Why can't I use 4Gb memory on my GA-MA78GPM-DS2H? 64bit linux

[Nighthog]

I have found a particular problem whit Linux and 4Gb memory whit this motherboard, the GA-MA78GPM-DS2H rev1.0 , similar problems I had whit a rev1.0&1.1 GA-MA78GM-S2H I had before it as well. (I managed to fry those two GA-MA78GM-S2H and as the third replacement I talked myself the better more expensive GA-MA78GPM-DS2H in return for free as replacement whit reason I should not get more fried boards whit that one instead )

I have a Phenom X4 9750 2.4Ghz(95W) and 2x2Gb corsairs XMS2 DDR2 800Mhz memory. Having the 4Gb installed shows no problems but I get problems when I want to try to use the full 4Gb memory.

There is a option in the Bios which decides if it should use below 4G or above. memory re-mapping or something like that.
When it's set to use below 4G I can't use the full 4Gb off memory and limited to something in the 3Gb, any linux distribution, memtest etc work all good whit this below 4Gb memory usage, but I want to use all off my memory so I try to set it to use above 4G re-mapping but this is a hell for anything but windows... only windows will work whit above 4G, memtest or any linux based thing will give crap in return and will be unusable. This GA-MA78GPM-DS2H works a little better than the GA-MA78GM-S2H I had earlier whit memtest working atleast whit above 4G choosen but the Linux stuff is still a no go in totality.
Whit above 4G anything to do whit linux kernels etc is full stop and freeze in various diffrent ways. The most colorfull ways was whit the GA-MA78GM-S2H giving some real nice abstract boxes of "Tv ant wars" in colour stills when trying to boot a linux dist. (as soon as I tried to start anything whit the linux stuff). (grub works all good though) The newer GA-MA78GPM-DS2H I get farther inside in the booting and such. I get as far as to login but as it should load the desktop it freezes whit the X cursor and nothing responds. You can't even start the live-cd's whit above 4G going on any off the various Ubuntu distributions I had burned. I had some seatool hdd thing that used some other linux kernel stuff and that doesn't work either whit the above 4G setting.

I reckon it's some bad coding in the works in the BIOS for the Gigabyte motherboards?

You guys know off anything?

[Kano]

Is BIOS up2date?

[Nighthog]

Is BIOS up2date?

There is only 1 Bios and it's the orginal first release on the Gigabyte website which came whit the board all ready.

So yes Bios is up2date as there hasn't been released any new ones than the first ones yet for this motherboard.

Is this a general problem whit Gigabyte or have I just been unlucky and had 3 boards that all had the same coding faults whit their Bios:es regarding usage of 4Gb or more memory?

[deanjo]

Are you using the IGP?

[mattmatteh]

i have the same board. read my other post. but having random lock ups with any amount of ram (2, 4, 6, 8). tried to reproduce and that just leaves me more puzzled.

deanjo, i am using the IGP. why would that matter ?

matt

[matt8]

This board is on my short list as it has an excellent set of features (incl. sideport memory), but mattmateh's experience seems highly negative and reported memory problems such as Nighthog's are troubling. Maybe I should get something else.

I have one question to both of you as well as others that have this mobo. I am reading many posts here and there that this mobo does get VERY hot. Is that the case with you mattmateh and Nighthog? BTW, a reviewer at newegg suggested this remedy for the NB heat problem:

BigDubbe@Newegg

Cons: The design on the North Bridge Heat sink is awful. I removed it to find a thin layer of cheap thermal paste, a sticky ring so keep it from moving , and a indent in the heat sink the size of the chip. I think it gets hot bc the sticky ring and indent causes all heat to get trapped. I fixed it by removing the ring and get rid of the cheap paste and applied some arctic silver 5. I am sure that would do the trick, but to be on the safe side I added a small fan. Now it runs at a cool 23C instead on the egg frying 86C it was at before.

This tip seems very helpful to fix the heating problem, but I am still not sure if I should go through with the hassle.

[mattmatteh]

i have an infrared laser temperature gun (borrowed from friend and is nothing fancy).

southbridge 95 F
northbridge 104 F
ram 88 F
cpu 80 F (while writing this), (cpu has larger zalman, better colling and difficult to see cpu area)
highest choke temperature 97 F

there is nothing in there that is hot to the touch and has never seemed hot. that means its really cool, or the heatsinks are not working ?

here is the report from sensors, i have no idea what this all means.

k8temp-pci-00c3
Adapter: PCI adapter
Core0 Temp:
+3 C
Core0 Temp:
-7 C
Core1 Temp:
+6 C
Core1 Temp:
+0 C

it8718-isa-0228
Adapter: ISA adapter
in0: +1.20 V (min = +0.00 V, max = +4.08 V)
in1: +1.95 V (min = +0.00 V, max = +4.08 V)
in2: +3.28 V (min = +0.00 V, max = +4.08 V)
in3: +2.94 V (min = +0.00 V, max = +4.08 V)
in4: +3.09 V (min = +0.00 V, max = +4.08 V)
in5: +3.15 V (min = +0.00 V, max = +4.08 V)
in6: +4.08 V (min = +0.00 V, max = +4.08 V) ALARM
in7: +0.00 V (min = +0.00 V, max = +4.08 V) ALARM
in8: +3.14 V
fan1: 940 RPM (min = 0 RPM)
fan2: 0 RPM (min = 0 RPM)
fan3: 0 RPM (min = 0 RPM)
fan4: 0 RPM (min = 0 RPM)
temp1: +30 C (low = +127 C, high = +127 C) sensor = thermistor
temp2: +25 C (low = +127 C, high = +127 C) sensor = diode
temp3: +28 C (low = +127 C, high = +127 C) sensor = thermistor
vid: +1.550 V

noose ~ #

[matt8]

Wow!

That's neat. I am certainly not an expert, but those readings look OK for low load level operation.

If it's not too much trouble, are you able to compile the latest kernel (which may be helpful for your purposes as well?) using multiple threads (say three tests, using: N, 2N, 4N, where N is the number of cores you have) an then measure the temperature? 60-70C after these tests should be OK readings, I guess (experts? Kano, if you're reading this, what do you think? BTW, Kano, I used to use Kanotix quite a bit; thanks for creating and developing it!)

[bridgman]

One thing to remember is that different points in the heat flow path will have different temperatures, ie the junction area of each chip will be hottest; the body of the die will be a bit cooler; chip package will be cooler again, and the heatsink will be cooler than any of them. If you think of a bunch of resistors wired in series (only you're talking thermal resistance instead of ohms) and measuring the voltage at different points it's exactly the same effect.

If the heatsink is not well attached to the chip, for example, the effective thermal resistance would be higher so the heatsink would be cooler and the chip would be hotter -- and of course it's chip temperature that matters. I'm not trying to scare you here; those temperatures seem reasonable.

[mattmatteh]

MAKEOPTS="-j3"

southbridge 95 F ( +0 )
northbridge 102 F ( -2 cpu fan is cooling it ? )
ram 97 F ( +9 )
cpu 83 F ( +3 )
highest choke temperature 102 F ( +5 )

noose ~ # sensors
k8temp-pci-00c3
Adapter: PCI adapter
Core0 Temp:
+13 C
Core0 Temp:
+8 C
Core1 Temp:
+16 C
Core1 Temp:
+15 C

it8718-isa-0228
Adapter: ISA adapter
in0: +1.22 V (min = +0.00 V, max = +4.08 V)
in1: +1.95 V (min = +0.00 V, max = +4.08 V)
in2: +3.28 V (min = +0.00 V, max = +4.08 V)
in3: +2.94 V (min = +0.00 V, max = +4.08 V)
in4: +3.07 V (min = +0.00 V, max = +4.08 V)
in5: +3.15 V (min = +0.00 V, max = +4.08 V)
in6: +4.08 V (min = +0.00 V, max = +4.08 V) ALARM
in7: +0.26 V (min = +0.00 V, max = +4.08 V)
in8: +3.14 V
fan1: 1275 RPM (min = 0 RPM)
fan2: 0 RPM (min = 0 RPM)
fan3: 0 RPM (min = 0 RPM)
fan4: 0 RPM (min = 0 RPM)
temp1: +33 C (low = +127 C, high = +127 C) sensor = thermistor
temp2: +36 C (low = +127 C, high = +127 C) sensor = diode
temp3: +36 C (low = +127 C, high = +127 C) sensor = thermistor
vid: +1.550 V

noose ~ # grep j3 /etc/make.conf


i am using 2.6.24-gentoo-r8. i tried vanilla 2.6.27.8 and didnt notice any difference in stablity. (i diconnected all but the mouse for usb).

also the zalman fan is quiet, when loading i can only notice a freqency change in the speed, it is not any louder at all.