Support Board
Date/Time: Mon, 06 Apr 2026 07:38:33 +0000
Performance of Numbers Bars Study - Benchmarks?
View Count: 306
| [2026-02-17 01:28:51] |
| User273277 - Posts: 132 |
|
The Numbers Bars study consistently adds 25ms to my Average Calculation and Draw Time in the Performance tab of the Chart Settings window. The VBP study also runs at 10ms, but I'm less concerned about that one. I have seen other threads regarding Numbers Bars and overall Sierra performance improvements like these: Performance Problems with Numbers Bars Recommendations to Improve Overall Sierra Chart Performance I have tested and changed tick size, number of days to display, removed all unused studies, tested OpenGL on/off, and am using a stripped down Windows install. The machine has 24 CPU threads, 64GB ecc ram, NVMe storage. I have tested different Nvidia and AMD GPU's, though older generations. CPU performance is about 20% off the fastest single core geekbench. The CPU cores usually peak in single digit usage, and GPU hits 30% with this study and OpenGL enabled. I have other charts with dozens of all kinds of built in studies, alerts, spreadsheets, and custom compiled code, and they all run fast. ie. Charts with 10, 20, 50 studies, run at 6ms, 10ms, 15ms. If I run one chart, with just the Numbers Bars study, it runs at 25ms. This is at full screen 4k resolution. The draw time improves the smaller the window is. I would happily upgrade the CPU to get the last 20% of silicon performance, or the GPU if a generational upgrade makes a huge difference with these calculations. But seeing as the performance is isolated to 1-2 studies, and is 10x to >25x slower then any other study I use, this seems more like an optimization issue in the study code, or user control. Any ideas, or benchmarks of drawtime at 4k with the numbers bars study with different cpu/gpu combinations out there? |
| [2026-02-17 03:46:08] |
| Sierra_Chart Engineering - Posts: 23438 |
|
Numbers Bars involves a lot of text and text is simply inefficient. If you remove the text that will make a difference. The problem is the text, and the use of a 4K monitor. Sierra Chart Support - Engineering Level Your definitive source for support. Other responses are from users. Try to keep your questions brief and to the point. Be aware of support policy: https://www.sierrachart.com/index.php?l=PostingInformation.php#GeneralInformation For the most reliable, advanced, and zero cost futures order routing, use the Teton service: Sierra Chart Teton Futures Order Routing Date Time Of Last Edit: 2026-02-17 03:46:18
|
| [2026-02-19 07:50:08] |
| User273277 - Posts: 132 |
|
In the "Numbers Bars" study, I removed the column 1 and column 2 numbers by setting "Column 1 Numbers Bars Text" to "None", and also "Column 2 Numbers Bars Text" to "None". There was no change to the Average Calculation and Draw Time in the Performance tab of the Chart Settings window. It looks like hiding the text in the study still has the text being drawn in the background somewhere? If I hide the "Numbers Bars" study, the time drops significantly. |
| [2026-02-19 07:50:27] |
| User273277 - Posts: 132 |
|
Side note. If I understand correctly, depending on exactly which methods are being used, OpenGL text relies heavily on pixel/texture fill rates, and memory bandwidth. I generated a table with an Nvidia RTX 4060 as baseline, ordered by combined improvement. It looks like text drawing performance could be 3-6x better with the latest hardware. GPU Pixel Fill Rate (GPixel/s) Texture Fill Rate (GTexel/s) Memory Bandwidth (GB/s) Multiplier vs. RTX 4060 (approx.) RTX 4060 118 236 272 1× (baseline) RTX 5060 120 300 448 ~1× fill rates, ~1.65× bandwidth RTX 5070 Ti 235 687 896 ~2× fill rates, ~3.3× bandwidth RX 7700 XT 244 550 432 ~2.1× fill rates, ~1.6× bandwidth (effective higher w/ Infinity Cache) RX 9070 323 564 640 ~2.7× fill rates, ~2.35× bandwidth RTX 5080 293 879 960 ~2.5× fill rates, ~3.5× bandwidth RX 9070 XT 380 760 640 ~3.2× fill rates, ~2.35× bandwidth RTX 5090 424 1637 1792 ~3.6× fill rates, ~6.6× bandwidth I know this is a ridiculously niche question, but does this hardware performance difference match what users are seeing real world for numbers bars calculation at 4k? |
| [2026-02-20 22:29:25] |
| Sierra_Chart Engineering - Posts: 23438 |
|
We just did a quick test now, under our debugger, and we do see the draw time for Numbers Bars go down by about 50%, when not displaying text within Numbers Bars. You cannot have a very dense display of text with so many individual text drawings, and then remove it and it makes no timing difference. It has to make a difference. If it does not, well that would be good but it is not logical. Sierra Chart Support - Engineering Level Your definitive source for support. Other responses are from users. Try to keep your questions brief and to the point. Be aware of support policy: https://www.sierrachart.com/index.php?l=PostingInformation.php#GeneralInformation For the most reliable, advanced, and zero cost futures order routing, use the Teton service: Sierra Chart Teton Futures Order Routing |
| [2026-02-21 23:09:58] |
| User273277 - Posts: 132 |
|
I agree it should make a difference. Please see the attached video where I set both column 1 and 2 text to "None" in the Numbers Bars study, and there is no change to draw time. The chart is zoomed out to exaggerate the draw time. Attached is a second video where hiding the study results in the expected large draw time reduction. Maybe it is something I am doing? |
Private File Private File |
| [2026-02-23 16:28:08] |
| User273277 - Posts: 132 |
|
I updated video drivers. OpenGL rendering improved a lot. From 40-45ms to 15ms consistently. I have never seen such a tripling in performance from a driver update. I don't know if it's AMD, OpenGL, or the combination with the specific calls Sierra uses to render text, but using the latest drivers seems extra important here. |
To post a message in this thread, you need to log in with your Sierra Chart account: