I know VMs aren't officially supported, but I saw how diligently the Slysoft folks tackled previous issues, so I figured it was worth at least asking.
I've encountered a really weird performance problem. I assume it's related to my running in a VM, since I haven't seen any other reports about it.
Essentially, enabling AnyDVD HD causes BD read performance to get stuck below 2.8X (often much worse)...but only if I let the BD drive spin down after the initial boot & AnyDVD scan.
If I start benchmarking right after boot, ripping starts at 3.37X and increases as expected. And if I start another benchmark before the drive has spun down, it stays fast. I've tried that up to 5 times, and it consistently works fine...until the drive spins down.
Other weird details:
- rebooting the VM is sufficient to "fix" it -- I don't have to reboot or power-cycle the entire machine.
- disabling and re-enabling (or quitting and rerunning) AnyDVD HD doesn't fix it.
- disabling AnyDVD causes the read speed to be normal as long as AnyDVD is disabled.
- Windows XP SP3 is always stuck slow.
- when things are stuck slow, one CPU seems to be much more active (~60% according to Task Manager) than normal (~30%). The other CPU is always ~10%.
I'm currently testing Windows XP SP1 in VMware Player 4.0 running on Ubuntu 10.04. (Windows Server 2003 also exhibits the same behavior.) Most of my benchmarking is done using Opti Drive Control 1.51 (to eliminate write speed as an issue). I did verify that this happens when ripping to ISO as well, though.
Things I haven't yet tested:
- VMware Player 3.1.5 (latest of the 3.x series)
- VirtualBox
- VMware ESXi (bare metal hypervisor)
- Windows running natively
I've attached a log from when things are fast as well as two from when things are slow. There may be something useful in the traffic files (which are noticeably smaller in the slow versions), but I realize in retrospect the logs weren't as controlled a contrast as they could be (e.g. different reboots of the VM).
If these provide enough information, great. If not, I'm happy to do a run of sequential tests in a single boot to eliminate some of the variables. Or if I can answer any questions, please let me know!
I've encountered a really weird performance problem. I assume it's related to my running in a VM, since I haven't seen any other reports about it.
Essentially, enabling AnyDVD HD causes BD read performance to get stuck below 2.8X (often much worse)...but only if I let the BD drive spin down after the initial boot & AnyDVD scan.
If I start benchmarking right after boot, ripping starts at 3.37X and increases as expected. And if I start another benchmark before the drive has spun down, it stays fast. I've tried that up to 5 times, and it consistently works fine...until the drive spins down.
Other weird details:
- rebooting the VM is sufficient to "fix" it -- I don't have to reboot or power-cycle the entire machine.
- disabling and re-enabling (or quitting and rerunning) AnyDVD HD doesn't fix it.
- disabling AnyDVD causes the read speed to be normal as long as AnyDVD is disabled.
- Windows XP SP3 is always stuck slow.
- when things are stuck slow, one CPU seems to be much more active (~60% according to Task Manager) than normal (~30%). The other CPU is always ~10%.
I'm currently testing Windows XP SP1 in VMware Player 4.0 running on Ubuntu 10.04. (Windows Server 2003 also exhibits the same behavior.) Most of my benchmarking is done using Opti Drive Control 1.51 (to eliminate write speed as an issue). I did verify that this happens when ripping to ISO as well, though.
Things I haven't yet tested:
- VMware Player 3.1.5 (latest of the 3.x series)
- VirtualBox
- VMware ESXi (bare metal hypervisor)
- Windows running natively
I've attached a log from when things are fast as well as two from when things are slow. There may be something useful in the traffic files (which are noticeably smaller in the slow versions), but I realize in retrospect the logs weren't as controlled a contrast as they could be (e.g. different reboots of the VM).
If these provide enough information, great. If not, I'm happy to do a run of sequential tests in a single boot to eliminate some of the variables. Or if I can answer any questions, please let me know!