Support Board
Support Board
Search Board
Control Panel
View My Posts / Threads
Direct MessagesDate/Time: Thu, 25 Apr 2024 05:17:40 +0000
[User Discussion] - The Ultimate PC of the year 2020 for Sierra Chart. CPU/GPU/CM/RAM/SSD
View Count: 12461
| [2021-12-24 19:10:07] |
| GravisHTG - Posts: 303 |
|
After all my experimenting I end up with a 5800X and I am using process lasso to distribut the cpu cores. I always dedicated 5/5 ( core/thread ) to SC and its ok , as long as I have GPU power. my GPU is a 1070ti. ( did not try better than this but I think SC always take all of it ) so the faster the better. Hope this can help. ( my next machine will be a laptop with at least 8/8 cpu and a 3060+ gpu ) And I will get a dedicated monster Desktop ( like AMD EPYC CPU or INTEL variant of it ) - but that is not in the short term. |
| [2022-01-27 22:08:02] |
| tdug - Posts: 124 |
|
Anyone try the bigger PNY GPS? Thinking of the PNY QUADRO RTX 4000??? I'm using about 30% of the GTX 8Gb1070ti and wondering if I should go for the 8gb pny 4000 Have changed my build to the i912900k CPU Alder lake and DDR5 64GV ram. But going for the PNY4000 would push the build cost up by 50% |
| [2022-01-27 22:53:40] |
| GravisHTG - Posts: 303 |
|
TDUG - we need to know if its worth it to spend on a Commercial grade GPU instead of the gamings one , Only testing will answer that. To go for a Quadro rtx 4000 instead of a normal RTX 2080-ti , i dont know if its worth it. As What Sierra chart use to calculate ? on the Benchmark part of things ( on gpu user bench ), RTX 2080 ti is 80% faster. even your 1070 ti is almost the same "speed" as the quadro. So , I would go for 6+ gb ram at least on the card. And for my next upgrade up from my 1070ti here, ill go to a 2080 ti or a 3080 ti. How is the new i912900k going btw ? Those CPU seems to be Monster cpu ! I just did some research , and unless Sierra chart use double-precision complex numbers for all calculation , a gaming gpu is far much better value for the money. Date Time Of Last Edit: 2022-01-28 07:12:10
|
| [2022-01-31 17:47:55] |
| tdug - Posts: 124 |
|
@GravisHTG Yep I agree on the PNY GPU./ I can get one for about 400 quid though. The 2080 looks nice. There was a delay with the prior 10980xe build. Cant get the STRIXX 299 Motherboard for another 4 weeks, so we went with a 12900k build 64GB of DDR5 4800RAM. So Im looking forward to getting that live and seeing how my ladders are on it. A 3080 would be nice but they are hard to get at reasonable money at the moment. |
| [2022-01-31 21:45:18] |
| RIFT - Posts: 124 |
|
I had a 3060ti. With 5 instances my gpu usage was 70-90% (when it was volatile). Now I'm using a 3080 and the gpu utilization is 30%. I think 3070 is a good choice to test if it is sufficient. I am now very satisfied with my PC (SC is finally running wonderfully): Ryzen 5900x (water cooling), 64 GB RAM (3800 Mhz CL16), PCI4 NVMe M2 SSD. Date Time Of Last Edit: 2022-01-31 21:59:24
|
| [2022-01-31 21:53:50] |
| GravisHTG - Posts: 303 |
|
Well I guess my next upgrade will be that Laptop with liquid cooling , 3080 Ti and the 12th gen Intel i9 I think this will be my best option for portability and efficacity. + the review on that new intel 12th gen laptop chip are very good. https://www.eluktronics.com/lpp Date Time Of Last Edit: 2022-01-31 21:55:35
|
| [2023-09-03 08:46:25] |
| tdug - Posts: 124 |
|
Coming back to this, I'm running more complex chartbooks now. 3 instances and noticing a small lag on the loading of chartbooks. Running the Intel extreme utility and getting 5.1ghz and I'm hitting 80 to 100% CPU usage sometimes. Think I'm going to upgrade to a 3080 gpu and see if that helps the chartbooks load up faster but the load seems to be on CPU mainly when loading the additional chartbooks. Ran through some advanced power settings for overclocking the i9 and it was running super fast but then crashed on a stress test. Anyone else on an i9 12900k? |
| [2023-09-03 10:14:30] |
| Dorian - Posts: 57 |
|
Hello, You must disable Hyper-Threading and E-cores in the BIOS. You can actually push your processor to 5.1 GHz with the XTU software, but only for P-cores. Disable the E-cores in the BIOS, it's important. This allows SierraChart to point to the P-cores and not the E-cores. The graphics card has nothing to do with it. I'm using an nVIDIA Quadro P600 and it runs at 10% on software. You can run SierraChart 3x software on the same computer. Instances don't work because it crashes graphics sometimes. I explained everything here 10x faster than 4 Instances !!! The 3 software programs must be divided into different processor cores. This is easily done with Task Manager. When you run the software 3x, they can be placed in the same processor core, which will make everything slow down, so you must put each SierraChart in different cores. You can also create shortcuts for each SierraChart .exe, and modify the shortcuts to ensure that each time you launch SierraChart, it goes to the correct CPU cores. This avoids having to make changes each time the software is launched. Everything is explained here https://trading-order-flow.fr/optimisation-sierra-chart/ You can find all the quotes for computers perfect for SierraChart here https://trading-order-flow.fr/workstation/ To translate into English, simply open the site on Google Chrome, then right-click then Translate into English. Date Time Of Last Edit: 2023-09-03 10:20:17
|
| [2023-09-30 01:01:08] |
| A77 - Posts: 8 |
|
One thing that admittedly has confused med Dorian is that I believe you said somewhere else that the base single core is what matters, not the overclock. Is this still true, or did I misunderstand? If base core is most wouldn't the ryzen 9 7900x be much better than intel i7 given that it has a base 4.7ghz? |
| [2023-09-30 22:20:11] |
| Dorian - Posts: 57 |
|
I said that we must prioritize the frequency of the processor, and not the number of cores. An overclock increases the base frequency. It's not a Turbo Boost. If the processor score is above 3500 in Single Threaded CPU, then the software will perform well. The AMD Ryzen 9 7900X should therefore perform well, but hyperthreading must be deactivated. https://www.cpubenchmark.net/cpu.php?cpu=AMD+Ryzen+9+7900X&id=5027 |
| [2023-10-01 08:20:46] |
| tdug - Posts: 124 |
|
Hi Dorian, are you saying you can get 5.1ghz stable running 3x instances? I optimised the core clocking in XTU with playing with power settings. It was running super super fast and smooth however when I ran a stress test it crashed the whole system. But the speed was incredible. So I might test diablo the e- cores as you said and assigning the specific cores for sierra. I'm on an i9 12900k Date Time Of Last Edit: 2023-10-01 08:21:30
|
| [2023-10-01 13:16:40] |
| Dorian - Posts: 57 |
|
According to the technical sheet of a 12900K, the P-cores can go up to 5.1 GHz and the E-cores to 3.9 GHz. Beyond these frequencies, we enter into pure overclocking, we must therefore touch on the voltage. Otherwise there is no need to touch the voltage, everything is automatic. Do not exceed these frequencies in XTU to have stability. https://ark.intel.com/content/www/us/en/ark/products/134599/intel-core-i912900k-processor-30m-cache-up-to-5-20-ghz.html Use only 2 RAM sticks with this processor and clocked at 4800 MHz (DDR5) or 3200 MHz (DDR4), no more. Everything is explained in this video. https://youtu.be/HOnCZnokdnA?si=NWsDNVOuiSWEnU4R Date Time Of Last Edit: 2023-10-01 13:20:36
|
| [2023-10-17 21:27:37] |
| User624595 - Posts: 176 |
|
For SC, is the base or max turbo boost clock more important? Are the PassMark single thread ratings using Turbo Boost, and are they accurately predictive of performance in SC? Is a 4060/i5 or 3060/i7 more effective? Any thoughts on something like these, or an alternative at this price ? https://www.costco.com/cyberpowerpc-gamer-xtreme-gaming-desktop---13th-gen-intel-core-i5-13400f----geforce-rtx-4060%2c-black.product.4000202717.html https://www.costco.com/ibuypower-slatemri7n3601-gaming-desktop---13th-gen-intel-core-i7-13700f----geforce-rtx-3060---windows-11.product.4000139335.html https://www.dell.com/en-us/shop/desktop-computers/alienware-aurora-r16-gaming-desktop/spd/alienware-aurora-r16-desktop/useahctor16i05 |
| [2023-10-18 09:56:03] |
| Dorian - Posts: 57 |
|
I run the software 3x, so that I have real-time information on all the graphics, this draws around 4 GHz on 2 software and 3 GHz on the last one. So you must favor the K version of the Intel processor and overclock the processor with the Intel XTU software. It is impossible to determine the true Passmark score, because some motherboards overclock the base frequency of all cores to the same turbo frequency (like ASUS Z590/Z690/Z790 STRIX or TUF series), and some people use very basic coolers that restrict the processor. I tested an i7-11700K, a W5-2465X, an i7-13700K. Sierra Chart always runs well with these processors, so you have to base yourself on the power of an 11700K and always take version K or X (for Xeon). Install Intel XTU to overclock all cores easily. You can also overclock the cores with any tool offered by ASUS in the STRIX or TUF version. But I noticed that the Single Thread score is indeed the real score to look at, because the test software will run a single core at full capacity, that is to say that even a crappy cooler will be enough for this test, so this reflects reality well. So to close the debate, a processor with a Single Thread score greater than 3500 will perform well for Sierra Chart with complex Studies. https://www.cpubenchmark.net/singleThread.html I have a basic graphics card (nVIDIA Quadro P600), it works well. The PCs offered do not have a processor with the K version, so it is not good. Date Time Of Last Edit: 2023-10-22 23:24:11
|
| [2023-10-18 10:05:51] |
| Dorian - Posts: 57 |
|
Turbo Boost will increase the frequency by 1 or 2 cores, so you must overclock it to the Turbo Boost frequency to have a stable and efficient processor. This is an interesting article. https://www.anandtech.com/show/16564/intels-new-adaptive-boost-technology-floating-turbo-comes-to-rocket-lake It's so complicated that it's better to push all the cores to the Turbo Boost frequency with the XTU software to have a guarantee that all the cores will push their frequency. |
| [2023-10-22 23:24:28] |
| Dorian - Posts: 57 |
|
But I noticed that the Single Thread score is indeed the real score to look at, because the test software will run a single core at full capacity, that is to say that even a crappy cooler will be enough for this test, so this reflects reality well. So to close the debate, a processor with a Single Thread score greater than 3500 will perform well for Sierra Chart with complex Studies. https://www.cpubenchmark.net/singleThread.html |
| [2023-10-23 21:34:17] |
| User624595 - Posts: 176 |
|
I run the software 3x, so that I have real-time information on all the graphics, this draws around 4 GHz on 2 software and 3 GHz on the last one. So you must favor the K version of the Intel processor and overclock the processor with the Intel XTU software.
It is impossible to determine the true Passmark score, because some motherboards overclock the base frequency of all cores to the same turbo frequency (like ASUS Z590/Z690/Z790 STRIX or TUF series), and some people use very basic coolers that restrict the processor. I tested an i7-11700K, a W5-2465X, an i7-13700K. Sierra Chart always runs well with these processors, so you have to base yourself on the power of an 11700K and always take version K or X (for Xeon). Install Intel XTU to overclock all cores easily. You can also overclock the cores with any tool offered by ASUS in the STRIX or TUF version. But I noticed that the Single Thread score is indeed the real score to look at, because the test software will run a single core at full capacity, that is to say that even a crappy cooler will be enough for this test, so this reflects reality well. So to close the debate, a processor with a Single Thread score greater than 3500 will perform well for Sierra Chart with complex Studies. https://www.cpubenchmark.net/singleThread.html I have a basic graphics card (nVIDIA Quadro P600), it works well. The PCs offered do not have a processor with the K version, so it is not good. Have you tested OpenGL with a higher end NVidia graphics card? That provided the biggest performance boost I have seen in SC, seemed more relevant than CPU to me. What was your experience with this? Is it definitive that even a 3 generation old 11700k because it's unlocked will perform better than a 13700f, which has a 25% higher single thread rating? |
| [2023-11-02 19:19:15] |
| User624595 - Posts: 176 |
|
@Dorian any comment on above?
|
To post a message in this thread, you need to log in with your Sierra Chart account: