Has anyone hacked any more RAM unto it yet?

[Link]

I'm also not entirely sure what he is talking about either.

That's exactly what i was talking about. Thanks Neko!

 

[mcobit]

He means drivers, the kernel, etc. Sure you could change them, but could you change everything else, assuming that modifying the hardware itself went OK?

No, no and no. Just put another piece of ram into your pc and see, if everything doesn't work anymore... I suppose not!


More RAM doesn't need any modification to existing software to support it!

 

[darfgarf]

No, no and no. Just put another piece of ram into your pc and see, if everything doesn't work anymore... I suppose not!


More RAM doesn't need any modification to existing software to support it!

does in the embedded world, or at least needs the kernel recompiling (it's been done for the dingoo)

 

[mcobit]

Are you sure about that? Ok, seems just too unplausible to me. Also I couldn't find any references.


Or maybe in the case of dingoo, they had to support modules by another vendor?

 

[Letalis Sonus]

We're talking about bare RAM chips here. A PC's RAM module consists out of more than just some chips sticked on a PCB. You have a unified interface and you have a SPD chip.

 

[mcobit]

Ok, so is the rammodule on the pop of Pandora just the bare chip or does it have any of those in it?

 

[darfgarf]

Are you sure about that? Ok, seems just too unplausible to me. Also I couldn't find any references.


Or maybe in the case of dingoo, they had to support modules by another vendor?

I've done it (credit goes to Stephanie for the hack) its all in what configuration flags you use, if you know the size of the ram its common to just specify it, on a PC with tram slots its a requirement that you don't know how much you have until boot

 

[Edglex]

No, no and no. Just put another piece of ram into your pc and see, if everything doesn't work anymore... I suppose not!


More RAM doesn't need any modification to existing software to support it!

does in the embedded world, or at least needs the kernel recompiling (it's been done for the dingoo)

On pretty much any device you can autodetect the amount of RAM (often in the bootloader). It's just a question of doing so vs statically telling the kernel/software how much to use. Example; the port of linux to the Psion devices (ARM7/710, mostly) used to use different kernels for every RAM configuration, until autodetect code was implemented. In practice autodetecting how much RAM your device has is not necessarily a priority.

 

[Exophase]

If the RAM configuration changes you may have to do more than auto-detect the capacity. Like if you go from using one chipselect to two, you may have to setup a second set of registers or another enable flag. Or if the speed rating changes you'll have to change the setup for that.

 

[Edglex]

If the RAM configuration changes you may have to do more than auto-detect the capacity. Like if you go from using one chipselect to two, you may have to setup a second set of registers or another enable flag. Or if the speed rating changes you'll have to change the setup for that.

This still can be (and probably is) done automatically as part of memory detection routines. At least, I don't see any reason why not. I'd imagine you do something like attempting to read and write different values to different chips using different chip selects, and see if it works? Speed differences between different chips on the same bus sounds a little odd to me, but again I guess if the hardware supports it you can probably cope.

 

[Exophase]

This still can be (and probably is) done automatically as part of memory detection routines. At least, I don't see any reason why not. I'd imagine you do something like attempting to read and write different values to different chips using different chip selects, and see if it works? Speed differences between different chips on the same bus sounds a little odd to me, but again I guess if the hardware supports it you can probably cope.

The chip select stuff can be done as part of the capacity, but the speed would require reading the SPD EEPROM.. if it's actually wired. I doubt you'd have different speeds selected, but if the new memory is slower you'd have to change it from whatever the old was, whether or not you're replacing the old or adding to it.