Forums before death by AOL, social media and spammers... "We can't have nice things"
|    alt.comp.os.windows-11    |    Steaming pile of horseshit Windows 11    |    4,969 messages    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
|    Message 4,248 of 4,969    |
|    Paul to Carlos E.R.    |
|    Re: What on earth does TurboTax need Win    |
|    30 Jan 26 14:16:09    |
      XPost: alt.comp.os.windows-10, misc.taxes       From: nospam@needed.invalid              On Thu, 1/29/2026 3:56 PM, Carlos E.R. wrote:              > I was saying this for Dennis (he is not happy about buying a new machine       every five years),       > because it is a way of running W11 in a machine that doesn't have supported       hardware,       > like no TPM. It will of course run slow.              I wish we had better benchmarks for "slowness", just so we could       identify factors that make it faster.              Linux host-to-guest file I/O can manage 600MB/sec inside a VM.              The Windows VMs seem to be lower than that, host-to-guest.       It can be hard to tell if paravirtualization is being used       when the hosting software has settings like "Default".              On both Linux host (TMPFS) and Windows host (OSFMount), I can       have a RAMDrive for the VM container to sit on. This does not       do anything for I/O particularly, but it reduces the seek time       to zero. It behaves more like an SSD, when all you own is a       single slow HDD. But you need a lot of RAM to do that. And       in the current RAMpocalypse I can't really advocate for this       any more. When a Windows Guest boots and scans the shit out of C:       you hardly notice.              At one time, virtual machine file I/O was down       around 1MB to 2MB/sec or so. And the graphics drew       so slowly, you could see individual pixels getting       painted row by row. To say it is slow today, it's       nothing compared to how it was in early days. We were       running x86 OS on top of a SPARC instruction set.              A modern VM could have unaccelerated graphics. The driver       is wrong or very wrong. The CPU takes up the slack (MESA       is doing some of the work via software path).              The Windows MBEC support can degrade performance on older       than 10th gen CPUs. 8th gen CPUs sorta work. 4th gen CPUs       some feature might be turned off.              One virtualization product won't allow more than 2 cores. Silly.              In a lot of these cases, there does not seem to be a lot of       traction to fix it. You can use "PCI Passthru" to have a       second GPU dedicated to the virtual machine, and then the       driver is no longer driving virtual graphics, it is       driving real graphics. The odds of that working are       pretty close to zero :-) Just the fact my new computers       don't have a PCI slot, rules out using my spare-dummy card       for graphics.              At some point, it's just better to say "screw it, I'm       going physical" instead of virtual. And just install       Windows 11 besides Windows 7, using Rufus for the boot       stick preparation, and using the "Custom" install option,       declare a 200GB partition and just slap it in.              I'm averaging around a day each for these little projects,       just to give some idea what sort of time allocation to expect.       You might have a dozen tabs open in your browser, with recipes       to "fix this or that". The AI can at least make you aware       of stuff you need to fix, even if the recipes aren't complete.              And Duckduckgo is turning out to be a better search than Google       plague search.              And where would I be if I didn't have two computers ?       You can't be hardware poor and expect a quick result.               Paul              --- SoupGate-Win32 v1.05        * Origin: you cannot sedate... all the things you hate (1:229/2)    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
(c) 1994, bbs@darkrealms.ca