Did you try buying a new bios chip and installing it yourself? I see them for the GD70 on ebay for around 11$ USD. Might be faster than RMA.
Printable View
Yup...this one is soldered..so that may not be the best way to do it as its going to kill your warranty.
hello guys
it's quite some time i'm reading you oc-maniacs and here i am asking some questions that would neccessary need a lot of experience with this board.....it's my first MSI and i have always run into Asrock or Asus bioses since now but i don't understand only a few sides of this board.....here is the spec of my new gaming system:
Enermax Modu 82+ 625W
555 C3
2x2GB Patriot Viper 1600 CL7 1,8v (is the XPM spec, but i keep them at 1333 Cl8 1,5V)
Ati 5770 1GB and so on....this is in a HTPC case so not so "windy"!
I'm stucked at: 3,8ghz a 1,37V (to stay large, but even at 1,35V it is ok with a 1hour Linx run...max tep 54°C) and 2400 NB..........
RAM for now i don't spend time because it is very annoying me that i cannot push NB to 2800........what about your NB settings for the tensions in the bios?for now i have 1.9 bios but i'm thinking to go to the 1.C.....
Another thing i don't understand is the ACC in this board........i unlocked my 555 to B55 and it was ok, did not linxed it but made a run of cinebench 11 and it was ok, but i have to put ACC into "auto" otherwise the system stay dual after the reset....is it possible to have a X3 instead of the X4?
thanks you all :p:
Mmmmmmm thanks for the reply! :welcome:
Yes i guess something like that but when i went to the "core control" after the unlocking and tried to "disable" one core i think that nothing changed....but maybe i made something wrong, these evening i will be more "on the job"...
Ah after putting 1,30V for the vNB i could reach 2600 of NB.....nothing to do with 2800, i got error reporting in OCCT linpack...maybe pushing to 1,33/1,35V i can but i have to understand if it worths the famous candle...at cinebench 11 the 2600NB let me own a 0,01 don't know about the games improvements (my main target!)....
I tried even to push a little on timings, i tried 1333 888 1T at 1,55V.....but occt rebooted my system so is not very stable...i will raise to 1,65 and see if 777 is possible otherwise i will keep 898 1T....yeah, this is very poor for you but i haed to buy a elc-hapo kit, don't want to bench :)
Ah, 3,8ghz the 555 can keep at 1,35V but i can maybe slow down a little more....
cheers
I am going to run a couple of air benches on this board with a 1090T....I have read that a few broke down with the load of a 1090T...so any inputs would be great guys....am looking at around 4.1-4.2 for some air runs...
I have already changed the tim on the NB and SB, but left the PWM area the way it came with the same thermal tape.
yep i confirm.....two boards on our forum left after oc on 1090T...the voltages were not very high, think that the ampers went up to much for the "old" AM3 mobos like 790....dunno about 890 boards....
you where right and i made right but during the post the X4 B55 write did not change so i guess that it was still X4 and not X3 like CPU-z told me after :ROTF:
thanks for the support!
I think i have found my specs: 555@X3 3,8ghz 1,38V (i can lower maybe) and NB 2600..... ddr3 at 1,6V (maybe i can lower) 1333@888 1T (it's a shame i know :rofl: ) and the 5770 for now stay at 900/1300 because is overheating in my htpc-case-full-of-things without any case fan (i'm installing a 120 in and 2x80 out)...
cheers!!!:up:
I'm not running crossfire, But i am running a 5870 and i can produce the same error as you have stated. As soon as the error hits the system will restart. I can open as many CPU-Z windows as i like without a hitch. My HT-Flood errors hit when i do any hard I/O task. I have 3 systems with this hardware and only one does this. I have not had the error since flashing to the newest bio's revison.
Indeed. Had I modded the board in some other way (aside from installing the mount hardware for my nh-d14) I might go that route, but since I haven't . . .
btw how poor was the thermal interface on the sb that you replaced? Do you think the thermal interface under the PWM sink is going to be a major problem? It would be pretty disappointing if 790FX-GD70s are prone to burning out when running an overclocked X6.
Well the one below the NB and SB was really very bad.....it was like pink chewing gum...the pwm area was covered with a thick thermal tape...I am still a bit worried that it may not be sufficient....anyway just got a 1090t today, lets see if it will work.
i cant believe how many of these have been dying w/ x6 lately, they must be having a busy RMA dep't lol
about the ht sync error, i really wish we knew more about this. i had it happen maybe 5 times total with my gd70/4870x2 & it was always when there is decent stress on the system. i wonder if it is board specific? like if you got a different gd70 madDias, with all your same hardware, would it reproduce the error every time still? very strange
http://i262.photobucket.com/albums/i...-5870error.jpg
Im not sure what causes this problem. Anyway I posted quite a few bench test results in this thread earlier with the GD70 and the 1090T Thuban but I was running two 4870x2s quadfire. No problems with the system what so ever. I had seen the sync flood error message at the end of post just before the system loaded the OS but that was very rare. I could do just about anything I wanted as far as clocking the GFX card or about anything else in the system.
Well the other day I thought I would do a little GFX upgrade to make another run a 30K in 3D06. After a lot of money spent and a few days later I had two new GFX cards sitting in a box in front of my door (5970 and a 5870).
I dont know it the sync flood error or whatever problem causes that error is the problem with my system but there is a problem. I also dont know if I did a clean install of the OS and everything else if that would fix it or not?
I seen the sync flood error message after the GFX upgrade. The system either drops to the desktop during a bench test, the screen goes black, it crashes then reboots or as in the above SS during 3D05 the screen goes black then sets that error message. When the message appears I can tap the Esc key and return to the desktop. Usually I get a bunch of white pixels on the screen as the bench test locks then sets the above error message.
This happens at default or mild GFX overclocks. All the above problems are with only one 5870 installed. When I install the 5970 or the 5970 and the 5870 together some bench tests such as vantage wont even start as the system just crashes then reboots. Im using the Corsair 1000W PSU and I dont see it as being the problem.
All the above problems happen during the graphic part of the bench tests not the CPU tests.
I am using the latest bios on the MSI website which I flashed after first trying the GFX cards with the beta bios.
I may pull the cards and install them in the Asus Crosshair III to see if the problem still occurs.
Thanks for confirming the HT sync flood issue.
I'm pretty sure it's a problem with the latest 1.c0 bios. With 1.b0 i could bench vantage at 4.3ghz, with 1.c0 i get the error either during gpu tests or definetly during cpu tests.
Even the new 890 gd70 shows a strange behaviour when a 5970 is installed, at least that's what dumo was experiencing.
A friend of mine has a gd70 that was a "winki edition" i guess it's another revision of the board than mine. But seeing that different people with the same setup are expierencing the same issue as me, i don't think it's worth swapping the mobo.
I hope that everyone, that is suffering from the ht sync flood error, is writing a ticket to the msi customer support. We need a new bios asap.
Ok its dead...so this board cannot handle a thuban at 4 ghz for even benchmarking. Took all of 40 minutes to die. Just saw one small spark and that was the end of it. Im sure the vermont is fried but since i cant see any damage im not going to take out the hs. Going in for rma tomorrow. Just a heads up to anyone else who wants to get a hexa fix.
damn that sucks dude.. its too bad they dont play well w/ x6, the gd70 gave me best results on my x4 by a mile compared to giga 890gpa-ud3h & crosshair IV formula
cannot handle 4ghz??
i even primed 4.1ghz@1.475v for over an hour and ran vantage at 4.3ghz@1.55v. i donīt know how u cooled ur pwm, but just because it has no fan on the heatsink doesnīt mean that it will run cool without ventilation. and a case fan isnīt really enough to cool the pwm heatsink.
this board can handle a thuban at 4ghz very well...
better knock on wood :p: at least 4 boards from XS have died for sure + frozen said 2 boards from his forum have died already, i think they have some power issues w/ x6 that can hopefully be fixed w/ bios updates
mav2000, did you have "active phase switching" enabled when it died?
Yes active phase was on. Anyway my sense is that ver 1 boards may be the issue and ver 1.1 may be better specced. The VRM Heatsink was extremely hot to the touch after the shutdown. And yes I did change out the crappy bubblegum on the NB and SB and my temps on both reduced by 7-10 degrees after this.
which bios is the best ?? last one from msi 1.c or beta 1.c2 ????
today i received a message from msi tech support:
"First sorry for late reply. We can duplicate this error issue and we are checking this with related department then we will get back to you as soon as possible.
Best regards,
MSI Technical Support Team
Posted: May 28, 2010 by MSI"
so they now know whatīs going on and i hope that weīll have a new bios in the next few days.
and for those that canīt believe that this board is capable of running a x6 at 4ghz for longer than just a benchmark:
http://www.abload.de/thumb/4.1ghzum5r.jpg
1hour prime95 blend with bios 1.C0
249x16,5@1.475vid
249x11 NB@1.3v
1660 DDR3 7-7-7-21 @1.65v
CīnīQ enabled (thatīs why the htt clocking, cīnīq only works up to multi 18.5 on the 1.C0 bios)
i donīt know how mav2000 killed his board... i also donīt know what active phase switching is.. is that something in the "green power" bios menu? i never used those features, cause they always just gave me a no boot.
Well...is yours a version 1.1 by any chance?