Hi all. I have a host VM (Ubuntu), that is capping out at a transfer speed of 100mbps,
The guest nic is detected as 1gbps, and i even tried setting
ethernet0.virtualDev = "vmxnet3"
to see if this made a difference, and while it is detected at 10gbps. the max throughput i can achieve is 100mbps.
The host machine is capable of much faster speeds.
I note the advice that throughput should be CPU limited but guest CPU is negligible, and host CPU is max 25% on any core.
Presumably id see 100% on at least 1 core if this were a CPU limit?
I also noticed that if I set a file transfer away sending from the guest to a file server at approx. 90 mbps then simultaneously run an internet speed test it will show 90 down/10 up. which really makes me feel like there is an artificial 100mbps limit somewhere.
Any ideas?
More details, please. There's no "artificial limitation" going on, but there's something else "throttling" your network speeds.
Workstation version?
Host operating system and version?
Host hardware?
Details on network configuration to whatever you're testing against.
Guest operating system version?
Bridged or NAT networking?
How are you measuring network speeds? What is the source of the transfer that's providing the measurement?
If you run the same (or similar) tests on your host, what's the speed you see there?
Not sure what's up here specifically, but I can definitely say we're not throttling anything (unless of course you're using Workstation Pro with the "bandwidth throttling" feature specifically...)
Attached screenshots taken with 1 Bridged and 1 NAT... the NAT ended up being faster as it happens, but they're both basically 1Gbps
I have a 14th gen i9 tho, so maybe in your case it's CPU bound.
It doesn't necessarily have to _peg_ the CPU for that IIRC.
Thanks for your efforts,
I take there's no significant CPU tax on NAT vs Bridged?
I don't think trying to bridge a WIFI connection worked when I tried.