- Joined
- Dec 6, 2002
When XP first came out I remember installing it on several systems with slow hardware with great success. One of the first computers I installed it on was a Pentium2 233 with 128megs of ram, and I remember thinking how well it ran on there. I had another system that was a Pentium3 533 with 128megs of rambus that saw several years of usage.
These days it seems like being on any system less than a ~2.4Ghz P4 with 1Gb ram is like pulling teeth. Even before loading the system up with programs, it will be sluggish for the first 20-30 mins after boot and any little program that runs in the background seems to nearly max the CPU. I’m not sure I could even imagine how slow it would be on some of the older systems that once successfully ran XP.
I understand if a 3+ year old system is slow running new programs, but it seems like a 3+ year old system is slow even just running 3+ year old programs, slower even than it was 3+ years ago. I mean like I’ll boot the system, click on Internet Explorer, and it won’t even pop-up for 5 minutes.
How much of the problem is just windows? I notice that the main issue with many older systems is SVCHOST using 100% CPU after boot for a good amount of time. This appears to be due to Microsoft/Windows Update/Automatic Updates. I’m not sure why exactly it needs to suck up the whole CPU while it scans for updates, or why it’s not smart enough to have the process at idle priority where it wouldn’t bog the entire system, but this seems to be a huge problem with older systems. Pentium3 systems tend to be completely unusable unless you disable automatic updates. Beyond just the SVCHOST issue, random windows services will randomly tax the CPU for seemingly no reason, often maxing it at 100% usage for up to several minutes. I don’t recall having any of these problems with XP back in 2001.
The computer that my father uses in his office seems to be immune from this issue despite being well beyond its prime. It’s a Dual Pentium3 Xeon 550Mhz (w/2Mb L2 Cache each) and 1Gb Reg PC100. It amazes me to this day that the system continues to be just as if not more responsive for office work than even some non-hyperthreading P4 systems I’ve worked on. This is part of the reason I think the “Windows services gone crazy” problems such as the SVCHOST issue is such a big part of the problem – despite his system only having 550Mhz processors, no process can really bog down more than one of them at a time vs. killing the one and only processor on a single-processor machine, even if it overall has more processor power.
Is this all just going to get worse and worse? I know that multi-threaded programs have yet to really take off, but what seems to have taken off is the assumption that all computers have extra virtual/real cores available so the amount of computing power random little background tasks take up has gone through the roof.
Building on the idea that SMP systems, even older ones, suffer less from these effects, would it be possible to somehow set it on a single-processor system where no process can use more than a certain percentage of the CPU?
These days it seems like being on any system less than a ~2.4Ghz P4 with 1Gb ram is like pulling teeth. Even before loading the system up with programs, it will be sluggish for the first 20-30 mins after boot and any little program that runs in the background seems to nearly max the CPU. I’m not sure I could even imagine how slow it would be on some of the older systems that once successfully ran XP.
I understand if a 3+ year old system is slow running new programs, but it seems like a 3+ year old system is slow even just running 3+ year old programs, slower even than it was 3+ years ago. I mean like I’ll boot the system, click on Internet Explorer, and it won’t even pop-up for 5 minutes.
How much of the problem is just windows? I notice that the main issue with many older systems is SVCHOST using 100% CPU after boot for a good amount of time. This appears to be due to Microsoft/Windows Update/Automatic Updates. I’m not sure why exactly it needs to suck up the whole CPU while it scans for updates, or why it’s not smart enough to have the process at idle priority where it wouldn’t bog the entire system, but this seems to be a huge problem with older systems. Pentium3 systems tend to be completely unusable unless you disable automatic updates. Beyond just the SVCHOST issue, random windows services will randomly tax the CPU for seemingly no reason, often maxing it at 100% usage for up to several minutes. I don’t recall having any of these problems with XP back in 2001.
The computer that my father uses in his office seems to be immune from this issue despite being well beyond its prime. It’s a Dual Pentium3 Xeon 550Mhz (w/2Mb L2 Cache each) and 1Gb Reg PC100. It amazes me to this day that the system continues to be just as if not more responsive for office work than even some non-hyperthreading P4 systems I’ve worked on. This is part of the reason I think the “Windows services gone crazy” problems such as the SVCHOST issue is such a big part of the problem – despite his system only having 550Mhz processors, no process can really bog down more than one of them at a time vs. killing the one and only processor on a single-processor machine, even if it overall has more processor power.
Is this all just going to get worse and worse? I know that multi-threaded programs have yet to really take off, but what seems to have taken off is the assumption that all computers have extra virtual/real cores available so the amount of computing power random little background tasks take up has gone through the roof.
Building on the idea that SMP systems, even older ones, suffer less from these effects, would it be possible to somehow set it on a single-processor system where no process can use more than a certain percentage of the CPU?