64 bit Windows and 3gb ram - problem?

[poachedeggs]

From one website:

"Who Shouldn't Go 64-Bit?
If you're not planning on going to 4GB of RAM anytime soon, you might
wanna hold back, __since you need 4GB of RAM to take full advantage of
64-bit's memory management.__ " [my underscoring]

I see many 64 bit Windows 7 laptops with 3gb, and occasionally see 2gb
64 bit systems in people's signatures on forums (that cute habit).
Can someone clarify things here for me? I'm building a pc for a
friend and my homebuilt never goes above 1.2 gb so I'm thinking of
putting one of its 2gb sticks in the friend's and buying two 1gb
sticks so we've both got 3gb. A cost-cutting thing partly so I can
get her a bigger hard drive than otherwise.

They will both have Ubuntu 64 bit on, which seldom goes about 300mb,
plus mine is triple-booting Windows 7 64 bit and 32 bit XP in a small
partition for old stuff, where the friends will have 32 bit XP
additionally.

Is the quote an absolute and not to be ignored, and if so, why these
64 bit 3gb machines?

Thanks.

 

[Loren Pechtel]

From one website:

"Who Shouldn't Go 64-Bit?
If you're not planning on going to 4GB of RAM anytime soon, you might
wanna hold back, __since you need 4GB of RAM to take full advantage of
64-bit's memory management.__ " [my underscoring]

I see many 64 bit Windows 7 laptops with 3gb, and occasionally see 2gb
64 bit systems in people's signatures on forums (that cute habit).
Can someone clarify things here for me? I'm building a pc for a
friend and my homebuilt never goes above 1.2 gb so I'm thinking of
putting one of its 2gb sticks in the friend's and buying two 1gb
sticks so we've both got 3gb. A cost-cutting thing partly so I can
get her a bigger hard drive than otherwise.

They will both have Ubuntu 64 bit on, which seldom goes about 300mb,
plus mine is triple-booting Windows 7 64 bit and 32 bit XP in a small
partition for old stuff, where the friends will have 32 bit XP
additionally.

Is the quote an absolute and not to be ignored, and if so, why these
64 bit 3gb machines?

Thanks.

64 bit takes up more memory as the pointers are bigger. If you don't
have enough memory to need 64 bit to get it all all you are doing is
slowing the system down compared to a 32 bit version.

 

[Conor]

64 bit takes up more memory as the pointers are bigger. If you don't
have enough memory to need 64 bit to get it all all you are doing is
slowing the system down compared to a 32 bit version.

Absolute utter ****ing bollocks.

 

[Conor]

Is the quote an absolute and not to be ignored, and if so, why these
64 bit 3gb machines?

Thanks.

No its not an absolute. You can run 64bit versions of Windows on 512MB
machines if you want.

 

[TVeblen]

No its not an absolute. You can run 64bit versions of Windows on 512MB
machines if you want.

This is a good read on the subject:
http://www.bit-tech.net/bits/2007/10/16/64-bit_more_than_just_the_ram/2

 

[Bug Dout]

Is the quote an absolute and not to be ignored

No. It's as idiotic as the first response you got to your question. 3GB
is just fine for Winx64, as is 2GB.

 

[RayLopez99]

From one website:

"Who Shouldn't Go 64-Bit?
If you're not planning on going to 4GB of RAM anytime soon, you might
wanna hold back, __since you need 4GB of RAM to take full advantage of
64-bit's memory management.__ " [my underscoring]

They will both have Ubuntu 64 bit on, which seldom goes about 300mb,
plus mine is triple-booting Windows 7 64 bit and 32 bit XP in a small
partition for old stuff, where the friends will have 32 bit XP
additionally.

Is the quote an absolute and not to be ignored, and if so, why these
64 bit 3gb machines?

Wow, you are ambitious--triple booting!

I have no informed comment, but I am somewhat suspicious that 64 bit
is completely totally backwards compatible with 32 bit. Why? Because
I code in Visual Studio and I notice that when a publish executables
for a program with the "Debug" option on, where extra stuff is added,
it sometimes behaves slightly differently than when executables are
published for "release". Sometimes, on very rare occasions, these
differences are so pronounced that the program will fail on Release
but not on Debug. Very rare but happens. So the analogy would be:
32 bit is like Debug mode and 64 bit is like release. Release is
"faster" but has "extra stuff" (actually taken out, but sometimes
stuff is combined to make it faster) that makes the 64 bit OS arguably
different than the 32 bit OS when it comes to handling programs.

I doubt anybody here save a hard core hardware engineer who designs
microprocessors could tell us definitively if my hunch is correct
however.

And, I suspect, just like Debug vs Release in Visual Studio, the
differences are slight and any hiccups very rare.

RL

 

[Loren Pechtel]

I have no informed comment, but I am somewhat suspicious that 64 bit
is completely totally backwards compatible with 32 bit. Why? Because
I code in Visual Studio and I notice that when a publish executables
for a program with the "Debug" option on, where extra stuff is added,
it sometimes behaves slightly differently than when executables are
published for "release". Sometimes, on very rare occasions, these
differences are so pronounced that the program will fail on Release
but not on Debug. Very rare but happens. So the analogy would be:
32 bit is like Debug mode and 64 bit is like release. Release is
"faster" but has "extra stuff" (actually taken out, but sometimes
stuff is combined to make it faster) that makes the 64 bit OS arguably
different than the 32 bit OS when it comes to handling programs.

I doubt anybody here save a hard core hardware engineer who designs
microprocessors could tell us definitively if my hunch is correct
however.

And, I suspect, just like Debug vs Release in Visual Studio, the
differences are slight and any hiccups very rare.

I've never compared VS debug and release versions but I have hit a
doozy of a bug of this sort with VB4--I can't imagine how they shipped
it. Compile in memory, fine. Compile to disk--32 bit integers would
be trashed if you wrote them to a file. 100% reproduceable.

Microsoft isn't alone in this, Borland shipped a DPMI library that
handled real mode pointers as pointers while in protected mode--and
the compiler would manipulate them by loading them into a
segmentffset pair. Generally BOOM---except if you single-stepped
with the debugger everything was fine, you got the correct output!

 

[Loren Pechtel]

Absolute utter ****ing bollocks.

How about explaining why then?

The actual need of 64 bit registers is rare and thus you won't gain
much from these. I know I use a hell of a lot more int's than
int64's. Rarely do you need to count to more than 2 billion with
integer types.