Running MultiCharts In A Virtual Machine
Posted: May 09 2019
I'm considering switching from Windows to Linux, and I'm hoping experienced Linux/VM users will chime in with some sage comments. I did a search and the last relevant discussion on this topic was 2014, maybe we can do an update on that. My biggest concern is the speed at which MC will run on a VM, and if there will be any impact on the real-time IQFeed data.
I started this process by first converting my Lenovo Thinkpad X1 Carbon Gen3 (i7-5600 2.6GHz 4 core, 8GB RAM, 250GB HD, Intel HD Graphics 5500) to Ubuntu 18.04. I've run Windows 10 with both VMWare and Gnome Boxes for VM software. Multicharts will run on both, but of course it is slow on this little laptop. I noticed jerky mouse movement and slightly slow screen refreshes. But in my very informal tests, the quote data seemed to come through just as fast as a non-VM version of Windows.
At some point I will convert my workstation to Ubuntu and see how it does. It is a much beefier machine (Asus WS X299 PRO/SE motherboard, Intel i9-7900 3.3GHz, 10 cores, 20 logical processors, 64GB RAM, Samsung 500GB 960 EVO M.2 SSD, nvidia quadro P2000). I would expect it to be much faster, but have no idea what performance change to expect.
My questions for others are:
1) Have you done this, and what was your experience?
2) Should I be considering a PCI passthrough setup to get native performance?
I started this process by first converting my Lenovo Thinkpad X1 Carbon Gen3 (i7-5600 2.6GHz 4 core, 8GB RAM, 250GB HD, Intel HD Graphics 5500) to Ubuntu 18.04. I've run Windows 10 with both VMWare and Gnome Boxes for VM software. Multicharts will run on both, but of course it is slow on this little laptop. I noticed jerky mouse movement and slightly slow screen refreshes. But in my very informal tests, the quote data seemed to come through just as fast as a non-VM version of Windows.
At some point I will convert my workstation to Ubuntu and see how it does. It is a much beefier machine (Asus WS X299 PRO/SE motherboard, Intel i9-7900 3.3GHz, 10 cores, 20 logical processors, 64GB RAM, Samsung 500GB 960 EVO M.2 SSD, nvidia quadro P2000). I would expect it to be much faster, but have no idea what performance change to expect.
My questions for others are:
1) Have you done this, and what was your experience?
2) Should I be considering a PCI passthrough setup to get native performance?