[Solved] AMD Athlon 64 X2 3250e on Foxconn A7GM-S 2.0

[refelian]

Apologies for the spam (see this thread).

So I am still trying to find a good use for this quirky laptop am2 processor. However in this case on a Foxconn A7GM-S 2.0.

So the error is the same as in the other thread. There are no acpi objects returned which results in CNQ not working.

latest rom

So what I was thinking is that with this specific model, it seems, I can adjust the cpu multiplier. So I am wondering would it help if I do a cpu-z extract for each different multiplier and use that to create the various p -states?

[LLC]

Reason in my opinion is that Agesa code see that (P[Max] VID = MaxVID + 00_0010b), may be it must be other.
I little read "BIOS and Kernel Developer’s Guide for AMD NPT Family 0Fh Processors" 32559 Rev. 3.08 July 2007, chapter 10.5 but nothing understand.
Do you mean that set the BIOS parameter “CPU Multiplier Adjust” (4,4.5, ..., 7.5) to “Fox Intelligent Stepping” (AGESA 3.7.1.2) work fine?
If only 4x multiplier work fine, test bios mod, link in your Private Messages.

[refelian]

Reason in my opinion is that Agesa code see that (P[Max] VID = MaxVID + 00_0010b), may be it must be other.
I little read "BIOS and Kernel Developer’s Guide for AMD NPT Family 0Fh Processors" 32559 Rev. 3.08 July 2007, chapter 10.5 but nothing understand.
Do you mean that set the BIOS parameter “CPU Multiplier Adjust” (4,4.5, ..., 7.5) to “Fox Intelligent Stepping” (AGESA 3.7.1.2) work fine?
If only 4x multiplier work fine, test bios mod, link in your Private Messages.

So I just did a test by changing the multipliers with the Fox Intelligent Stepping. The 4 and 4,5 multiplier don't work (it doesnt post), but the rest does. Perhaps the voltage had to be different for the 4 and 4,5.

I also tried out your test bios and it doesnt boot in to windows or ubuntu (luckily still in freedos). The error in windows is really confusing (inaccessible boot device????). It seems in linux it breaks when it tries to read the pss objects (strangely enough disabling cool n quiet also doesnt help).

[LLC]

Perhaps the voltage had to be different for the 4 and 4,5.

The voltage range is small, so it's not from a lack of power. On the MSI Board, they decided to start with 5x. New test bios in PM.

[refelian]

Perhaps the voltage had to be different for the 4 and 4,5.

The voltage range is small, so it's not from a lack of power. On the MSI Board, they decided to start with 5x. New test bios in PM.

With TEST_5X it boots and I can see the frequency change between the two states (5 and 7,5) in Windows with CPU-Z. The Linux kernel is not giving any warnings with this rom. The powernow_k8 driver recognizes fid 0x2 and 0x7. Lscpu shows that the 0x2 state is being used as well. These are the results of the test.

With TEST_PSS it also boots to Windows. Linux kernel recognizes that there is 1 C- state but also complains about no compatible PSS object. Included here are the results of the test.

[LLC]

Okay, I got the results, so they can be deleted. You can share this mod here using your file quota in this forum.

[refelian]

Attached the 3250e supported Foxconn A7GMS 2.0 bios. It supports two states allowing it to clock back from 1500 to 1000 mhz.

Thanks LLC its going to become a useful 2nd low power server.