Hey, Bill.
It sounds like you have thing pretty much figured out, then. That's
good. I do not know that much about how Partition Magic works, but I
have seen other partitioning utilities leave a tiny "blank" partition in
the front like you describe. Sometimes, I wonder if they do this on
purpose, or whether it is by accident. At least you have gotten around it.
You have the hard drives hooked up to a Promise PCI IDE controller?
Yes, that would explain the lack of DMA checkboxes for the hard drives
in the Win98 Device Manager. This is normal behavior. Many 3rd party
controllers do this. The controller will run the drives in DMA mode.
One possibility about the 30 second pause during the Win98 boot. It may
not be the hard drive. It may a network card searching for a DHCP
server on the network. If there is no DHCP server present, Windows 98
will keep searching for a while until it gives up. One way to test this
is the assign a static IP address to network card and see if it takes
less time to boot.
--Alex
bill wrote:
> Alex Zorrilla wrote:
>
>> In Windows 98, if you go to Control Panel --> System and look at the
>> Performance tab, does everything seem to be in order, or do you get
>> some message about "compatability mode"?
>>
>> Also, in the Win98 Device Manager, do you see both hard drives, or
>> only one? Do they have the DMA checkboxes checked? Does the Device
>> Manager show the correct IDE controller drivers? Also, what is the
>> "I:" drive? A CD-ROM? Is that working correctly?
>>
>> If you boot to the Win98 Safe Mode Command Prompt Only, can you access
>> the partitions correctly? What about from a boot diskette?
>>
>> --Alex
>>
>>
>
> Thanks, Alex,
>
> Looks OK today. Apparently, there is some kind of murky definitional
> disagreement as between MSDOS FDISK and Linux fdisk, on the one hand,
> and PowerQuest's Partition Magic on the other.
>
> For some reason, PM insists on inserting a single-cylinder "unallocated"
> first primary partition *in front of* what is supposed to be the first
> primary partion on my second drive, and that prevents MSDOS/Windows98
> (but not Windows 2000) from detecting it.
>
> The only way I could find to get rid of the damn thing was to go
> into Linux and use fdisk to delete both partitions from the second drive
> and then create the full-size primary partition that I wanted starting
> on cylinder 1. I then booted PM from floppies and formatted that
> partition ane restired the contects from backup. Everthing seems to be
> OK now except that it still takes Windows 30 seconds to come up compared
> to its "normal" 10 seconds on my PC.
>
> But at least I am back in business.
>
> Thanks again.
>
> Bill
> p.s. I don't see anything about DMA access for my IDE hard drives in
> Device manager. Could this be because I have them connected to a Promise
> Ultra100 controller?
>
>
>>
>>
>> bill wrote:
>>
>>> Alex Zorrilla wrote:
>>>
>>>> Kyle is right. Of course, there is also the issue of how much
>>>> memory you use on a constant basis. If you regularly exceed 255-256
>>>> MB, then it will still be worth adding the extra RAM, since even
>>>> uncached RAM is still faster than a swapfile on the hard drive.
>>>>
>>>> On the other hand, if your regular usage is only 128 MB, then adding
>>>> the extra RAM above 256 MB may actually slow you down a little.
>>>>
>>>> I believe the absolute limit for RAM on this motherboard is 768 MB.
>>>>
>>>> --Alex
>>>>
>>>>
>>>>
>>>> Kyle wrote:
>>>>
>>>>> "bill" <> wrote in message
>>>>> news:a6tcf.14016$q%. m...
>>>>> | Alex Zorrilla wrote:
>>>>> | > Hi, Mariano.
>>>>> | >
>>>>> | > Farmuse is right. It would help to know which motherboard you are
>>>>> | > using. If it is something like a VA-503+ (since that is the most
>>>>> common
>>>>> | > motherboard on this newsgroup), then the individual memory chips
>>>>> cannot
>>>>> | > be larger than 16 MB. For example, if you have a 256 MB DIMM,
>>>>> then it
>>>>> | > must consist of 16 chips, each of which is 16 MB. That would make
>>>>> it
>>>>> | > double-sided, with 8 chips on each side.
>>>>> | >
>>>>> | > Some motherboards have different limitations. As a general rule
>>>>> of
>>>>> | > thumb, if the motherboard can support only CPUs with an FSB of 100
>>>>> MHz,
>>>>> | > then the limit is usually 16 MB/chip. If the motherboard can
>>>>> support
>>>>> | > CPUs with an FSB of 133 MHz, then the limit is usually 32 MB/chip.
>>>>> If
>>>>> | > the motherboard can support only 66 MHz FSB, then it varies.
>>>>> | >
>>>>> |
>>>>> | Along these same lines, right now I have 2 128-MB ECC Pc100 sdram
>>>>> | sticks in my VA-503+ and I want to increase my total ram as much as
>>>>> | I can, but I seem to dimly recall reading somewhere that it is not
>>>>> | useful to put more than 384MB ram on this board. Is that correct and
>>>>> | , if not, what is the max useful total ram I can install?
>>>>> |
>>>>> | Bill
>>>>>
>>>>>
>>>>> 256 meg is the practical limit of the mobo cache (if it has a 1 Mbyte
>>>>> cache). There is technical information that reveals the "real" limit
>>>>> is 255 MB. However, a k6+ cpu will overcome this limit nicely caching
>>>>> the entire memory space. With the onboard k6+ CPU cache the
>>>>> performance is still good beyond the memory limitations of the mobo's
>>>>> memory caching subsystem.
>>>
>>>
>>>
>>>
>>> Ok, thank you gentlemen, I guess I'll stay with 256MB.
>>> Slamming into uncached RAM on linux is like running into a patch
>>> of concrete on an ice-skating rink.
>>>
>>> Bill
>>>
>>> p.s. New problem: I have had two hard drives for a long time,
>>> partitioned as follows:
>>>
>>> 1. C [E,F,G,H] J K
>>> 2. [D]
>>>
>>> [...] represents an extended partition.
>>>
>>> I have a dual boot setup:
>>>
>>> Windows 98 is on C:
>>>
>>> Windows 2000 is on J:
>>>
>>> (J: and K: are NTFS partitions)
>>>
>>> Windows2000 is no problem,
>>> but today (for the first time in a long time)
>>> I tried to boot Windows98.
>>>
>>> When I select Windows98 from the BOOT.INI menu,
>>> of course, I get the Windows98 boot menu,
>>> but when I select Normal from the Windows98 boot menu,
>>> it takes an ***extremely*** long time to bring up the DeskTop --
>>> and not only that, drive D is not detected, so Windows98
>>> interprets the boot drive as
>>>
>>> C [D,E,F,G]
>>>
>>> which screws up the registry.
>>>
>>> Does anyone have any clues about what might be causing this
>>> problem?
>>>
>>> Again, Windows2000 is working perfectly.
>>>
>>> Thanks again,
>>>
>>> Bill