_clamp_separate GL_AMD_sample_positions GL_ARB_ES2_compatibility GL_ARB_base_instance GL_ARB_blend_func_extended GL_ARB_compressed_texture_pixel_storage GL_ARB_debug_output GL_ARB_draw_indirect GL_ARB_explicit_attrib_location GL_ARB_get_program_binary GL_ARB_gpu_shader5 GL_ARB_gpu_shader_fp64 GL_ARB_occlusion_query2 GL_ARB_robustness GL_ARB_sampler_objects GL_ARB_separate_shader_objects GL_ARB_shader_bit_encoding GL_ARB_shader_precision GL_ARB_shader_stencil_export GL_ARB_shader_subroutine GL_ARB_shading_language_packing GL_ARB_tessellation_shader GL_ARB_texture_compression_bptc GL_ARB_texture_rgb10_a2ui GL_ARB_texture_swizzle GL_ARB_timer_query GL_ARB_transform_feedback2 GL_ARB_transform_feedback3 GL_ARB_vertex_attrib_64bit GL_ARB_vertex_type_2_10_10_10_rev GL_ARB_viewport_array GL_EXT_shader_image_load_store GL_EXT_texture_sRGB_decode GL_EXT_vertex_attrib_64bit GL_AMD_multi_draw_indirect GL_AMD_pinned_memory GL_AMD_vertex_shader_layer GL_AMD_vertex_shader_viewport_index GL_ARB_conservative_depth GL_ARB_internalformat_query GL_ARB_shader_atomic_counters GL_ARB_shader_image_load_store GL_ARB_shading_language_420pack GL_ARB_texture_storage GL_ARB_transform_feedback_instanced GL_EXT_debug_label GL_AMD_query_buffer_object GL_AMD_shader_trinary_minmax GL_ARB_ES3_compatibility GL_ARB_arrays_of_arrays GL_ARB_clear_buffer_object GL_ARB_compute_shader GL_ARB_copy_image GL_ARB_explicit_uniform_location GL_ARB_fragment_layer_viewport GL_ARB_framebuffer_no_attachments GL_ARB_internalformat_query2 GL_ARB_invalidate_subdata GL_ARB_multi_draw_indirect GL_ARB_program_interface_query GL_ARB_shader_image_size GL_ARB_shader_storage_buffer_object GL_ARB_stencil_texturing GL_ARB_texture_buffer_range GL_ARB_texture_query_levels GL_ARB_texture_storage_multisample GL_ARB_texture_view GL_ARB_vertex_attrib_binding GL_EXT_copy_buffer GL_KHR_debug GL_KHR_texture_compression_astc_ldr GL_AMD_gpu_shader_half_float GL_AMD_gpu_shader_int64 GL_ARB_ES3_1_compatibility GL_ARB_bindless_texture GL_ARB_buffer_storage GL_ARB_clear_texture GL_ARB_enhanced_layouts GL_ARB_indirect_parameters GL_ARB_multi_bind GL_ARB_query_buffer_object GL_ARB_seamless_cubemap_per_texture GL_ARB_shader_draw_parameters GL_ARB_shader_group_vote GL_ARB_sparse_texture GL_ARB_texture_mirror_clamp_to_edge GL_ARB_texture_stencil8 GL_ARB_vertex_type_10f_11f_11f_rev GL_EXT_shader_integer_mix GL_EXT_timer_query GL_KHR_blend_equation_advanced GL_KHR_blend_equation_advanced_coherent GL_NV_shader_atomic_int64 GL_ARB_clip_control GL_ARB_conditional_render_inverted GL_ARB_cull_distance GL_ARB_derivative_control GL_ARB_direct_state_access GL_ARB_get_texture_sub_image GL_ARB_pipeline_statistics_query GL_ARB_shader_texture_image_samples GL_ARB_sparse_buffer GL_ARB_texture_barrier GL_ARB_transform_feedback_overflow_query GL_EXT_polygon_offset_clamp GL_EXT_shader_image_load_formatted GL_KHR_context_flush_control GL_KHR_robust_buffer_access_behavior GL_KHR_robustness GL_OVR_multiview GL_ARB_ES3_2_compatibility GL_ARB_fragment_shader_interlock GL_ARB_gpu_shader_int64 GL_ARB_parallel_shader_compile GL_ARB_post_depth_coverage GL_ARB_sample_locations GL_ARB_shader_atomic_counter_ops GL_ARB_shader_ballot GL_ARB_shader_clock GL_ARB_shader_viewport_layer_array GL_ARB_texture_filter_minmax GL_EXT_texture_sRGB_R8 GL_EXT_texture_sRGB_RG8 GL_EXT_texture_shadow_lod GL_KHR_no_error GL_OVR_multiview2 GL_ARB_gl_spirv GL_EXT_memory_object GL_EXT_memory_object_win32 GL_EXT_semaphore GL_EXT_semaphore_win32 GL_AMD_gpu_shader_int16 GL_ARB_polygon_offset_clamp GL_ARB_spirv_extensions GL_ARB_texture_filter_anisotropic GL_EXT_texture_compression_bptc GL_KHR_parallel_shader_compile GL_KHR_shader_subgroup GL_EXT_nonuniform_qualifier GL_NV_mesh_shader GL_NV_timeline_semaphore

Report · Apr 05, 2025

"Init State: GPU for Export supported by default

User Preference: Auto"

With newer GPUs, more of the work of rendering photos is done in the GPU than in the CPU, and it's normal to see the CPU well below 50% utilization. When LR renders a raw, it first does preparatory work in the CPU before handing it off to the GPU. So when you initiate a batch export, the CPU spikes up because LR is preparing the first set of photos to hand off to the GPU. But then as the export proceeds, the load on the CPU evens out as the processing pipeline balances out between CPU and GPU.

Report · Apr 05, 2025

@johnrellis

The OP is building 1:1 previews not exporting the images. Also, preview building is currently handled by the CPU.

Report · Apr 05, 2025

@Ian Lyons is correct - I am building 1:1 previews. My understanding is that the GPU is not involved. The GPU was also almost completely idle during preview generation.

Report · Apr 05, 2025

[This post contains formatting and embedded images that don't appear in email. View the post in your Web browser.]

Right, thanks for the correction, I don't know why I thought you were doing exports.

Make sure you have the option Preferences > Performance > Generate Previews In Parallel selected. On my Mac building 1:1 previews of raws, CPU utilization is about 50% with that option off and close to 100% with it on. In the screenshot below, you can see precisely when I set it on while building previews:

Report · Apr 05, 2025

"Generate Previews in parallel" was already selected, so unfortunately that is not the cause. I should also mention that these 1:1 previews were of "finished" images each with many edits and noise reduction.

Report · Apr 05, 2025

"DirectX: AMD Radeon RX 6800 XT (32.0.13031.3015)"

It likely isn't related to this issue, but there is a slightly newer graphics driver available from 3/20/25 (32.0.13031.4034):

https://www.amd.com/en/support/downloads/drivers.html/graphics/radeon-rx/radeon-rx-6000-series/amd-r...

Report · Apr 05, 2025

Thanks to @Ian Lyons and @johnrellis for your replies so far. My current theory after further experiments is that there may be a memory leak.

Based on John's last comment, I reconfirmed that parallel preview generation was set. I disabled that and ran a small batch of 50 photos. CPU utilization was around 20%. I then reenabled preview generation. I was able to see a small batch of 50 photos generate previews in parallel with 100% CPU utilization. I then tried a batch of 1000 photos. Those 1000 photos were running at 100% CPU utilization until about 656 previews in, then utilization went down (and process memory went down about 11GB). Eventually, memory usage crept back up. It took 14-15 minutes total to generate these previews.

Finally, I tried a second (different) batch of 1000 photos to generate 1000 previews for. The second batch also started off at 100% utilization but started slowing down around 4 minutes in. System memory usage pushed up to 44.4GB (out of 64GB total), then perf dropped and memory usage dropped. System memory usage kept bouncing between 42GB and 44GB, and CPU utilization dropped each time that happened. I eventually noticed the Lightroom.exe process was reaching 32GB, then dropping 2-3GB, then reaching 32GB again. Each time the memory usage dropped, CPU utilization dropped temporarily. But it was happening frequently so it was impacting performance. After the second batch of 1000 photos finished (16m30s or so) Lightroom.exe was using 28,881,236K (~28GB) of memory.

I can rerun my initial experiment with the full 4750 previews, but I suspect the memory issue may have been happening during that run, made even worse by the large number of photos being processed. These photos have a large number of edits, AI selections, etc and most have noise reduction applied.

Report · Apr 05, 2025

[This post contains formatting and embedded images that don't appear in email. View the post in your Web browser.]

"System memory usage pushed up to 44.4GB"

That is unusually high. I ran a test on my Mac LR 14.2, building 1:1 previews for 4245 raws (proprietary and DNGs), with lots of masking, AI Remove, and Denoise. CPU utilization was constantly near 100%, and memory bounced between 3 GB and 5 GB:

If you can identify a small set of photos that repeatedly causes this high memory usage, it would be worthwhile to file a bug report and make the photos available to Adobe.

Report · Apr 07, 2025

I believe Windows and Mac OS manage memory differently. How much memory would preview generation of a large batch of files be expected to use on Windows?

Report · Apr 07, 2025

I did a quick test on my Windows 11 test machine with 784 random raws. Task Manager reported LR using up to 12 GB of memory, far less than 40+ GB you're seeing, and CPU utilization was consistently above 90%.

Report · Apr 08, 2025

What type of raw files, what CPU, and how much system memory were used? I suspect the size of the raw files would impact memory usage significantly, and the total amount of system memory would likely impact how much memory LRC chooses to use as a max limit. I ask about the CPU in case there could be some difference between Intel or AMD processors. My Canon R5 and R5II RAW files are 45MP, and my AMD Ryzen 9950x system has 64GB total RAM.