Window7 install on dodgy PC

[Bob I]

Yes, I think you're probably right. XP might have been able to run OK in
256M (I never played much when XP originally came out), but by the time
of SP2 and SP3, I suspect 512M was minimal, and 1G what was really
required.

The machine in question came with SP2, and decided to upgrade to SP3 when I
was staying at a B&B in a strange town, and I wanted to read my e-mail quickly
using their WiFi before we packed up and left. So our five minute departure
time stretched to four hours while we waited forr it to finsih downloading and
installing it. We were lucky they didn't charge us for another day's stay.

That slowed the machine considerably, but so did the new versions of other
programs.

Regarding the startup and shutdown times you provided, I agree with
what Ken said above. Disk swapping will seriously slow things down,
but slowness to that degree implies that something was seriously
wrong.

Agreed. Not necessarily malware, but certainly some error in
settings/configuration. (Or just conceivably a failing hard disc: modern
ones have error-correction circuitry that can conceal problems for a
very long time, but make the disc [seem to] run exceedingly slowly.
Rare, but I've seen it at least once: hardly ever showed a single error,
but the machine was taking a quarter of an hour to boot, and also slowed
to a crawl whenever it accessed the disc, though was fast enough when it
wasn't doing so.)

Changing the display to use fewer fancy graphic gizmos helped a little, but
not much.

Another thing that will cause that, is having a string of read errors
will trip the driver to revert to PIO access from DMA. That will result
in a very noticeable performance decline in areas dominated by drive
access. Reasserting DMA access quickly corrects that issue. Another one
was caused by some useless features of "Norton Antivirus" back then and
you would wait forever after start up for it to finish loading/scanning
before the PC would have acceptable performance. Not sure if Symantic
had released some update that just pushed a 256MB box over the edge or
512MB covered up the disk thrashing better. But tossing Norton fixed
the problem there.

 

[Justin]

The original OS was Vista (shudder) Which is why (I think) BF installed
Win7. Yeah. I could get an OEM, or upgrade, or other workaround, but
didn't think that it was worth the hassle. I'd prefer to just start again.

Bramblestick

Forget about installing Vista and then installing that upgrade version of 7.
Wipe the drive, start installing Windows 7, and when it asks for the
key, quit, reboot and boot from the Windows 7 DVD. Choose the upgrade
option, and proceed normally.
All the install wants to do is see if there is some incarnation of
Windows already installed. It can be Windows XP, Windows Vista, or even
a botched install of Windows 7.

If you don't want to deal with this bullshit constantly, get a Mac.
Good prices for students and if that isn't an option there is a
refurbished section of the Apple store with cheaper machines with full
AppleCare.

 

[Gene E. Bloch]

The machine in question came with SP2, and decided to upgrade to SP3 when I
was staying at a B&B in a strange town, and I wanted to read my e-mail
quickly
using their WiFi before we packed up and left. So our five minute departure
time stretched to four hours while we waited forr it to finsih downloading
and
installing it. We were lucky they didn't charge us for another day's stay.

You have given me an idea, which I'll share.

Before going on a trip, configure Windows Update to "Notify and let me
choose..." when to install updates. Your tale is only one reason to do
that, since on various WiFi hotspots, one might worry about security
and corruption issues as well.

At least you didn't miss an airplane flight

BTW, I always have Update configured to notify and let me choose when
to update, even on desktop machines.

 

[Gene E. Bloch]

It has for some time struck me as odd anyway that we continue to use the - 8
bit - byte anyway, for RAM at least: RAM modules are many bytes wide.

In the Intel family of processors, the size of a byte has to do with
the architecture of the CPU and its instructions, and with how data is
accessed, not the width of the data path.

Instructions are built up in units of 8-bit bytes, and they address
memory in the same units.

Character codes and data transmission protocols also are based on
eight-bit units.

Most every processor I've dealt with in the last few decades is the
same way. It's been a long time since I've seen a machine that uses
another organization.

 

[Gene E. Bloch]

Gene E. Bloch wrote:

I can't vouch for all of them, but they seem to come
with a single NTFS partition. So the MBR would be defined.

I probably told you at some point, about the *BIOS*
on an old computer here, that will not tolerate an all-zero MBR.
If you do that, and leave the zeroed drive connected, the
computer will not boot. You have to disconnect the zeroed
drive, to complete boot. Without that, basically the
BIOS freezes up.

In that case, you need to put that drive in a USB enclosure, wait
for boot to complete (so Windows is in control), and then work on the
drive over USB. Then you can use Disk Management and so on.
Once "repaired", then it's safe to connect directly to
the motherboard again.

I've learned all of this, the hard way :-( By getting
stuck, and having to "bail like crazy".

Paul

Thanks for the info - also for the reminder about that all-zero
problem.

 

[J. P. Gilliver (John)]

In the Intel family of processors, the size of a byte has to do with

Hmm. To me, the size of a byte is always 8 bits. However, I suspect you
mean the size of instruction word ...

the architecture of the CPU and its instructions, and with how data is
accessed, not the width of the data path.

.... which I agree is not necessarily the same as the data width.

Instructions are built up in units of 8-bit bytes, and they address
memory in the same units.

They probably always are these days. I was involved with some processors
(not all on one IC! They were "bit slice" processors, used in DSP) where
the instruction word was _not_ a multiple of 8 bits wide. (I remember 4
bits of it were for the sequencer chip, and so on.)

Character codes and data transmission protocols also are based on
eight-bit units.

Most every processor I've dealt with in the last few decades is the
same way. It's been a long time since I've seen a machine that uses
another organization.

Indeed - '80s and '90s when we were working on those roll-your-own DSPs.
(2900 chipset - 2910 sequencer, 2901 and 2903 ALUs, and I think we used
mostly ordinary TTL chips for most of the rest of it, such as address
generators.)

 

[Gene E. Bloch]

Hmm. To me, the size of a byte is always 8 bits. However, I suspect you mean
the size of instruction word ...

No, I mean a byte. And it was a direct reply to your remark about
continuing to use the 8-bit byte. It seemed to me that you meant why do
we continue to use 8 bits as the size of a byte, rather than, say the
64 bits you mentioned.

... which I agree is not necessarily the same as the data width.

They probably always are these days. I was involved with some processors (not
all on one IC! They were "bit slice" processors, used in DSP) where the
instruction word was _not_ a multiple of 8 bits wide. (I remember 4 bits of
it were for the sequencer chip, and so on.)

I did specifically refer to the Intel architecture, because I know that
one, and I don't know a lot about other architectures. Although as you
can see I still speculated below

 

[charlie]

No, I mean a byte. And it was a direct reply to your remark about
continuing to use the 8-bit byte. It seemed to me that you meant why do
we continue to use 8 bits as the size of a byte, rather than, say the 64
bits you mentioned.




I did specifically refer to the Intel architecture, because I know that
one, and I don't know a lot about other architectures. Although as you
can see I still speculated below

Don't forget "Nibbles" <G>

I missed the early 4 bit microprocessors, since the technology I worked
with changed from TTL logic and function chips to Minicomputers
overnight. (HP2100 and it's successors, on into minicomps with
microprocessors.)

One system, on F-15's, used a microprocessor for almost each sub
assembly, and shared main ram memory with the system micro processor
controller. (the "logical" sub assemblies were a receiver, various
signal processors, signal generators, modulation generators, and
transmitter modules, as well as communications, and transmitter voltage
controls. The system also responded to external control via a
Manchester data buss.

My end of the stick had mainly to do with the test stations, which were
minicomputer controlled, with stimulus and response from DC to above
10Ghz, and engineering/calibration lab measurement accuracy. Getting
serious about calibration meant flying in NBS transfer standards.

We even came up with a scheme that used a Z80 based laptop to test the
systems "in situ", using mostly commercial software. The "customer" was
in the big expensive test station mode at the time, and didn't want to
"confuse" the funding source, so the laptops were never officially
approved. Besides, the laptops of the time would not meet the required
environmental specifications.

 

[Ken Blake]

No, I mean a byte. And it was a direct reply to your remark about
continuing to use the 8-bit byte. It seemed to me that you meant why do
we continue to use 8 bits as the size of a byte, rather than, say the
64 bits you mentioned.

A byte is eight bits by definition. To talk about a 64-bit byte would
be like talking about a 64-ounce quart.

 

[Paul]

I hate to correct spelling here, but since this is a matter of
definition, I'll just mention that it's normally spelled "nybble," as
opposed to "nibble," just as it's "byte" instead of "bite."

Variations are allowed. Where I worked, it was *never* spelled with a "y".
And we got many opportunities to use it, in computer equipment hardware
design and chip designs.

http://en.wikipedia.org/wiki/Nibble

I expect this is a locality issue, like which side of the ocean you're on.

Byte was always spelled with a "y", no exceptions.

Paul

 

[Char Jackson]

I hate to correct spelling here, but since this is a matter of
definition, I'll just mention that it's normally spelled "nybble," as
opposed to "nibble," just as it's "byte" instead of "bite."

Like you, I prefer the nybble spelling, but Wikipedia lists a few
alternatives to nybble, including nibble and even nyble. Nybble is the
only one that looks 'right' to me.

<http://en.wikipedia.org/wiki/Nybble>

 

[Gene E. Bloch]

On Tue, 06 Dec 2011 18:15:43 -0800, Gene E. Bloch

A byte is eight bits by definition. To talk about a 64-bit byte would
be like talking about a 64-ounce quart.

A byte is not necessarily 8 bits by definition, although 8-bit bytes
are 8 bits by definition.

Here's what the American Heritage Dictionary, 5th edition says:

byte, n.

1. A unit of data equal to eight bits. Computer memory is often
expressed in megabytes or gigabytes.

2. A set of bits constituting the smallest unit of addressable memory
in a given computer, typically eight bits.

I've even seen it used to represent the width of a character code even
in computers that aren't byte-adressable.

And really, your post *did* read to me as though you wondered whether
bytes should be made longer than eight bits to match the wider data
busses. I now realize you didn't mean that.

 

[blank]

write it out: mega-, not Mega-

Not correct John. Multiples of the basic measurement are begun in upper case
(except k for some unknown reason). Megavolts for example. Submultiples have
lower case, millivolts for example. Hence MV cannot be confused with mV.
Incidentally Volts is begun in upper case because it is named after Volta.

 

[J. P. Gilliver (John)]

Not correct John. Multiples of the basic measurement are begun in upper case
(except k for some unknown reason). Megavolts for example. Submultiples have
lower case, millivolts for example. Hence MV cannot be confused with mV.
Incidentally Volts is begun in upper case because it is named after Volta.

What you say is true _for the abbreviations_ - mV versus MV - not for
the words written out in full, millivolts and megavolts. However
(especially since Microsoft invented "Title Case"), people do tend to
Capitalise Everything In Sight These Days.

Incidentally, the units, as opposed to the scientists they are named
for, do _not_ have a capital, certainly when written out as a word -
volt, watt, newton - and, in some cases, not even when abbreviated (that
is up to the committee who allocate such things, who _do_ tend to give
capitals unless one is already assigned).