Athlon 64 4000+ mobile on MSI K8T Neo-FIS2R

[TheWiz]

My best guess it that the ACPI module is messed up. The only shot you have at this point would be to find the CLOSEST motherbaord possible with same chipset and characteristics that I can use to replace the ACPITBL.

TheWiz

[Peregrine]

Yes, this board can take up to 1GB per DIMM slot, and up to a total of 2GB.

My best guess it that the ACPI module is messed up. The only shot you have at this point would be to find the CLOSEST motherbaord possible with same chipset and characteristics that I can use to replace the ACPITBL.

TheWiz

oh crap... that's not gonna be an easy task.

I know that Gateway laptops MX7515 and MX7525 were using this exact CPU, but I doubt the boards would be anywhere close to this one, plus I can't find BIOS for those anywhere, anyway.

I also know that according to THIS Turion 64 (which should be identical to this CPU we're trying) works fine on MSI RS482M-IL, but the chipset on this is made by nvidia, while the one on mine is VIA. Also, from their test picture, it looks like they were also using only a single stick of RAM. But, while official Turion CPU compatibility for my board says it's under testing, for this board MSI lists the Turions as compatible. So let me check what other "closest" motherboards MSI lists as supported for Turion CPUs.

---

here's what I found so far. The AMN4000BKX5BU CPU we're trying to get to work has a core stepping version E5. So do all the Turion 64 (socket 754) CPUs, and that's what I'm basing the identical functionality on. Our CPU is just a bit faster (multiplier of 13) than the fastest of the Turion's (multiplier of 12).

According to MSI, my board is not officially tested for Turion CPUs:

K8T Neo-FSR/ FIS2R

However, the following 2 boards based on identical chipset VIA K8T800 are officially supporting all Turion CPUs:

K8T Neo-V
K8T Neo-V V2.0

I don't know what other difference these boards have, and I don't know how ACPITBL from those boards' BIOS would affect this one, but perhaps you would have more insight on this.

[TheWiz]

My hunch was correct. I looked at the Neo-V AMI bios and the microcode tables were the same as yours, meaning the support changes are somewhere in the P-STATES.

If you have hotswap capabilities or guaranteed BIOS recovery methods, I would try directly flashing the K8T Neo-V. If not, I will have to think about it some more.

TheWiz

[Peregrine]

hmm.. wouldn't the P-states be applicable to PowerNow and Cool'n'Quiet technologies only? Meaning, wouldn't they only affect the lowest and highest multipliers? I'm not sure how that would relate to memory. If you remember from my earlier posts, while at BIOS v.7.1 we're not getting to POST with 2 sticks of RAM, at v.2.1, it was powering on only to get stuck at NVRam check. And with a single stick of RAM it was working fine at the correct multiplier 13. So would P-states have any relation to NVCheck ?

This board doesn't have dual-bios, and I looked at the AMI BIOS recovery instructions on your great website and even created the recovery floppy before flashing the modded BIOS yesterday, but I'm still not sure that I would get an option to recover if I flashed a wrong BIOS. So if possible, I'd prefer to avoid flashing BIOS for a totally different board. I checked the specs and there are quite a few differences between the 2 boards. Is there any way you could just adopt the P-states from that other BIOS into the 7.1 for this board?

[Peregrine]

So... any hopes left for this project? Please..

[TheWiz]

Not to worry,still thinking vigoriously on this thread

TheWiz

[banana19]

Not to worry,still thinking vigoriously on this thread

TheWiz

have you tried putting ram only in slots 1 and 3, or turning off dual channel ram mode(if it has it) and turning off that turbo cache feature it looks like the bios has?

Nick

[Peregrine]

Not to worry,still thinking vigoriously on this thread

TheWiz

Thank you!

have you tried putting ram only in slots 1 and 3, or turning off dual channel ram mode(if it has it) and turning off that turbo cache feature it looks like the bios has?

Nick

Yes, I tried various combinations of slots, all to no avail. This mobo (socket 754) doesn't have dual channel. The memory runs in single channel. And all my tests are with memory set to auto, so there's no turbo enabled at that time.

[banana19]

Strange about the memory issue because according to http://www.crucial.com/store/listparts.a...o-FIS2R%29 you can have upt o a gig per slot!

speed reporting issue.... athlon mobile cpus will report differnt speeds based on power source... im thinking thats why your getting 4200+ descriptions.

im thinking the memory controller in the chip is just not compatible with your board. looks like single sided and double sided memory even makes some kind of difference.

http://www.msi.com/index.php?func=proddesc&maincat_no=1&cat2_no=&cat3_no=&prod_no=565

If I wanted to take a chance on a bad flash I would flash the attached Asus bios... not sure if it would even solv your problem but its does have the same chipset and i/o controller.....chances are better then most it will not brick the board.


Nick

[Peregrine]

The 4200+ description was only recognized by BIOS 1.9, which we're not gonna use for this CPU. I'm using v.1.9 for the current CPU 3000+, because it's the best one for that. But for 4000+ CPU we're definitely going to use the latest v.7.1 which does correctly recognize this CPU as 4000+.

Right, that's exactly the problem. The board doesn't properly interpret the memory controller from the new CPU 4000+. Single vs. double sided shouldn't make any difference here. The board should allow to run 2 double-sided sticks totaling up to 2GB of either DDR400 or DDR333.

Thank you Nick for that Asus BIOS. Can you tell me exactly which board it is from so I can compare the specs? I'll keep that idea as the last resort, though, that is, when the Wiz runs out of ideas