@Vlademil Klippet med kommentarer nämner redan att de jobbar på att förbättra UI. Mycket mer matnyttig info där.
Däremot så finns det något som folk bör vara medvetna om med 3.0. CIG har släppt tech specs för SQ42 nu och det är minimum 16GB ram. Även med 16GB ram så kan man få CPU spikes och heavy page file use om man har för lite RAM ledigt för spelet vilket gör det nästintill ospelbart. Har man inte Spel + system installerat på en SSD så bör man definitivt skaffa mer än 16GB ram. Så SC är oerhört RAM hungrigt. Även hastigheten på RAM påstås göra skillnad men det har jag ej sett tester på.
Här är ett bra test som gjorts på 3.0, med fulla servrar, med en testrigg med olika RAM mängder. Det bekräftar det som nämnts ovan. https://robertsspaceindustries.com/spectrum/community/SC/foru...
Själv har jag 32GB och precis som testet anger så är det fullständigt stabilt utan CPU spikes etc trots nätkoden.
Star Citizen Performance Test (RAM)
discussion
December 7th at 16:15
This test was designed to determine performance metrics of the Star Citizen game as it related to RAM amounts. It was limited only to RAM tests as the team wondered what effect RAM had on the game. The team consists of very enthusiastic players made up of 5 people, all working in Information Technology, with a diverse background in fields such as Network Administration, Network Engineering, Server system Design and Administration, Systems Hardware Configuration Specialist, and Performance Management Director. These tests were performed at the job site, with permission of the Department Director and were carried out with lab hardware and software that was released for the purpose of this test by the Department Director. The test was performed after hours, during non-peak time, and was performed by employees not on company time.
System
The tests were run by the same system with different RAM configurations.
CPU: Intel(R) Core(TM) i5-7600K CPU @ 3.80GHz
Base speed: 3.80 GHz
Sockets: 1
Cores: 4
Logical processors: 4
L1 cache: 256 KB
L2 cache: 1.0 MB
L3 cache: 6.0 MB
GPU: NVIDIA GeForce GTX 1060 6GB
Driver version: 23.21.13.8843
Driver date: 11/26/2017
DirectX version: 12 (FL 12.1)
Dedicated GPU memory: 6.0 GB
Shared GPU memory: 8.0 GB
GPU Memory: 14.0 GB
SSD: Interface: SATA III 6Gb/s
Max Sequential Read/Write Speed: 550/520 MB/s
Max Random Read/Write Speed: 90,000/90,000 IOPS
Controller: Phison
NAND Type: TLC
All RAM was of the same brand and model number to ensure a consistent test.
System Software Configuration
Windows 10 Enterprise Edition running the Fall Creators Update OS Build 16229.64
Clean install from USB drive provided directly by Microsoft and was not used previously.
Default Hard Drive File System Configuration on an internal SSD
Second internal SSD for dedicated installation of Star Citizen PTU Launcher and Star Citizen Game.
Both SSD’s used in the test were freshly installed from sealed packaging and never used previously.
No Anti-virus software installed. Built-in Windows Defender Disabled
Windows Firewall Enabled
Windows UAC Enabled – Set to default.
Windows Account role for testing: Administrators
Network
The System was connected to an isolated network consisting of 1 PC (test system), connected to a Ubiquiti EdgeSwitch -150W Which was connected to a Ubiquiti EdgeRouter X SFP. The EdgeRouter was fitted with a SFP+ Fiber module provided by Ubiquiti and a Fiber line connection provided by Bandwidth Solutions and is graded at OC3 (155.52Mbps). The test computer is connected to the EdgeSwitch with a CAT6 1Gbps.The Test computer was given dedicated priority (QoS) of no less then 75Mbps of bandwidth (guaranteed first in / first out priority) with the remaining 80Mbps available to the test computer, as needed, on standard priority.
Terminology
Interval(s): An interval is a measure of time equaling 1 second, the sample rate for this test.
CPU Spike: For the purpose of this test, a CPU spike is considered any event that occurs which results in the CPU increasing more then 30% between intervals. CPU Spike may or may not cause a pause.
Pause: Any event that causes input by the tester to stop responding AND the interface does not update to the next frame.
4GB RAM Test
Unfortunately this test could not be completed. It took over 15 minutes to get to the main menu and once there, the mouse pointer was so slow that it took almost 2 minutes to get to the crusader button. We decided to cancel this test as clearly, it was not going to perform well, if at all.
NOTE: When attempting to quit the game at the main menu, it took almost 10 minutes for the game to close.
CPU usage was at 100% the entire time sc.exe was running. Additionally, 98% of RAM usage was consumed and calls to the Windows paging file was extremely high. CPU spike tests could not be performed as the CPU was never less then 100%.
About 12 seconds after sc.exe successfully terminated, a Windows Blue Screen event occurred.
8GB RAM Test
This test was the first successful test to be performed. Playability was low in this state as every 10-15 seconds a pause was observed, lasting for 1-2 seconds each. While the test lasted for the entire test duration, it was hard to play. Flight was achievable but took a lot of skill. CPU Spike tests found that a spike occurred every 10-15 seconds for no longer then 1 second each. It was also observed that the CPU spike would occur when the paging file received a very high activity and may be linked to this event.
CPU usage was no less then 85% the entire time sc.exe was running. RAM usage was no less then 85% and calls to the Windows paging file was very high.
16GB RAM Test
This was the first test to be considered stable. Playability was moderate to high with pauses observed once every 1-4 minutes lasting an average of 1 second each. The longest pause was 3 seconds and the shortest pause was less then 1 second. It was found that during each pause, the paging file received a high activity level and may be linked.
CPU usage did not exceed 83%. At the lowest level, CPU usage was at 38% with an average of 45%. Calls to the Windows paging file was moderate to high.
24GB RAM Test
This test was the first test to not experience any CPU spikes. That being said, pauses were observed once every 5-10 minutes lasting no longer then 1 second each and averaging 639ms. Although it was observed that the paging file received traffic during a pause, it was considered low and did not exceed 10% of available IOPS. It is unlikely that pauses can be linked to any paging file activity.
CPU usage did not exceed 60% and averaged at 39%. Calls to the Windows paging file were low.
32GB RAM Test
This test was extremely stable. Playability was very high and pauses were only experienced twice in the 30 minute test cycle. Each pause lasted no more then 500ms and no paging activity was observed during the two pauses.
CPU usage did not exceed 52% and averaged 37%. Calls to the windows paging file occurred 7 distinct times. All 7 were batch calls.
64GB RAM Test
This was the first Extreme configuration test performed. Playability was very high a pause was only observed 1 time. Best guess was a small bottleneck issue on the network that caused the pause as network traffic spiked at that time.
CPU usage did not exceed 53% and averaged 39%. Calls to the windows paging file occurred 2 times and was in batching mode during both calls.
128GB RAM Test
This was the second Extreme configuration test performed. Unfortunately, the test was not successful as sc.exe would CTD during the loading screen to the main menu (before it would call full screen mode). This however, is not unusual as there are currently no games on the market capable of running with this amount of memory.
There are no CPU usage or paging file observations to report as this test was not able to complete.
EDIT: This was accidently missed when I posted this originally. In order to ensure a consistent experiment, we ensured that we were only running our tests on a completely full server (50 players). this was done by joining a group with a second lab account (another player in the lab) and having that person keep checking the "join group" screen which will show the number of players on the server as well as the maximum players permitted on the server, which at the time of this test was 50 players per server.
Case: Fractal Design Define S | Fans: 4x Noctua NF-a14 PWM140mm | Motherboard: Asus ROG Maximus XI Hero | CPU: i9-9900K | CPU-Cooler: Noctua NH-D15S dual fan | PSU: Corsair AX1500i | GPU: ASUS ROG STRIX GeForce RTX 3060Ti OC V2 8GB | RAM: 32 GB DDR4 Kingston HyperX Fury Black 2666mhz Cl15 | Storage: 1x Kingston Fury Renegade M.2 2TB SSD 1x OCZ Vector 180 480GB SSD, 1x OCZ Vector 512GB SSD | Screen: 24" Acer Predator XB240HA G-Sync 144hz | Sound Card: Asus Xonar Essence STX II | Aux: Cherry MX 3.0S Gaming Keyboard. EVGA X10 Carbon