You are not logged in.
So, this may be due to a variety of issues, but I wanted to document my experience with AL32 and two old motherboards.
My question is for the second motherboard, has anyone seen this behavior?
The first motherboard has more widespread instability issues, and may be due to CPU voltages (the capacitors around the cpu are all bulging and may have leaked).
In windows 98, I get as far as the login sometimes, but it always hangs. With the AL32 live image, it will sometimes hang on the grub splash screen, the memtest, or while trying to boot the system. Oddly, at one point I got about 1 hour into a memory test before it hung.
The motherboard is a ECS Elitegroup K7SEM and the processor is an AMD Duron D700AUT1B at 700MHz. I have also had issues with RAM (errors reported in MemTest), but I have 2x256MB sticks installed.
The second motherboard is a 2006 PC Chips M825G. The CPU is the same but I have a single 512MB stick of RAM.
This one is newer and works perfectly fine for my tests with KolibriOS.
Windows 98 and 2000 setup disks can't seem to get to the installation process... *edit* I should mention this may be an issue with the IDE drive and keyboard I have hooked up. KolibriOS worked perfectly.
This board will not launch the grub bootloader correctly as it displays, but then immediately hangs. More specifically, the boot screen is displaying indefinitely, but the keyboard (and mouse) do not interact with it, as far as even toggling caps-lock.
Why this board is not allowing even the bootloader to proceed is baffling me. Perhaps the AL32 disk is just too new overall for a Duron processor.
tl;dr: old technology is annoying
Last edited by atg (2020-11-12 03:56:23)
Offline
IIRC the old iive images at the very least uses syslinux for the boot menu, not grub. I couldn't get grub to run on this machine either, so I looked to see what the iso was using and discovered it was using syslinux, so I used that. That might have changed more recently, but I've not noticed yet.
Architecture: pentium4, Testing repos: Yes, Hardware: EeePC 901+2GB RAM+OS half on the SD card.
Offline
As I read the wikipedia page, this processor does not have sse.
However, you need sse for the i686 branch.
So you can only try the i486 branch, which unfortunately misses out most of the graphics packages, is hard to install (because we do not have a current install iso), but otherwise runs fine.
Offline
IIRC the old iive images at the very least uses syslinux for the boot menu, not grub. I couldn't get grub to run on this machine either, so I looked to see what the iso was using and discovered it was using syslinux, so I used that. That might have changed more recently, but I've not noticed yet.
I am being silly, of course it is syslinux. Sorry about that!
As I read the wikipedia page, this processor does not have sse.
Thanks very much for pointing that out to me! When I am free I will attempt an older release / see if I can check the processor flags with one of the OS's I can get to boot.
RE: The PC Chips board:
As a follow up to my original post, I got a little further with Windows 98, although 2000 is still giving me trouble.
With AL32 I can now get past the boot menu and into memtest, but the kernel still doesn't load. I switched some of my hardware: cdrom, hdd, keyboard. That seems to be what win98 and syslinux was mad about.
I will look into this further with the info provided here. Thanks again, everyone.
Offline
FWIW, capacititors near a chip are normally used either as a low pass filter to block voltage spikes and dips, or more generally to keep the chip powered when it suddenly asks for more power than the external circuitry can supply. Electrolytic caps I would suspect would generally fail open if the electrolyte drains, which in most basic applications of those two techniques wouldn't stop voltage getting through, but would be more liable to brownouts and spurious errors.
Architecture: pentium4, Testing repos: Yes, Hardware: EeePC 901+2GB RAM+OS half on the SD card.
Offline
FWIW, capacititors near a chip are normally used either as a low pass filter to block voltage spikes and dips, or more generally to keep the chip powered when it suddenly asks for more power than the external circuitry can supply. Electrolytic caps I would suspect would generally fail open if the electrolyte drains, which in most basic applications of those two techniques wouldn't stop voltage getting through, but would be more liable to brownouts and spurious errors.
Yeah absolutely. I noticed they were bulging and I figured I would have to re-cap it especially with the instability. Booting would presumably be more erratic than just a memory test, so the spikes in voltage make sense.
I'm going to re-cap it, but in the meantime I got a newer board in the PC Chips board which doesn't seem to have /those/ issues. (I have a rather large list of components I need to order atm, so in lieu of paying twice for shipping I figured I would get a known-good board as well)
Also, thanks for corroborating the notion of unstable voltages !
Offline