The issue is not known other then it is with the Phenom chip itself some way shape or form. I wish for it to be fixed also but as much testing that has been done there is nothing that points to what it could be doing to cause the system to freeze/
Printable View
The issue is not known other then it is with the Phenom chip itself some way shape or form. I wish for it to be fixed also but as much testing that has been done there is nothing that points to what it could be doing to cause the system to freeze/
I would like my OC to be enabled all the time, not by some program.Quote:
Originally Posted by KTE
What I meant by NB, is that it ran at 2034MHz where as stock is 1800MHz, so I don't quite understand why I would have the problem also at stock...
Well, I already figured out my OC wasn't stable at 2,6GHz, but at least one person at a Finnish forum also had problem with NB climbing too high with his 9750, he has lowered the multi so that NB and HT results at about 1750MHz and as far as I understand, 3GHz seems stable for him so far. His motherboard is MSI K9A2 Platinum, and he also reverted back to BIOS 1.1 to get P-States and NB multi working again.
By the way, does anyone happen to know the default volts for NB and HT? Would be nice to know.
PS. Of course, I could try to raise the volts for my memory too. Had it at 2.0V.
That isn't what I answered to. You said:Quote:
Originally Posted by Suosaaski
"It might be that I'll have to revert to an older BIOS in order to change the NB multi to something lower, just to find out if it was NB causing the problems or what."
And to that I said use Bar_Edit and you can find out very easily by dropping the multi. ;)
Takes less than 10 seconds to change it and then to reboot, the reboot time and you'll have a lower multi to see if that hinders you or otherwise.
When you say NB, what do you mean, CPU-NB or MB-NB?Quote:
By the way, does anyone happen to know the default volts for NB and HT? Would be nice to know.
MB NB is 1.10v
HT 1.20v
The one that needs to be adjusted via P-states. so that would be CPU NBQuote:
Originally Posted by KTE
And when I sead "OC to be enabled all the time" I meant via BIOS settings, not by program like bar_edit, but you are correct, bar edit would be an easy way to check the nb multi.
If you're looking for some repeatable Linux benchmarks, here they are. ;)
PTS: http://global.phoronix-test-suite.com/
I've ran many [automatic] benchmarks, haven't been able to merge the total results files yet but they're all uploaded - you'll find them all under my username. Here they are.
Attachment 77668
Stock Volts on all hardware
2000MHz NB, 2000MHz HT, 800MHz 4-4-4-12-20 RAM Unganged
http://imageupload.com/out.php/i111011_1.png
http://imageupload.com/out.php/i111316_1.png
http://imageupload.com/out.php/i111002_6.png
http://imageupload.com/out.php/i111001_5.png
http://imageupload.com/out.php/i110999_4.png
http://imageupload.com/out.php/i110998_3.png
http://imageupload.com/out.php/i110997_2.png
http://imageupload.com/out.php/i110996_1.png
http://imageupload.com/out.php/i111004_8.png
http://imageupload.com/out.php/i111315_1.png
http://imageupload.com/out.php/i111009_1.png
http://imageupload.com/out.php/i111010_1.png
http://imageupload.com/out.php/i111008_10.png
http://imageupload.com/out.php/i111005_9.png
http://imageupload.com/out.php/i111317_1.png
http://imageupload.com/out.php/i111447_1.png
http://imageupload.com/out.php/i111448_2.png
http://imageupload.com/out.php/i111449_3.png
http://imageupload.com/out.php/i111003_7.png
http://imageupload.com/out.php/i110995_1.png
Hows that? :D
I'll try running more - the Kernel compile benchmark placed the highest load I've seen so far on Phenom.
I know what you meant but you wanted to "check" if lower NB multi gives you any more stability and I mentioned an easy and quick method to do so with that. If it works you can then flash 1.13b which most of us use, and that will give you NB multi options back so you can use it in the BIOS. If you want such options, you will have to learn them though, they are not difficult, they just need time spent on like with anything - probably 30 minutes maximum for a learner.Quote:
Originally Posted by Suosaaski
I'm not sure what NB VID is stock on your CPU, open AMD Power Monitor and it will mention it. :)
Hey KTE: MSI accepted my RMA of the board that died during the bios flash, I was a pretty loyal customer before, now those guys are my heros!! ;) The replacement is on it's way back now.
Is the 1.13 Beta still the best OC/Fully Functional bios (P-States etc)?
Thanks.... Dave
Great news :D
1.13b is still the best - but try 1.4 aswell.
In Linux I've found a way to get Vcore; Screenlets, grab Sensors and AllCoreUsage screenlets.
The first is not useful due to the CPU load it places apart from when running the voltage/temperature sensors it picks up. It registers the real Vcore from the I/O chip.
The second shows loads across all cores conveniently, CPU usage is minimal.
You can see them running top right in my ss below:
http://img227.imageshack.us/img227/1...shotir4.th.png
nice and cool looking desktop BTW...
http://global.phoronix-test-suite.co...51-10040-30430
:D
64bit vs. 32bit might make an big impact here, also i ran at 2.8GHz/2.25GHz.
Try to run the universe suite kte. How do you submit under your username?
EDIT: Saw you ran already, here's my result http://global.phoronix-test-suite.co...19379-691-1761
Thanks :)Quote:
Originally Posted by jonspd
The beauty of it is still hidden though... Compiz at its best, cube, fire, rain, water, skydom, fade, wobbling windows, pull back windows, exploding windows, transparent windows, burning windows -- you name it ;)
Most of the applets are hidden too, but whilst I really hated, yes hated, many Ubuntu versions before it for the massive problems - this one really is attached to me, so much so that I can't be bothered rebooting into Windows :p:
Jeeez SpecViewPerf9, Quake4, Enemy Territory Quake Wars and Doom3 are taking eons to download in that PTS for Linux :(
Also they take a lot of space, beware - I might have to add GB to that partition.
Keep an eye on the above graphs, more will be added, maybe 2.8G and 2.9G too and I'll try adding 4450e and Q6600 on there if I can...
Hehe, yeah, with multi-threaded ones, 64b does make a good difference, but wow, see the scaling, large decrease in time there. Can you download and run v0.5 please, Michael says the results between different versions shouldn't be compared.Quote:
Originally Posted by justapost
The above shots I posted are from the universe suite and others. :yepp:Quote:
Try to run the universe suite kte. How do you submit under your username?
Or are you saying, you're trying to run that? :shrug:
You have to first register at: http://global.phoronix-test-suite.com/
Wait for email validation, click the validation link and then open console:
cd to the phoronix-test-suite dir
type:Then you can type your login name and password - after that it will always ask you to upload anything you run automatically.Code:./phoronix-test-suite login
When it says "launch PTS to view your results" answer no and then it'll ask you to upload results automatically.
If you have an old result you want to submit, then first find that result.
For me say 2700-2000 (a result I saved), they are saved automatically under:
/home/kte/.phoronix-test-suite/test-results/2700-2000 [kte is my OS username there]
Then open console, cd to phoronix-test-suite dir
Now find where your results are saved - the name will be the filename you gave before the benchmark [special identification] and it will be an XML file in the same dir as the composite.xml file, i.e. in the example above:
Then you have to type in console;Code:/home/kte/.phoronix-test-suite/test-results/2700-2000/test-1.xml
It'll be uploaded under your username then. Hope it helps ;)Code:./phoronix-test-suite upload /home/kte/.phoronix-test-suite/test-results/2700-2000/test-1.xml
Edit, seen your new results just now.. what speed and tRFC is your RAM in?
I'm going to keep it at plain settings for now without any hardware tweaking, but I'll give your settings a try if you let me know them fully under 32b.
Thank you for the detailed info, saved me some time. :) I created three different accounts till now, activated all of em and none works.Quote:
Originally Posted by KTE
I get "PTS Global Account not found" instead. :shrug:
I ran the above test at 12,5*225 CPU, 10x225 NB and Mem at 600MHz 5-7-7-18-31 105ns timings.
I rerun the test now with 220 ref HT and will run at 13,5x200, 10x200, 400 later.
Here are the 0.5.0 results, the new version made the difference
12,5x220 : http://global.phoronix-test-suite.co...8208-5683-8972
630.79s
13,5x200 http://global.phoronix-test-suite.co...906-17635-2096
615,63s
EDIT: universe http://global.phoronix-test-suite.co...586-17145-7140
BTW: The linux kernel comes with an module for the msr registers.
modprobe msr -> /dev/cpu/[0-3]/msr
It's a character device which is rw. I wrote a small script in python to read the tlbcache info few month ago. Did not manage to write to this device but with a small c-program it should be no problem.
Also nice http://ozlabs.org/~jk/code/bitfield/
Give me a few minutes- that last run of yours was very helpful, I'll now run the same tests and compare under 32b ;)
When and where does it say this? In console?Quote:
Originally Posted by justapost
Have you manually logged in before running a benchmark?
I just had an error, I ran your benchmark and uploaded it - it should show up alongside your last universe benchmark but it doesn't. I have the relevant files saved though. So 64b vs 32b [yours/mine]:
http://img3.freeimagehosting.net/uploads/ce41bbac84.png
http://img3.freeimagehosting.net/uploads/16446a311e.png
http://img3.freeimagehosting.net/uploads/7060b20365.png
http://img3.freeimagehosting.net/uploads/3488ef2b60.png
http://img3.freeimagehosting.net/uploads/ac23e4d981.png
http://img3.freeimagehosting.net/uploads/40afdbc5fd.png
http://img3.freeimagehosting.net/uploads/6b4aa45435.png
http://img3.freeimagehosting.net/uploads/6d60a29d84.png
http://img3.freeimagehosting.net/uploads/6afee22f8c.png
http://img3.freeimagehosting.net/uploads/8525b8dbdb.png
Nice, thanks for the info. I'll play around with it later. :)Quote:
BTW: The linux kernel comes with an module for the msr registers.
modprobe msr -> /dev/cpu/[0-3]/msr
It's a character device which is rw. I wrote a small script in python to read the tlbcache info few month ago. Did not manage to write to this device but with a small c-program it should be no problem.
Also nice http://ozlabs.org/~jk/code/bitfield/
Played alittle with bitfield and the msr module.
Here is a code snippet reading the p-state-0 msr register
Here is the bitfield definition for the p-state register (not all voltages are in it)Code:import struct
f=open("/dev/cpu/0/msr","r")
f.seek(0xc0010064);
(msr0,)=struct.unpack('L',f.read(8));
print "0x%x" % (msr0)
f.close();
/etc/bitfield/bitfield.d/phenom.conf or /etc/bitfield.d/phenom.conf
If you call bitfield with the result of the python script like thisCode:[P-State]
name: P-State Registers
width: 64
field: 58:63 CpuFid
field: 55:57 CpuDid
field: 48:54 CpuVid
value: 0 1.5500V
value: 1 1.5375V
value: 2 1.5250V
value: 3 1.5125V
value: 4 1.5000V
value: 5 1.4875V
value: 6 1.4750V
value: 7 1.4625V
value: 8 1.4500V
value: 9 1.4375V
value: 10 1.4250V
value: 11 1.4125V
value: 12 1.4000V
value: 13 1.3875V
value: 14 1.3750V
value: 15 1.3625V
value: 16 1.3500V
value: 17 1.3372V
value: 18 1.3250V
value: 19 1.3125V
value: 20 1.3000V
value: 22 1.275V
value: 24 1.250V
value: 26 1.225V
value: 28 1.200V
value: 30 1.175V
value: 32 1.150V
value: 34 1.125V
value: 36 1.100V
value: 38 1.075V
value: 40 1.050V
value: 42 1.025V
value: 44 1.000V
value: 46 0.975V
value: 48 0.950V
value: 50 0.925V
value: 52 0.900V
value: 54 0.875V
value: 56 0.850V
value: 58 0.825V
value: 60 0.800V
value: 62 0.775V
value: 64 0.750V
value: 66 0.725V
value: 68 0.700V
value: 70 0.675V
value: 72 0.650V
value: 74 0.625V
value: 76 0.600V
value: 78 0.575V
value: 80 0.550V
value: 82 0.525V
field: 42:47 SBZ
field: 41 NbDid
field: 39:40 SBZ
field: 32:38 NbVid
value: 0 1.5500V
value: 1 1.5375V
value: 2 1.5250V
value: 3 1.5125V
value: 4 1.5000V
value: 5 1.4875V
value: 6 1.4750V
value: 7 1.4625V
value: 8 1.4500V
value: 9 1.4375V
value: 10 1.4250V
value: 11 1.4125V
value: 12 1.4000V
value: 13 1.3875V
value: 14 1.3750V
value: 15 1.3625V
value: 16 1.3500V
value: 17 1.3372V
value: 18 1.3250V
value: 19 1.3125V
value: 20 1.3000V
value: 22 1.275V
value: 24 1.250V
value: 26 1.225V
value: 28 1.200V
value: 30 1.175V
value: 32 1.150V
value: 34 1.125V
value: 36 1.100V
value: 38 1.075V
value: 40 1.050V
value: 42 1.025V
value: 44 1.000V
value: 46 0.975V
value: 48 0.950V
value: 50 0.925V
value: 52 0.900V
value: 54 0.875V
value: 56 0.850V
value: 58 0.825V
value: 60 0.800V
value: 62 0.775V
value: 64 0.750V
value: 66 0.725V
value: 68 0.700V
value: 70 0.675V
value: 72 0.650V
value: 74 0.625V
value: 76 0.600V
value: 78 0.575V
value: 80 0.550V
value: 82 0.525V
field: 24:31 IddValue
field: 22:23 IddDiv
value: 0 1A
value: 1 10A
value: 2 100A
field: 1:21 SBZ
field: 0 PStateEn
value: 0 Invalid
value: 1 Valid
It showsCode:bitfield P-State 0x800001ee2800280b
Gotta figure out now how to access the pci-registers. Should be nice to read out the memory timings. :)Code:decoding as P-State Registers
0x800001ee2800280b [9223374159239718923]
CpuFid: 0xb
CpuDid: 0x0
CpuVid: 0x14 [1.3000V]
SBZ: 0x0
NbDid: 0x0
SBZ: 0x0
NbVid: 0x14 [1.3000V]
IddValue: 0xee
IddDiv: 0x1 [10A]
SBZ: 0x0
PStateEn: 0x1 [Valid]
EDIT
This gives a dump of the pci registers.Code:lspic -s 00:18.2 -xxx
Bitfield stuff for the DRAM timings high registers
Bitfield stuff for the DRAM timings high registersCode:[DRAM-THR]
name: DRAM Timings high register
width: 32
field: 28:31 TrwtWB read to write turnaround for opportunistic write bursting
value: 0 3 clocks
value: 1 4 clocks
value: 2 5 clocks
value: 3 6 clocks
value: 4 7 clocks
value: 5 8 clocks
value: 6 9 clocks
value: 7 10 clocks
field: 24:27 TrwtTO read to write turnaround for data, DQS convention
value: 0 Reserved
value: 1 3 clocks
value: 2 4 clocks
value: 3 5 clocks
value: 4 6 clocks
value: 5 7 clocks
value: 6 8 clocks
value: 7 9 clocks
field: 22:23 Twtr internal DRAM write to read command delay
value: 1 1 clocks
value: 2 2 clocks
value: 3 3 clocks
field: 20:21 Twrrd write to read DIMM terminantion turnaround
value: 0 1 clocks
value: 1 2 clocks
value: 2 3 clocks
value: 3 4 clocks
field: 18:19 Twrwr write to write timing
value: 0 1 clocks
value: 1 2 clocks
value: 2 3 clocks
value: 3 4 clocks
field: 16:17 Trdrd read to read timing
value: 0 2 clocks
value: 1 3 clocks
value: 2 4 clocks
value: 3 5 clocks
field: 14:15 Tref refresh rate
value: 0 Undefined behaviour
value: 1 Reserved
value: 2 Every 7.8ms
value: 3 Every 3.9ms
field: 13 DisAutoRefresh disable automatic refresh
value: 0 Automatic refresh is enabled
value: 1 Automatic refresh is disabled
field: 12 Reserved
field: 9:11 Trfc0
value: 0 75ns
value: 1 105ns
value: 2 127.5ns
value: 3 195ns
value: 4 327.5ns
field: 6:8 Trfc1
value: 0 75ns
value: 1 105ns
value: 2 127.5ns
value: 3 195ns
value: 4 327.5ns
field: 3:5 Trfc2
value: 0 75ns
value: 1 105ns
value: 2 127.5ns
value: 3 195ns
value: 4 327.5ns
field: 0:2 Trfc3
value: 0 75ns
value: 1 105ns
value: 2 127.5ns
value: 3 195ns
value: 4 327.5ns
Code snippet to read the pci registers (here F2x8c DRAM timings high register)Code:[DRAM-TLR]
name: DRAM Timing Low Register
width: 32
field: 28:31 Tcl CAS latency
value: 2 3 clocks
value: 3 4 clocks
value: 4 5 clocks
value: 5 6 clocks
field: 25:27 Trcd RAS to CAS delay
value: 0 3 clocks
value: 1 4 clocks
value: 2 5 clocks
value: 3 6 clocks
field: 22:24 Trp row precharge time
value: 0 3 clocks
value: 1 3 clocks
value: 2 4 clocks
value: 3 4 clocks
value: 4 5 clocks
value: 5 5 clocks
value: 6 6 clocks
value: 7 6 clocks
field: 20:21 Trtp read to precharge time
value: 0 2-4 clocks
value: 1 3-5 clocks
field: 16:19 Tras: row active strobe (DDR2/DDR2-1066)
value: 0 NA/15 clocks
value: 1 NA/16 clocks
value: 2 5/17 clocks
value: 3 6/18 clocks
value: 4 7/19 clocks
value: 5 8/20 clocks
value: 6 9/21 clocks
value: 7 10/22 clocks
value: 8 11/23 clocks
value: 9 12/24 clocks
value: 10 13/25 clocks
value: 11 14/26 clocks
value: 12 15/27 clocks
value: 13 16/28 clocks
value: 14 17/29 clocks
value: 15 18/30 clocks
field: 12:15 Trc row cycle time
value: 0 11 clocks
value: 1 12 clocks
value: 2 13 clocks
value: 3 14 clocks
value: 4 15 clocks
value: 5 16 clocks
value: 6 17 clocks
value: 7 18 clocks
value: 8 19 clocks
value: 9 20 clocks
value: 10 21 clocks
value: 11 22 clocks
value: 12 23 clocks
value: 13 24 clocks
value: 14 25 clocks
value: 15 26 clocks
value: 16 27 clocks
value: 17 28 clocks
value: 18 29 clocks
value: 19 30 clocks
value: 20 31 clocks
value: 21 32 clocks
value: 22 33 clocks
value: 23 34 clocks
value: 24 35 clocks
value: 25 36 clocks
value: 26 37 clocks
value: 27 38 clocks
value: 28 39 clocks
value: 29 40 clocks
value: 30 41 clocks
value: 31 42 clocks
field: 10:11 Twr write recovery time
value: 0 3 clocks
value: 1 4 clocks
value: 2 5 clocks
value: 3 6 clocks
field: 8:9 Trrd row to row delay (DDR2/DDR2-1066)
value: 0 2/4 clocks
value: 1 3/5 clocks
value: 2 4/6 clocks
value: 3 5/7 clocks
Output for DRAM HighCode:import struct
f=open("/proc/bus/pci/00/18.2","r")
f.seek(0x8c);
(reg,)=struct.unpack('I',f.read(4));
print "0x%x" % (reg)
f.close();
Output for DRAM LowCode:decoding as DRAM Timings high register
0x00020345 [131909]
TrwtWB read to write turnaround for opportunistic write bursting: 0x5 [8 clocks]
TrwtTO read to write turnaround for data, DQS convention: 0x4 [6 clocks]
Twtr internal DRAM write to read command delay: 0x3 [3 clocks]
Twrrd write to read DIMM terminantion turnaround: 0x0 [1 clocks]
Twrwr write to write timing: 0x0 [1 clocks]
Trdrd read to read timing: 0x0 [2 clocks]
Tref refresh rate: 0x2 [Every 7.8ms]
DisAutoRefresh disable automatic refresh: 0x0 [Automatic refresh is enabled]
Reserved: 0x0
Trfc0: 0x0 [75ns]
Trfc1: 0x0 [75ns]
Trfc2: 0x0 [75ns]
Trfc3: 0x0 [75ns]
Now i'm fixed. :) I'll try to modify the bitfield code now so the program can be used to read MSR and PCI registers directly.Code:decoding as DRAM Timing Low Register
0xae599113 [2925105427]
Tcl CAS latency: 0x3 [4 clocks]
Trcd RAS to CAS delay: 0x1 [4 clocks]
Trp row precharge time: 0x2 [4 clocks]
Trtp read to precharge time: 0x0 [2-4 clocks]
Tras: row active strobe (DDR2/DDR2-1066): 0x9 [12/24 clocks]
Trc row cycle time: 0x9 [20 clocks]
Twr write recovery time: 0x1 [4 clocks]
Trrd row to row delay (DDR2/DDR2-1066): 0x1 [3/5 clocks]
NoߺߺkeyboardߺsorryߺMSIߺboardߺdeadߺYESߺTOTALLYߺDEAD ߲ߡߺߺߡDiedߺbyߺfullߺloadߺᴈᴏᴏᴏᴍᴢߺoneߺpointߺfourߺvolts ߺusingߺAbitߺboardߺnow:) :) :) :( :( :( http://global.phoronix-test-suite.co...352-6784-27732:down: :down: :down: phenomߺworkingߺfineߺthough
wtf
bring it back to life LOL
o well send a rma like dave had to do is it really dead?
Non working 9850BE here. :mad: sensors-detect -> C1Quote:
Originally Posted by KTE
9600BE working fine. :) Not that i think this is funny, the only thing beside stock was the cpu-multi at 13,5.
Man, under load I went away and came back seeing a flash from the board :down:Quote:
Originally Posted by jonspd
Tested 3 PSUs, 3 RAM sticks and 2 CPUs, all worked in other two boards but did not work in MSI.
Then checked it all out alone - it powers on but then switches off with a flash as if there's a short circuit somewhere.. there isn't anywhere but on the board, near a capacitor. Looks like on of the circuitry wires has broken walls. definitely a dead board.
I can't RMA, skanky MSI are asking me to pay $60 for the RMA, testing and replacement where by UK statutory law they have to provide me 12 months warranty at no cost :down:
They claim its their EU repair center that imposes these charges by default.
The above SuperPi 1M benchmark was at 3000MHz CPU, 2400MHz NB, 800MHz RAM 4-4-4-8-15 Unganged.
The rest of the benchmarks I ran, the Universe suite, are below [that server is "eating" all our uploaded results]:
http://img3.freeimagehosting.net/uploads/8600c83819.png
http://img3.freeimagehosting.net/uploads/800c0a2569.png
http://img3.freeimagehosting.net/uploads/68624b7f10.png
http://img3.freeimagehosting.net/uploads/cc5b845804.png
http://img3.freeimagehosting.net/uploads/e16acd9e5a.png
http://img3.freeimagehosting.net/uploads/213d18a442.png
http://img3.freeimagehosting.net/uploads/da213d4518.png
http://img3.freeimagehosting.net/uploads/ab48d13d57.png
http://img3.freeimagehosting.net/uploads/4848cd95a9.png
http://img3.freeimagehosting.net/uploads/4a0121ffe0.png
I am currently on Abit AX78 at 2.7GHz stock volts although only installed one DIMM, first BIOS: http://global.phoronix-test-suite.co...64-32241-32236
The PWM has no problem so far, they are actually running very cool compared to what my MSI was running.
Thank you for all the info Achim :up:Quote:
Originally Posted by justapost
But I have a lot of work to do now before I can get to play about with that.
I still have C1 errors on all boards, they only go away if I run Corsair 800 RAM, no 1066 or 1000 RAM will boot, all of them give C1 for a while. Try and get hold of some 800 rated compatible RAM :yepp:
NP, I had this on my todo list for a while and needed to take a break from typo3 coding. That stuff works fine and only reads registers thru kernel modules.Quote:
Originally Posted by KTE
I was aware for your issue and tried a stick of G.Skill 6400 ram with no luck. Also tried the cpu in the M3A without success, definately a dead chip here.Quote:
Originally Posted by KTE
Hope your K9A2 can be brought back somehow.
Damn.. so totally dead?
When you get the C1, does it startup and then get stuck?
Update.. stock volts, benched, no issues of freezing so far on the Abit AX78: http://www.xtremesystems.org/forums/...&postcount=232
So far it's tagging the same route as the MSI which froze after 18 hours. ;)
X264 bench, Q6600 and 9850: http://www.xtremesystems.org/forums/...9&postcount=78
then you would say that your Abit is the better board? Include the first Bios an NB-Multi?
It behaves exactly like my non working 9500. I power on the board, a led between the dimm sockest starts to glow and the status leds jump to C1.Quote:
Originally Posted by KTE
The M3A ran the 9850BE fine at stock before and does not post now. I'm sure i have no corrupted bios problems there. :confused:
Hold on I need to test more..
AX78 pulls less power than the MSI board though, at the same settings.
-Managed to test and yup, this board also showed idle issues which means its Phenom itself for sure. 210x13.5 stock all volts failed after around 8 hours, although passed all benching fine.
-Ahhhh just had the idle freeze issue again, but I was at fairly low volts -> 1.34v 2.9GHz. Could be voltage related yet, although it was similar - never failed benching, failed during Firefox, after 1hr.
-Now at 210x13, stock volts on all - working fine... so far.
Edit: Yeah Achim, sorry bud but it definitely sounds like a dead chip. How though and why? :(
KTE, have you tried running individual cores at higher multis to see if the freezing can be associated to a single core? Unfortunately i'm out of freezers here atm but that whould have been the next thing i planed to investigate further.