# Is it normal for the Silhouette OFX server to use 100% of one CPU core while idle? **URL:** \ **Category:** Flame Questions\ **Tags:** silhouette, borisfx, linux, flame\ **Created:** [August 1, 2026, 12:23am UTC](https://forum.logik.tv/t/is-it-normal-for-the-silhouette-ofx-server-to-use-100-of-one-cpu-core-while-idle/14729 "2026-08-01T00:23:53Z")\ **Posts on this page:** 6\ **Page:** 1
**Author:** ![totoxan](https://avatars.discourse-cdn.com/v4/letter/t/7ea924/32.png) [@totoxan](https://forum.logik.tv/u/totoxan)\ **Post date:** [August 1, 2026, 12:23am UTC](https://forum.logik.tv/t/is-it-normal-for-the-silhouette-ofx-server-to-use-100-of-one-cpu-core-while-idle/14729/1 "2026-08-01T00:23:53Z")
Hi everyone, I’m a Flame Linux user and I’m trying to understand whether this is expected behaviour or whether something is wrong with my setup. ### System - Flame 2027.1 - Rocky Linux 9.7 (Autodesk ISO) - DKU 21.0.1 - NVIDIA RTX 4080 16GB - NVIDIA Driver 580.159.04 - CUDA 13.0 ### What I observed If I open a project that contains **BorisFX Silhouette OFX nodes** , Flame immediately starts the following process: ```auto ``` ```auto /opt/BorisFX/SilhouettePlugin/silhouette -flush -server=SfxRndr-flame ``` The interesting part is that **I do not open the Silhouette editor**. I simply open the Flame project. However, the Silhouette process immediately starts consuming almost one full CPU core continuously. Example: ```auto ``` ```auto PID CPU 2440054 98-100% ``` `top -H` shows: ```auto ``` ```auto Threads: 7 total 1 running 6 sleeping silhouette 99.9% QDBusConnection 0% QXcbEventQueue 0% sentry-http 0% cuda thread 0% ``` So it appears that only the main Silhouette thread is busy. ### GPU usage GPU usage looks normal. ```auto ``` ```auto Flame GPU memory : ~1.2 GB GPU Utilization : 0-1% ``` There is no heavy GPU activity at this stage. ### Additional information I noticed this behaviour on both: - Flame 2027.0.1 - Flame 2027.1 I don’t remember seeing this behaviour with an older Silhouette release (before the newer UI/logo update), although I haven’t performed a controlled comparison yet. I have also experienced several Flame crashes recently while working on projects containing Silhouette nodes, but **I cannot confirm whether these crashes are related to this high CPU usage**. ### My question Is this normal behaviour for the Silhouette server? Does every Linux user see one CPU core at nearly 100% whenever a project containing Silhouette OFX is opened, even when the editor has never been opened? Could someone running Flame 2027.x on Linux with BorisFX Silhouette please check if they see the same behaviour? The same behaviour was observed on both Flame 2027.0.1 and Flame 2027.1 after updating to the newer Silhouette OFX release. I’d really appreciate hearing from anyone running Flame on Linux with BorisFX Silhouette. Thanks! ---
**Author:** ![philm](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.logik.tv/philm/32/9286_2.png) [@philm](https://forum.logik.tv/u/philm)\ **Post date:** [August 1, 2026, 5:28am UTC](https://forum.logik.tv/t/is-it-normal-for-the-silhouette-ofx-server-to-use-100-of-one-cpu-core-while-idle/14729/2 "2026-08-01T05:28:58Z")
Is it the only batch group on the desktop? Is this in an active batch group? A closed batch group? A saved batch group in a library? ---
**Author:** ![totoxan](https://avatars.discourse-cdn.com/v4/letter/t/7ea924/32.png) [@totoxan](https://forum.logik.tv/u/totoxan)\ **Post date:** [August 1, 2026, 6:23am UTC](https://forum.logik.tv/t/is-it-normal-for-the-silhouette-ofx-server-to-use-100-of-one-cpu-core-while-idle/14729/3 "2026-08-01T06:23:55Z")
Thanks for your reply. The Batch Group is **active** on the Desktop. The behavior I observed is: 1. Launch Flame. 2. Open a project. 3. Open an **active Batch Group** that contains one or more BorisFX Silhouette OFX nodes. 4. **Without opening the Silhouette editor** , the `silhouette -server=SfxRndr-flame` process starts immediately and continuously uses nearly one full CPU core (~100%). If I open a project or Batch Group that does **not** contain any Silhouette OFX nodes, the Silhouette process is not started. I haven’t tested the behavior with a closed Batch Group or with a saved Batch Group in a library yet, but I can test those as well if that would be helpful. ---
**Author:** ![MLandon](https://avatars.discourse-cdn.com/v4/letter/m/e495f1/32.png) [@MLandon](https://forum.logik.tv/u/MLandon)\ **Post date:** [August 1, 2026, 4:17pm UTC](https://forum.logik.tv/t/is-it-normal-for-the-silhouette-ofx-server-to-use-100-of-one-cpu-core-while-idle/14729/4 "2026-08-01T16:17:02Z")
You may want to report this to the BorisFX Forums. They would be very interested in seeing this. Here’s the link: > **[Silhouette](https://forum.borisfx.com/c/silhouette/31)** > > A community forum for users of Silhouette to interact together. Please share questions, feature requests, and creative examples of your work. Personally, I haven’t encountered this, but I have experienced issues with the 2026 Silhouette release using the ML roto tools on over 4k shots over 100 frames long. Marco and the team are working with me on resolving this issue. ---
**Author:** ![totoxan](https://avatars.discourse-cdn.com/v4/letter/t/7ea924/32.png) [@totoxan](https://forum.logik.tv/u/totoxan)\ **Post date:** [August 1, 2026, 7:12pm UTC](https://forum.logik.tv/t/is-it-normal-for-the-silhouette-ofx-server-to-use-100-of-one-cpu-core-while-idle/14729/5 "2026-08-01T19:12:14Z")
Thanks for the suggestion! I wasn’t aware that BorisFX has its own forum. I’ll keep investigating this and will also consider posting my findings there if the behavior remains reproducible. I really appreciate your advice. Thanks! ---
**Author:** ![totoxan](https://avatars.discourse-cdn.com/v4/letter/t/7ea924/32.png) [@totoxan](https://forum.logik.tv/u/totoxan)\ **Post date:** [August 3, 2026, 11:36pm UTC](https://forum.logik.tv/t/is-it-normal-for-the-silhouette-ofx-server-to-use-100-of-one-cpu-core-while-idle/14729/6 "2026-08-03T23:36:00Z")
**Update** I contacted the Boris FX team and received clarification from Paul Miller (Silhouette lead developer). The `silhouette -flush -server=SfxRndr-flame` process is expected behavior. It is the shared Silhouette rendering server used by all Silhouette OFX instances and remains alive until Flame exits. However, Paul also mentioned that although the server continuously polls for work, **it should not normally appear as consuming 100% of one CPU core** , and he will investigate that behavior. Based on this information, it seems the helper process itself is not likely to be the cause of the crashes I previously experienced. Since upgrading to Flame 2027.1, stability has improved significantly on my system. GPU memory management also appears much better than in 2027.0.1. At this point, I suspect my remaining crashes may be unrelated and perhaps tied to a specific Batch setup or another issue rather than the Silhouette server itself. Thanks everyone for the suggestions.