Support Board
Date/Time: Thu, 08 Jan 2026 01:50:12 +0000
Post From: OS Timers vs SC Timers discussion again... SC Timers chart update limitations
| [2026-01-05 23:57:24] |
| User719512 - Posts: 384 |
|
Sierra Engineering & Community, I previously experienced issues with Sierra Chart timers causing unresponsive charts when they were first released. I recently tested them again with version 2852, and the problem no longer occurs. My system configuration has remained on Windows 11 with the latest updates throughout. Current details include: - Windows 11 (25H2 - Build 26200.7462) - Processor: AMD Ryzen 9 5950X - Graphics: NVIDIA GeForce RTX 5070 Ti (previously RTX 2080) - OpenGL: Enabled While I cannot pinpoint the exact change that resolved the timer issue, I have since updated to the latest drivers, firmware for my Samsung drives, and BIOS. Additionally, I upgraded the drive hosting the Sierra Chart installation from SATA6 to an M.2 SSD (Samsung 990 PRO with Heatsink, 1TB), as fast markets and high data volumes could potentially create bottlenecks during disk writes, though this may not have been directly related to the timer problem. Over the past 1.5 months, I addressed some random system crashes (unrelated to Sierra Chart and possibly linked to Windows Update KB5066835) by making adjustments that could mitigate Ryzen-specific issues exposed by the update. My system is now stable again, and the Sierra timers function as expected. It's unclear if these changes are connected, but notable tweaks include disabling CPU throttling (C-states) in the BIOS, setting Windows Power Plan to "Best Performance," and configuring the minimum processor state to 100%. To monitor and optimize my system, I installed the following tools: - HWiNFO64 - LatencyMon - CrystalDiskInfo Are timer issues more prevalent on Intel CPUs, or primarily with AMD Ryzen? Could they be related to OpenGL, video card drivers, or latency factors? In conclusion, with very fast chart refresh intervals, I observe no discernible difference between Sierra timers and OS timers, at least not to the naked eye. To quantify this, I developed custom studies using QueryPerformanceCounter to measure study update and GDI DrawToChart frequencies. These confirm Sierra Chart's excellent performance. For instance: - A basic Depth of Market (DOM) study calculates in approximately 7.5 μs. - A Volume by Price study calculates in approximately 7.0 μs. On a single basic chart refreshing every 200 ms, the DOM updates at rates such as: - 10 ms interval: ~55 study updates/sec - 20 ms interval: ~28 study updates/sec - 50 ms interval: ~14 study updates/sec This level of performance is impressive, considering the overhead from chart drawing, quote data writing, and general OS tasks. I hope this feedback is helpful for troubleshooting similar issues. |