Support Board
Date/Time: Sun, 02 Mar 2025 20:49:20 +0000
SC market data connections and SC instances
View Count: 1063
| [2019-06-09 21:10:38] |
| User12089 - Posts: 365 |
|
I am subscribed for 3 SC instances (ie N number of SC instances on 3 separate physical machines) - that works ok I am also subscribed for 3 simultaneous SC market data connections - however it seems that is limited to ONLY 3 SC instances (ie doesn't work like the above) each of which is directly connected to the SC Data feed and even when the 3 SC instances are running on the same physical machine and each SC instances uses (subscribes for) a different non-overlapping set of symbols Why is that? SC Data - All Services | Using username 2 | 2019-06-09 20:51:58.504 SC Data - All Services | Connecting to the server ds1.sierracharts.com. Port 443 | 2019-06-09 20:51:58.504 DTC Client socket (2) | Creating socket. | 2019-06-09 20:51:58.505 DTC Client socket (2) | Connecting to IP: 8.18.161.135. | 2019-06-09 20:51:58.505 SC Data - All Services | Connected to server. | 2019-06-09 20:51:58.616 SC Data - All Services | Starting socket receive thread. | 2019-06-09 20:51:58.616 SC Data - All Services | Sending encoding request to server: Binary VLS | 2019-06-09 20:51:58.616 SC Data - All Services | Setting DTC encoding to Binary VLS | 2019-06-09 20:51:58.738 SC Data - All Services | Sending logon request message. | 2019-06-09 20:51:58.738 SC Data - All Services | Received logon response. | 2019-06-09 20:51:58.878 SC Data - All Services | Server Name: SC Real-Time Server. | 2019-06-09 20:51:58.878 SC Data - All Services | Server protocol version: 8. Client protocol version: 8 | 2019-06-09 20:51:58.878 SC Data - All Services | Logon error received from server: Limited to one connection per username for the requested data source. | 2019-06-09 20:51:58.878 SC Data - All Services | Disconnected from the server. | 2019-06-09 20:51:58.878 SC Data - All Services | Data feed disconnected. Will reconnect in 2 seconds. | 2019-06-09 20:51:58.878 DTC Client socket (2) | Shutdown started. Waiting for graceful close. | 2019-06-09 20:51:58.878 DTC Client socket (2) | Socket gracefully closed by remote side. | 2019-06-09 20:51:58.987 DTC Client socket (2) | Closed. | 2019-06-09 20:51:58.987 |
| [2019-06-10 01:07:42] |
| |
|
- however it seems that is limited to ONLY 3 SC instances Yes, this is the case.There might be some changes related to this in the coming months. But only applying to CME and later EUREX data. If you want to share the data on the same system, then just use File >> New Instance to launch other instances which use that feed. 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: 2019-06-10 01:08:42
|
| [2019-06-10 11:34:41] |
| User12089 - Posts: 365 |
|
My main driver is not "sharing data", but achieving higher level of parallel processing, considering the single threaded nature of SC. (keep that in mind for the "upcoming changes") I got machines (laptops) with 32 GB RAM and 8 CPU cores each as most of that remains idle with one SC instance .... On top of that if I stuff that single instance with gazzilion of complex charts everything grounds to a halt in terms of chart refresh with real-time market data not to mention the responsiveness of the UI and also trading ie submitting orders etc from SC to IB So the winning infrastructure architecture to take advantage of modern hardware and overcome the single threaded SC software architecture seems to be: 1. Launch one SC instance with NO charts whatsoever - it will establish the market data connection to the SC data feed and trading/broker connection to e.g. IB TWS. 2. Then (install and) launch as many SC instances as your machine can support and stuff them with whatever charts you need and connect each of these instances as sub-instances to the one in point 1 above (I prefer separately installed SC instances instead of launching sub-instances from the main instance - keeps everything more tidy and manageable and decoupled) (reiterating - keep the above in mind during the "upcoming changes", we do need to take advantage of modern hardware or else make SC multithreaded ie every chart in a chart book runs on seeprate thread ) Date Time Of Last Edit: 2019-06-10 11:52:15
|
| [2019-06-10 23:29:54] |
| |
|
The only two things that run on the primary thread in Sierra Chart which takes a significant amount of CPU time are study calculations and graphics. Graphics is going to be moved to the GPU with OpenGL support we are working on. Market data processing, chart data loading, and other file operations are already done on background threads. Sierra Chart is already multithreaded. 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: 2019-06-10 23:30:25
|
| [2019-06-11 09:36:16] |
| User12089 - Posts: 365 |
|
well not quite (see below which is pasted from what I have written earlier in this post) - and that's with the latest SC version as of today. So ONLY running multiple instances of SC solves the below problems .... (And I have not heard anything about parallelizing study calcs as well as the parallel feeding with data of each chart and study calcs in your answer earlier. And separately, "OpenGL" support for GPU may help with visualization (separate from Study Calculations, but probably most likely with one chart at a time - here we have MANY charts each doing its own unique visualizations in parallel with the rest. As none of these vizualizations are particularly complex - GPU is for complex visualizations e.g. 3D animation silulations etc like e,g, in electronic games, physics simulations etc. As each of these simulations are run one at a time .... not at the same time on the GPU) ). You need to either leave the multiple instance of SC (which works quite well) or make sure the sudy calcs AND visualizations of EACH chart are truly parallel "My main driver is not "sharing data", but achieving higher level of parallel processing, considering the single threaded nature of SC. (keep that in mind for the "upcoming changes") I got machines (laptops) with 32 GB RAM and 8 CPU cores each as most of that remains idle with one SC instance .... On top of that if I stuff that single instance with gazzilion of complex charts everything grounds to a halt in terms of chart refresh with real-time market data not to mention the responsiveness of the UI and also trading ie submitting orders etc from SC to IB" Date Time Of Last Edit: 2019-06-11 12:51:31
|
To post a message in this thread, you need to log in with your Sierra Chart account: