Copy link to clipboard
Copied
Hi.
Seit einiger Zeit (ab Lightroom Classic 11) läuft der Export recht schleppend. Auffällig dabei, die CPU-Auslastung der 12 Kerne liegt maximal und nur kurzzeitig bei ca. 40%. Früher lief die CPU während eines Exports mehrerer Fotos mit nahezu 100 %.
Vorab vielen Dank!
Hier die Rechner-Spezifikationen:
AMD Ryzen 9 3900X 12-Core Processor 3.79 GHz
NVIDIA GeForce RTX 2060 Super 8GB
System-SSD: Samsung 970 EVO Plus 500GB M.2
Daten-SSD: Samsung 970 EVO Plus 1TB M.2
Mainboard: ASUS Prime X570-P
Windows 11 22H2
Systeminformationen aus Lighroom kopiert:
Lightroom Classic-Version: 12.1
Lizenz: Creative Cloud
Spracheinstellung: de
Betriebssystem: Windows 10 - Business Edition
Version: 10.0.22621
Anwendungsarchitektur: x64
Systemarchitektur: x64
Anzahl logischer Prozessoren: 24
Prozessorgeschwindigkeit: 3,7GHz
SQLite-Version: 3.36.0
CPU-Auslastung: 0,0%
Integrierter Speicher: 65441,6 MB
Dedizierter GPU-Speicher, der von Lightroom verwendet wird: 4562,4MB / 8031,0MB (56%)
Für Lightroom verfügbarer phys. Speicher: 65441,6 MB
Von Lightroom verwendeter phys. Speicher: 9648,3 MB (14,7%)
Von Lightroom verwendeter virtueller Speicher: 15835,4 MB
Anzahl GDI-Objekte: 1156
Anzahl BENUTZER-Objekte: 4012
Anzahl Prozess-Handles: 3560
Cache-Speichergröße: 305,6MB
Interne Camera Raw-Version: 15.1 [ 1329 ]
Maximale Anzahl Threads, die Camera Raw verwendet: 5
Camera Raw SIMD-Optimierung: SSE2,AVX,AVX2
Virtueller Speicher in Camera Raw: 2376MB / 32720MB (7%)
Physischer Speicher in Camera Raw: -7353MB / 65441MB (-11%)
DPI-Einstellung des Systems: 96 DPI
Desktop-Komposition aktiviert: Ja
Standardvorschaugröße: 2560 Pixel
Monitore/Anzeigegeräte: 1) 2560x1440, 2) 1200x1920
Eingabetypen: Multitouch: Nein, integrierte Toucheingabe: Nein, integrierter Stift: Nein, externe Toucheingabe: Nein, externer Stift: Nein, Tastatur: Nein
Informationen zum Grafikprozessor:
DirectX: NVIDIA GeForce RTX 2060 SUPER (31.0.15.1694)
Anfangsstatus: GPU zum Exportieren wird standardmäßig unterstützt
Benutzerpräferenz: GPU zum Exportieren aktiviert
Anwendungsordner: C:\Program Files\Adobe\Adobe Lightroom Classic
Installierte Zusatzmodule:
1) AdobeStock
2) Flickr
3) Nikon Tether-Zusatzmodul
Config.lua-Flags: None
Adapter Nr. 1: Anbieter : 10de
Gerät : 1f06
Subsystem : 87a01043
Version : a1
Grafikspeicher : 8031
Adapter Nr. 2: Anbieter : 1414
Gerät : 8c
Subsystem : 0
Version : 0
Grafikspeicher : 0
AudioDeviceIOBlockSize: 1024
AudioDeviceName: $$$/dvaaudiodevice/SystemDefaultAndEffectiveDeviceName=System Default - Lautsprecher (Realtek(R) Audio)#{comment}DVAAU-4201250: Open the audio hardware preferences page.
AudioDeviceNumberOfChannels: 2
AudioDeviceSampleRate: 48000
Build: LR5x19
Direct2DEnabled: false
-------------------------------------------------------------------------------------
Post edited by Moderator to remove unnecessary system information
"Export time of 40 photos (RAW to JPEG) without using the GPU (disabled): 1:03 min. CPU stays at 100%
Export time of 40 photos (RAW to JPEG) GPU enabled: 30 sek. During export the CPU is about 30-40%, GPU max 39%."
Those results look normal -- 1.6 secs/photo without the GPU, 0.7 secs/photo with the GPU. On my 3.5-year-old Macbook Pro 8-core Intel i9 using an AMD Radeon Pro Vega 20 4 GB, I observe about 1.2 secs/photo, with roughly similar CPU and GPU usage.
When using the GPU, it may not be poss
...Copy link to clipboard
Copied
I'll help you figure this out. I'll need more info. The GPU driver on your machine is version 516.94 & the GPU is supported to be used at export. Please update the GPU drivers first & then follow these steps:
On Windows, the GPU info & Logs location is %APPDATA%\Adobe\CameraRaw. A shortcut to get to this folder on Windows is:
If the above steps don't help, share a screenshot of the export settings you are using, the type of image you are exporting & how much free storage you have in the C drive.
Thanks!
Sameer K
Copy link to clipboard
Copied
Hi Sameer K,
thank you for your support. Unfortunately it didn't solve the problem. The max CPU-workload during export still remains at around 30 - 40%, so the export needs more time than normal (screenshot attached).
GL_VENDOR: NVIDIA Corporation
GL_VERSION: 4.6.0 NVIDIA 528.49
GPUDeviceEnabled: false
OGLEnabled: true
Free storage at C-Drive: 331GB of 464GB
Export settings ataached.
Copy link to clipboard
Copied
By the way:
The export settings refer to .cr3 (Canon R5 RAW File) to jpeg.
Copy link to clipboard
Copied
[This post contains formatting and embedded images that don't appear in email. View the post in your Web browser.]
"User preference: GPU for exporting enabled"
"The max CPU-workload during export still remains at around 30 - 40%"
Your preferences have enabled the option Preferences > Performance > Use GPU for Export. Your screenshot shows CPU at 40% and GPU at 14%:
That's typical when LR uses the GPU for exporting -- CPU usage is reduced and GPU increased (but LR never seems to use more than 15-20% of the GPU for exporting -- it still has to use the CPU for much of the computation).
Use a watch or the computer clock to time the export of 40 photos, with and without the option GPU For Export enabled, and let us know the times. Sometimes LR is slower when it uses the GPU than when it doesn't.
Copy link to clipboard
Copied
Thanks John,
i did the test you recommended to me.
Export time of 40 photos (RAW to JPEG) without using the GPU (disabled): 1:03 min. CPU stays at 100%
Export time of 40 photos (RAW to JPEG) GPU enabled: 30 sek. During export the CPU is about 30-40%, GPU max 39%.
Printscreens attached.
Copy link to clipboard
Copied
"Export time of 40 photos (RAW to JPEG) without using the GPU (disabled): 1:03 min. CPU stays at 100%
Export time of 40 photos (RAW to JPEG) GPU enabled: 30 sek. During export the CPU is about 30-40%, GPU max 39%."
Those results look normal -- 1.6 secs/photo without the GPU, 0.7 secs/photo with the GPU. On my 3.5-year-old Macbook Pro 8-core Intel i9 using an AMD Radeon Pro Vega 20 4 GB, I observe about 1.2 secs/photo, with roughly similar CPU and GPU usage.
When using the GPU, it may not be possible for LR to use more CPU than what you observe (39%), because there could be other limitations, e.g. memory bandwidth with the CPU and GPU. But that's just a guess.