595 reputation
412
bio website lightspeed.ca
location
age
visits member for 5 years, 3 months
seen Feb 12 at 16:44

Aug
17
comment Windows XP running as a guest on Linux VirtualBox is very slow
@kyle: Post that as an answer. It appears to be part of the solution.
Aug
17
comment Windows XP running as a guest on Linux VirtualBox is very slow
Disabling Hyperthreading in BIOS and switching to the Non-OSE version appears to have the desired effect. Windows Task Manager shows that the CPU usage is significantly less than before, both when idle and under load. Disk IO never seems to have been a problem.
Aug
17
comment Windows XP running as a guest on Linux VirtualBox is very slow
@MattH: it also makes it so that you can assign multiple CPUs to a guest OS. That may not actually provide better performance, but it's a feature that's disabled for me.
Aug
16
comment Windows XP running as a guest on Linux VirtualBox is very slow
I've also ensured that VirtualBox OSE host software and the Windows XP guest additions are the same version.
Aug
16
comment Windows XP running as a guest on Linux VirtualBox is very slow
Nope. Loading the database over the network is definitely slower (I timed a particularly slow search with a stopwatch for each test case). Filemaker also complains that it would be slower if I load its files over the network, and that rings true.
Aug
16
comment Windows XP running as a guest on Linux VirtualBox is very slow
Ok, I'll give this a try, but that sounds overly complicated for a situation that requires simplicity. Windows really shouldn't need to swap to disk in my situation, since I've allocated it 2 GB of RAM and the WTM says it uses all of 101MB of PF usage. Physical memory shows that 1.5 GB of RAM is free. If Filemaker really wanted to, it could load the entire 750 MB database into RAM.
Aug
15
comment Windows XP running as a guest on Linux VirtualBox is very slow
Actually, there was a bit of detail in the discussion that you may have missed. These CPUs are the old, first generation Xeon processors, model 2 according to /proc/cpuinfo. They're single-core but multithreaded, giving the kernel the impression that it has 4 CPUs to work with. Also, the original server has 1 single-core Pentium 4 cpu, but runs much faster with a native install of XP. In my experience with this Filemaker database, its performance is quite dependant on CPU speed, and your suggestion not to run Windows guests without VT is probably exactly where the problem is.
Aug
9
comment Windows XP running as a guest on Linux VirtualBox is very slow
@mailq let us continue this discussion in chat
Aug
9
comment Windows XP running as a guest on Linux VirtualBox is very slow
@mailq: Yes, Virtualbox Guest Additions has been installed. Why would Ubuntu server edition make a difference? And /proc/cpuinfo says nothing of VT-x, and neither does the BIOS.
Aug
9
comment Windows XP running as a guest on Linux VirtualBox is very slow
@Shane Madden: Well, if you have any suggestions for changing the situation, I'm all ears. :)
Aug
9
comment Windows XP running as a guest on Linux VirtualBox is very slow
@zoredache: I don't think it's disk IO that's the problem, and one CPU gets maxed out when the Filemaker server is properly busy. But at the same time, the CPU being used changes nearly constantly, which I find kind of funny.
May
4
comment Why aren't there many monitors with higher than 1920x1080 resolution?
I think that the best response to this is to get two smaller HD monitors for multitasking. My 22" 1080p monitor at work is almost wide enough to have two full windows side-by-side, but I had a lot more horizontal resolution on the two 17" CRTs that this thing replaced.
Feb
11
comment Is IPv6 'faster' than IPv4?
Oh don't worry. Before the end of this year, the whole world will see the value of IPv6. In spades. And on the front page of every newspaper.
Feb
10
comment Is IPv6 'faster' than IPv4?
You sure can. If you wanted to, you could assign 2000 IPv6 addresses to every square meter of the disc of our galaxy. There are 2^128 possible IP addresses in this scheme, or over 3x10^38. This is more than a billion billion times the total number of IPv4 addresses. You could even assign IPs to every single component of every single household item ever made in the entire history of humanity, until the end of humanity itself.
Feb
10
comment How to avoid exposing my MAC address when using IPv6?
I think when Arin says "customer" they mean "ISP". Any ISP (including very, very large ones) can allocate a single /64 for their entire network and be done with it. No further routing required. But allocating blocks of IP addresses numbering in the trillions to joe-average residential customers is downright foolhardy.
Feb
10
comment How to avoid exposing my MAC address when using IPv6?
eh, what? I can't even think of a reason to assign a /64 to a single residential client (beyond autoconfiguration, and even that is pointless), let alone a /48. They say that the only possible reason for IP exhaustion in IPv6 would be astonishingly poor address allocation, and it looks like your ISP qualifies.
Feb
10
comment How to avoid exposing my MAC address when using IPv6?
A single /48 will be sufficient for every last customer they have until the end of time, regardless of growth (2^80 hosts!). While this might seem like overkill, it certainly makes routing simple. A /96 network would be sufficient to provide IP addresses for the entire IPv4 internet. Any network smaller than a /64 wouldn't allow MAC autoconfiguration.
Jun
9
comment Windows XP doesn't like small subnets?
Actually, it's not the IP address that's the problem, it's Windows XP. Using the IP address 65.110.7.22 and the subnet mask 255.255.255.252 gives the same error message. But it works on another computer. It's probably a difference in the XP updates.
Jun
9
comment Windows XP doesn't like small subnets?
You're right. I was looking at the wrong line on my subnet table. :)
Feb
19
comment How does a sound card determine if headphones are plugged in?
cough Actually, it's more to the effect that the plug-in acts as a switch. When something is plugged into it, it completes the circuit (just like a light switch) and a) your headphones work, and b) the sound card knows it's plugged in by the fact that the circuit is complete. You don't have to muck around with measuring the resistance of the connection - that's complicated. Mostly, the fact that the computer knows at all is in software and drivers.