Support Board
Support Board
Search Board
Control Panel
View My Posts / Threads
Direct MessagesDate/Time: Fri, 26 Apr 2024 14:06:02 +0000
[User Discussion] - Versions past 1981 have slow chart update
View Count: 1205
| [2019-10-11 15:07:30] |
| User852559 - Posts: 40 |
|
Hello, I have uploaded a video comparing side by side a version of Sierra Chart running pre 1982 vs post 1982. I have compared this in both Denali and CQG instances and the visual difference in chart updates is clearly slower in versions post 1982. The video is only comparing CQG and Denali, but I can confirm I have tested this solely in Denali versions and have isolated the same problem with Denali, so it is not limited to a specific data feed. This video comparison was made between the two data feeds just for login limitation purposes. I have tested all versions up to 1982 and past 1982 and 1981 is the last version which updates with the same speed as before. This speed is essential because as you can see sometimes the market is trading offer side and versions in post 1982 show the market trading bid side for half a second or more. This difference in speed is especially noticeable when the market is moving quickly. The difference is not noticeable on trading DOMs. This issue seems limited to charts. I have run the same comparison on blank charts with no studies and there is still a noticeable difference, so this is not related to the studies I am running on the chart. I have also limited the depth levels and calculations to 10 a side on the Denali feed and this made no difference either. I am using only one chart in the instance with a chart update interval of 100, number of days to load is 2, and my machine is more than capable. It is imperative that this issue be isolated and fixed as this difference in speed is very detrimental to short term trading. No settings were changed between comparisons between 1981 and 1982 when using Denali and there is still a noticeable speed difference. Please look into this and advise how to remedy this solution. Thanks! Edit: I forgot to label the video - The chart on the left is pre-1982 and the chart on the right is post-1982. Date Time Of Last Edit: 2019-10-11 15:25:15
|
Private File |
| [2019-10-11 17:15:25] |
| |
|
We really just recommend adjusting the Chart Update Interval. The updating speed depends on that. Refer to: Chart Settings: Chart Update Interval in Milliseconds (Chart >> Chart Settings >> Display >> Chart Update Interval menu) It could not be an issue issue with a particular version. That would not make any sense. That is not anything you will ever convince us of. We do not notice anything unusual in the video. But one thing to have a look at though, is check the CPU usage between the two versions and see if you notice higher usage when you have this "problem". 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, *change* to the Teton service: Sierra Chart Teton Futures Order Routing Date Time Of Last Edit: 2019-10-11 17:16:46
|
| [2019-10-11 18:07:21] |
| User852559 - Posts: 40 |
|
You recommend to not use lower than 100 for chart update interval and increasing it would only make it update slower correct? The CPU usage is the same between the different installations. I have tested this thoroughly and exhausted all troubleshooting measures you have outlined in your documentation. I have uploaded in this post a slowed recording of the same video above and highlighted with my mouse some of the discrepancies. The delay in post-1982 is slight but it is there nonetheless. I have isolated it to the update between 1981 and 1982. Nothing else has changed. Everything else has remained constant. Any version past 1981 this same thing occurs. Thanks. |
Private File |
| [2019-10-14 17:43:23] |
| User852559 - Posts: 40 |
|
Hello, I've uploaded 2 more comparison videos from Friday's 2:30 EST close on CL. This discrepancy in update speed is best visible during times of rapid movement, so this is a perfect opportunity to demonstrate what I'm attempting to explain in my original post. The first video is in real time and the second video is slowed down. The chart on the left is once again pre-1982 instance and the chart on the right is post-1982 instance. As you can see versions prior to 1981 are updating the bid/ask quantities much more frequently than versions post-1982. And the trades are being displayed quicker as well. Please do a similar comparison yourself this week where you watch an instance pre-1982 and post-1982 side by side during a period of rapid movement such as the 2:30 PM EST close on CL. Perhaps then you will be able to see this for yourself and look into the cause of the discrepancy. Thanks. |
Attachment Deleted. Private File Private File |
| [2019-10-14 18:11:44] |
| |
|
We plan to put together a basic test to look at this but there is nothing that has changed, which would cause any problem. We are quite certain there is nothing for us to resolve here and there is some other explanation in your particular case. Really that is 110% certain to us. And we also doubt we would be able to reproduce this either. Do not post any further information. We will get to this when we can. 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, *change* to the Teton service: Sierra Chart Teton Futures Order Routing |
| [2019-10-20 18:28:24] |
| User852559 - Posts: 40 |
|
Did you test this yet? Upon further research the delay between the pre-1982 and post-1982 versions is a consistent 300ms. Did you implement a background management of the chart interval that you had previously discussed? Perhaps this is in relation to this. Either way a 300ms difference is unacceptable especially when a key selling point of using the SC feed is low latency.
|
| [2019-10-21 11:33:45] |
| |
|
No we have not gotten to this. A proper test for it has to be developed. And also we do not believe we will be able to reproduce this.
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, *change* to the Teton service: Sierra Chart Teton Futures Order Routing |
| [2019-11-04 15:02:30] |
| User852559 - Posts: 40 |
|
I just wanted to touch base with you and let you know that after further troubleshooting and testing I have detected the likely source of the problem. I had a few Color Bar Based on Alert Condition studies applied to the charts in question. These had a number of bars to calculate setting set to something high like 50000 on a 10 second chart (number of days to load set to 2). Lowering this setting for these studies to 2000 has pretty much eliminated the problem. In versions prior to 1982 this was never an issue, so something seems to have changed between 1981 and 1982 that was causing this to be an issue regardless of the power of my machine (this was tested on a brand new build with top of the line components and dedicated m2 SSD). I just wanted to inform any users having a similar issue to consider this as a possible solution.
|
| [2019-11-04 20:51:19] |
| |
|
Well of course we knew that there was never a problem to begin with related to the chart updating. We did say that it is 110% certain previously. It is too involved to analyze your particular case and that is not something we would do. It is very unlikely there is anything wrong on the Sierra Chart side.
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, *change* to the Teton service: Sierra Chart Teton Futures Order Routing Date Time Of Last Edit: 2019-11-04 20:52:26
|
| [2020-04-15 19:57:22] |
| User104854 - Posts: 441 |
|
I too have had this same exact issue. I use a decent amount of color bars and whenever I upgrade to a version past 1978 the performance drops considerably. I've never tried 1981. Is there anything that has been adjusted since this update that would cause this? Is it possible to resolve this on a future update? Something was done since 1981 that affected the overall performance and it sounds like it had to do with the amount of bars that were inputted. If there are only 500 bars on a chart and the number inputted is 50,000 it didn't seem to matter with version 1978. Now it does matter. Please help. User852559, is there anything else that you have seen or done that has helped to resolve this issue since your last post? I really appreciate you posting this. Thank you. Date Time Of Last Edit: 2020-04-15 20:03:20
|
| [2020-04-16 07:16:51] |
| |
|
Please try the very latest version which is now 2086. We do make ongoing performance improvements. We think in regards to post #10, the problem in your case could be that the Color Bar Based on Alert Condition study is being based on studies Which are updating themselves at indexes earlier than the last updated chart bar, which is causing more calculation time. There was a problem under certain conditions that was resolved months ago where the starting bar index for calculations for a study was not always being set back to an earlier index as it should have been based upon the studies that it is dependent upon. So the behavior you have now is normal and you just have to work on improving the efficiency of your study configuration. And yes you do want to reduce the number of bars that the study is calculating over. This is important. 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, *change* to the Teton service: Sierra Chart Teton Futures Order Routing Date Time Of Last Edit: 2020-04-16 07:18:20
|
| [2020-04-17 17:53:00] |
| User104854 - Posts: 441 |
|
I've reduced the number of bars to calculate and it has improved quite a bit. I'm all for improving the efficiency of my studies. I don't understand what you mean when you say a color bar is based on studies which are updating at indexes earlier than the last updated chart bar. Do you have an example possibly of what you mean and what a possible solution might be? Thank you for your help. One last thing, I still get a lot of error messages that read 'color bar based on alert condition, out of bounds array condition at index 0' but it doesn't reference the actual color bar that it is referring to. Could this be a problem with the efficiency? Not sure how I would fix this. Date Time Of Last Edit: 2020-04-17 17:56:20
|
To post a message in this thread, you need to log in with your Sierra Chart account: