[Emperor]

Hi Finas,

thank you very, very much for the helpful tool. I have not thought any more, anybody is responding to my thread!
I just downloaded and executed the tool with the following status report printed:

J
Tested memory: 131072
Memoryblocks OK: 54987
Memory Blocks ERROR: 0
Blocks not able to allocate: 76085
Number of CRC errors: 0

(By the way, what does the "J" mean in the left upper corner of the screen?)
It confirms that our XBoxes don't recognize or I will better say don't activate the additional 64MB of RAM in the native XBox OS mode.

As we have several XBoxes with RAM upgrade (and one of it was upgraded by a professional company) and none of them activates / recognizes the whole RAM, I think that our problem must be a matter of SW. I am quite sure, we and the company haven't made any mistake during upgrade. The Samsung chips soldered onto the empty foot prints on the PCB are the same as the original ones, but a little faster in access time.
We have checked every pin to its neighbours and against ground and found no error / short circuit and Gentoox runs and uses the whole memory amount!

We all have identical machines, board revision 1.4 and the Aladin blue chip with I think 256 kB Flash, capable to store only one BIOS at a time. As BIOS we all use a modified M8 BIOS (modified with the EBVtool). Modified are only some color settings and the paths and names, where to load the default executables.

In another thread somewhere in the web I have read, that XBMC has to be compiled with a special switch to make it use the additional RAM. Are the compilations of T3CH done with this switch beeing active?
As your experience tells me, T3CH has apparently activated this switch.

Another tool I have tried in the mean time was mame128. It also doesn't recognize the upgrade.

My theory is, that the M8 BIOS or the modchip is responsible for not activating the whole RAM:
The Pentium-III processor has an integrated MMU. And the MMU I think is not programmed correctly by the BIOS to map the address space above the standard 64MB to the new available physical space.
For example I had a "weired" PC in the past with a curious BIOS, there you had to validate / select the detected RAM after an upgrade, otherwise it was not visible to any SW.
GentooX I think starts during its Bootloader its own BIOS (a modified cromwell or somewhat) and does a machine initializing of its own.
During this procedure the RAM is auto detected and configured correctly and therefore becomes visible.

As Linux recognizes the whole RAM, I think also its proved, that the aladin modchip is not responsible for our RAM ignoration.

So my questions are:

1) is my theory of the MMU configuration right
2) is there therefore a special BIOS or BIOS patch necessary to activate the RAM above the standard 64 MB
3) what BIOS are the people using, which can access 128 MB through XBMC
3) are there additional solderings (hidden switches / jumpers) to be done besides a one to one soldering of each RAM chip pin to the spare locations prints
4) are there additional wirerings necessary to one of the chipset chips to e.g. connect the new chip select signals (I dont't believe this, since Linux even works ...)
5) is there anybody on this planet using a simple standard aladin modchip with M8 BIOS and having 128 MB within XBMC

Thanks so far and it would give me great pleasure if you or anybody could help me and answer my questions.