Re: Cube G4 ram into eMac 700



Curmudgeon <leave@xxxxxxxx> writes:

Ram Speed doesn't "Switch" it is a HARDWARE Propagation Time designed
into the Ram chips when they are built. If you have PC100 Chips
they WILL function at PC100 Speeds. If you have PC133 Chips, they will
function at PC133 speeds, as well as ANY Slower Speed. What the Module
says it is, is programmed into a separate memory location, on board
the module, and read during Memory Initialization.

Well, thanks for your ever-so-pleasant reply, You, but these modules do
say PC133 on them, but according to my friends at OWC fail to report it
to the System Profiler application, reporting instead to have "toggled"
down to PC100. If they were really running at 100 MHz, I'd expect that
I'd see the sort of system crashes others have predicted (or at least a
noticeable sluggishness in general) but the machine is solid as a rock.

If You would like to take it up with the folks at OWC, please feel free
to give them a call.


That's not how it works. First of all, RAM does not care about "Megahertz"
at least not directly. They are concerned about time.

Take this oversimplifed example..

You need to read a byte of data at address 00FFAA23BAD.

Step 1) Push 00FFAA23BAD to the address bus handling such. [Systems can
have more than one such bus...]

2) Wait

3) Read the data bus at that address..


You wait, because the address bus takes some time to quiet down, the chip
then takes some time to find that data and put in on the data bus, then
you want that bus's noise to die down, etc...

Better RAM does not take as long for its step; its data is up sooner.
So, if your CPU can read it sooner, it's "faster". Now, if you don't
read it as fast as you could, but wait a while, as a slower system would,
that's OK too. If you waited long enough, the data would eventually bleed
away except for refresh...

[Dynamic RAM is a series of tubes no wait... it's millions of little
capacitors holding a [un]charge. Keeping them correctly [un]charged over
time is another aspect called "refresh".......

We talk memory bus speed for the benefit of the humans designing systems.
Somewhere there's a spec that gives times for the dozen-plus access
times on PC100 vs PC133.

I've used PC133 in PC100 systems with no difficulties, but that does
not mean it's universally true.




--
A host is a host from coast to coast.................wb8foz@xxxxxxx
& no one will talk to a host that's close........[v].(301) 56-LINUX
Unless the host (that isn't close).........................pob 1433
is busy, hung or dead....................................20915-1433
.



Relevant Pages

  • Re: IEEE-488 bus
    ... more ram (heck, i have 512mb ram sitting here for a year since my pc ... The bus would take into account not only what we just discussed (dual ... Real chips used or FPGA, ... Minimum system is the C64 architecture and glue logic. ...
    (comp.sys.cbm)
  • Re: What kind of RAM?
    ... No warrantee. ... not mentioned is BUS speeds. ... I understand you want to know your RAM upgrade options. ... The RAM chips your system supports do not need to be matched for size ...
    (microsoft.public.windowsxp.basics)
  • Re: [PATCH 00/10] AXFS: Advanced XIP filesystem
    ... The NOR is not on the RAM bus, ... I'm not sure if cache is an option with this device - but would it ... ARM's cache and see if that speeds up streaming read time. ...
    (Linux-Kernel)
  • Re: Funky RAM
    ... very same RAM chips had been in the box for nearly a year and I'd not ... I sorted mine by just lowering the speed the RAM was running in the ... I had mine slightly 'tweaked' for gaming, I lowered the speeds and ... > POST sees all of the chips just fine, and the install is finished, is there ...
    (RedHat)
  • Re: Help RAM limits for Windowsxp Professional version
    ... Identical chips are always best; ... "stall" because two different size RAM sticks are used? ... If two chips have different speeds, ... laptop has a 1 Gig and one 512; they have to be in pairs in that one. ...
    (microsoft.public.windowsxp.help_and_support)