the Abit IN9 32X-MAX thread

holy thread resurrection, batman.. hehe.. I know people in this thread ran the fsb on this board at 475mhz at least.. speaking of the devil,

OfficerM - how's that quad core treating you??
 
Hey REV... I guess you know I had to RMA my first IN9 32x-max mobo.. I just got my "new" one back from RMA and plugged my quadcore in.. viola.. Works like a champ. I am bumping into a few issues but it's not the motherboard.. This quad is a little FIN-UH-KEE ...lol.

First off.. I am able to achieve a solid overclock at 3.2ghz on stock vcore. No problem.
I have successfully ran the 9x, 10x, 11x and 12x multipliers.. The 8x and 13x multipliers hate me.. lmfao. I froze EVERY SINGLE TIME when I used those multipliers, regardless of the final CPU speed.

I know that I am not having very good results in getting a HIGH FSB like I could with my E6600.. But, in Everest Ultimate Edition, when I benchmark, I noticed my RAM scores went down a little.. BUT.. I absolutely CRUSHED my old CPU scores.. Just simply crushed them!

My CPU ZLib score on Everest, with my E6600 at 3.6ghz was 42,319.
My CPU ZLib score on Everest, with my QX6700 at 3.4ghz is 87,889!

HUGE CPU improvement! Will I be able to absolutely use all 4 cores right now.. Not a chance. But, I can rest assured knowing that any program / upcoming game(s) that take advantage of multicores, I will get the most out of my system. I am VERY happy with my setup.. I ran 3dmark06 for the first time yesterday and again, I was VERY pleased with the results! Based on Futuremarks tally, at the moment, I rank roughly in the top 1% !!! Sure there are systems out there that would just crush my score, but for me, I am NOT vapochilling or using liquid nitrogen, etc.. LOL! Here are my results.. (look at how many system configurations that have been submitted and look how many submissions that are faster than me.. Not many ahead of me..)

3dmark06-3.jpg



3dmark06-3-1.jpg



3dmark06-3-2.jpg
 
officermartinez said:
Hey REV... I guess you know I had to RMA my first IN9 32x-max mobo.. I just got my "new" one back from RMA and plugged my quadcore in.. viola.. Works like a champ. I am bumping into a few issues but it's not the motherboard.. This quad is a little FIN-UH-KEE ...lol.

First off.. I am able to achieve a solid overclock at 3.2ghz on stock vcore. No problem.
I have successfully ran the 9x, 10x, 11x and 12x multipliers.. The 8x and 13x multipliers hate me.. lmfao. I froze EVERY SINGLE TIME when I used those multipliers, regardless of the final CPU speed.
Click to expand...

Are you saying that the issue of overclocking quad cores on the IN9 has been fixed? :confused:
 
defiant007 said:
Are you saying that the issue of overclocking quad cores on the IN9 has been fixed? :confused:
Click to expand...

Well, since I just got this quadcore up and running, I don't know exactly what issue(s) you are talking about. I am vaguely familiar with some motherboards not completely being compatible or having issues correctly recognizing the quadcore. I am familiar with multiplier not being recognized by two of the cores when using a higher multiplier than 10. I am also aware of the limitations of achieving a HUGE FSB like the C2Ds..

After I updated my BIOS to version 1.1, I haven't really encountered any problems. It boots just fine. I WAS finally able to use the 13x multiplier but could not achieve any high FSB so my memory suffered as a result. The 8x multiplier, however, was still unattainable. When selected, I could NOT get the 8x multiplier to even boot up. I tried every possible combo for almost 6 hours.. I could only gather that as soon as my BUS SPEED went up that high, it refused to boot. On my E6600, this setting was perfect.. On my QX6700, it's a different story.. Compare the screen shots below..


E6600&



CPU-ZQuadcore3.jpg



I noticed something as I was loading those pictures.. My E6600, I had my RAM at a 1:1 ratio.. I set my RAM on a 1:1 ratio in the BIOS with the quadcore but CPU-Z reports it as a 1:2 ratio.. What gives? I understand that my BUS SPEED (effectively doubled) equals my RAM frequency of 533. What I don't understand is why my quadcore hates higher FSB's? I would LOVE to jack up those settings but I haven't had any success. Can anyone shed some light on this issue?
 
The fastest I seem to be able to run my FSB is 1333. Works out well with 10x multi.
 
Saaa-weet!

You ~should~ (in theory) be able to get to about 325 ~to~ 333 or so on the FSB, I think... that seems to be a common hard-deck for the FSB on most 680i boards with quad cores... but with that unlocked multi you should be able to get a good clock out of it.. I mean, at 12x and 325 FSB that's 3.9ghz.. very much not too shabby.. but if the chip and board are willing.. I think more mature bios releases will help with this greatly.. hopefully..

your 1:2 ratio looks right on.. with a low FSB like that, @ 266, 266*2(dual channel) = 533 and *2 again(ratio) = 1066 which is what your memory is rated for.. when you change the FSB to something other than 266 I think the mobo doesn't try and autoset the divider.. well, maybe.. at the various strap levels (266 for 1066, and 333 for 1333) the mobo will try and auto set thing like memory dividers and spread spectrum settings.. or I have seen options un-grey themselves at those frequencies..

anyways.. super cool that you got things running again! with time and better bioses I am sure you get even more OC out of that monster.. :D
 
officermartinez said:
What I don't understand is why my quadcore hates higher FSB's? I would LOVE to jack up those settings but I haven't had any success. Can anyone shed some light on this issue?
Click to expand...

It's a physical problem with the motherboards. Check out this, there's a pic on page 2
http://www.vr-zone.com/?i=4756

I'm ready to order my rig this weekend and not sure what to do for a mobo. I really like the Abit but Evga has fixed the quadcore problem but getting a good board from them seems to be a crap shoot, not to mention the whole RAM eating thing. I emailed Abit about the quadcore issue and they wouldn't even email me back. I'm gonna start dual core and go quad later so I don't wanna loose mhz in the process. Abit Sean over at Hexus is asking for company input on the forums but for some reason my account never gets validated. So color me frustrated on the quadcore front. Also check this out.
http://forum.abit-usa.com/showthread.php?t=121615
 
Well, I have not heard an official ruling if these non-factory spec 680i boards do indeed suffer from the same hardware limitation.. but do to the fact that the behave the same way one can surmize this is the case.. I still think if you can muster 333 on the fsb and use multipliers you can get a crap-ton of performance out of a quad.. better than a high clocked (with high fsb, like 450+) dual core...
 
Has anyone gotten a code 26? I'm running a 3DFuzion 7600GS and I get a error code 26 on either master or slave slot. It will boot (most of the time) when I use the 3rd slot but will complain and tells me to move it to either the master or slave slots.

Is there a incompatibility issue?
 
I'd just do Shamino's fix for the IN9 but I don't think my soldering skills are that high. Hey officermartinez, are all four of your cores clocking or just two? Thanks.
 
officermartinez said:
Well, since I just got this quadcore up and running, I don't know exactly what issue(s) you are talking about. I am vaguely familiar with some motherboards not completely being compatible or having issues correctly recognizing the quadcore. I am familiar with multiplier not being recognized by two of the cores when using a higher multiplier than 10. I am also aware of the limitations of achieving a HUGE FSB like the C2Ds..
Click to expand...

My understanding was that it was only possible to overclock 2 of the 4 cores, but from what you are saying it sounds like it is possible by upping the FSB, but not if you up the multiplier.

This is basically a showstopper for me, I really wanted to buy this board but for the amount of money they want it should really work properly. From reading the abit forums it sounds like this is an issue with the IN9, the AW9D and the QuadGT.

Considering getting an revision 2 evga board as they seem to be producing better results when overclocking quad cores.
 
You can overclock the quad. There are two issues to consider though.

1. It seems all 680i mobos cannot take the the fsb much beyond 333 with a quad.
2. At all multi's 10x and below, all cores are clocked evenly. Once you go above 10x multi, only two of the cores will be running at the higher multi but the other two will still be running at 10x.

For me, 333 fsb with 10x multi gives me 3.3GHz on the cpu. Running 4 instances of P95 draws 125+amps according to uGuru. CPU temps hit 70c accoring to coretemp. Basically, even though I can set 12x multi and go 3.8Ghz and prime just fine, only two cores are actually running 3.8 and the other two are like, I forget, but like 3.2. Playing LockOn when clocked like that causes massive warping! I mean, the plane warps 2 miles at a time! Everything else I've tried doesnt seem to mind the uneven clocks.

Still though, I'll wait until abit fixes it before going above 10x multi, plus I think I may need better cooling.
 
defiant007 said:
This is basically a showstopper for me, I really wanted to buy this board but for the amount of money they want it should really work properly. From reading the abit forums it sounds like this is an issue with the IN9, the AW9D and the QuadGT.
Click to expand...
it's solved on the QuadGT (which fyi will run a quad ~450fsb) & appears to be a common code fix so is expected to be so in the others v. quickly.
 
BUFF said:
it's solved on the QuadGT (which fyi will run a quad ~450fsb) & appears to be a common code fix so is expected to be so in the others v. quickly.
Click to expand...

Well if abit can release a fix to the IN9 by the time Nvidia and ATI release their refresh parts then I will consider getting the IN9. If the problem isnt solved by then I will look at getting the evga board instead
 
Slammin said:
The fastest I seem to be able to run my FSB is 1333. Works out well with 10x multi.
Click to expand...

I've been toying with the settings all day today.. So far, I think I am going to settle for this setting.. It seems to play nice and behave. I have my VCORE up to 1.3650 volts.. Here is a screen shot..


Quadcore3.jpg
 
defiant007 said:
My understanding was that it was only possible to overclock 2 of the 4 cores, but from what you are saying it sounds like it is possible by upping the FSB, but not if you up the multiplier.

This is basically a showstopper for me, I really wanted to buy this board but for the amount of money they want it should really work properly. From reading the abit forums it sounds like this is an issue with the IN9, the AW9D and the QuadGT.

Considering getting an revision 2 evga board as they seem to be producing better results when overclocking quad cores.
Click to expand...


Basically the first two cores were getting overclocked.. The other two cores were lower (although they did overclock a little over the stock speed of 2.66ghz). But that was on a 12x multiplier.. I moved my multiplier back down to 10x and noticed ALL 4 cores overclocked equally. My temps went up just a bit (roughly +3c overall). Very acceptable temp increase, IMO. I am going to play around with my current 3.333ghz setting.. Benchmarking, gaming, etc..
 
***UPDATE*** yowzers.. I've been Orthos testing and Holy Crap Batman.. I've since had to add more VCORE.. I was at 1.3650 volts.. Now I am at 1.4050 volts.. That VCORE, coupled with the FSB Strap of 1333 has been the minimum settings to get this QX6700 stable. Check out how HOT this thing is getting!


Quadcore3-1.jpg
 
officermartinez said:
***UPDATE*** yowzers.. I've been Orthos testing and Holy Crap Batman.. I've since had to add more VCORE.. I was at 1.3650 volts.. Now I am at 1.4050 volts.. That VCORE, coupled with the FSB Strap of 1333 has been the minimum settings to get this QX6700 stable. Check out how HOT this thing is getting!


QX6700&
Click to expand...


Yeah, 1333 and 1.38vcore is 100% stable for me too. You can even set the cpu strap to 1333 as well, though I've not seen any effect from doing that - good or bad. I do set it to 1333 though. Nice memory clocks btw!

I'd check the TM (AS5?) spread on your block mount. You have a huge temperature difference between your two cores. I can tell you that the half grain of rice method is not good for these quads, and though my coretemp never goes above 70c, I don't have a lot of confidence that I've used the correct amount of AS5.

Also, the diff between my two cores never exceeds 5c, so maybe I'm ok now, but if I could get the two higher cores to read what the two lower ones read at load, I'd be tickled :)

At idle, the diff between the two is 2-3c per coretemp.

edit: oops! Looks like you are only priming two cores?
 
aldy402 said:
possible to fit a 60mm fan with those included clips?

thanks in advance
Click to expand...


Yes. Just use twisty ties with the clips. Thats what I did :)

These little fans are whiny though. I've opted to control all my fans with the bios and mobo fan headers this time. The is a new one for me because I'd gotten into the habit of NEVER using mobo fan headers, but right now I'm running 5 fans off the mobo and controlling them all via bios and I like this a lot. In the past, I've read, and have run across some mobo's that had problems when you had too much draw on the mobo fan headers. Not seeing any of that now, and uGuru/bios is really great at handling it so far with this mobo. Silent until loaded.
 
Slammin said:
Yes. Just use twisty ties with the clips. Thats what I did :)

These little fans are whiny though. I've opted to control all my fans with the bios and mobo fan headers this time. The is a new one for me because I'd gotten into the habit of NEVER using mobo fan headers, but right now I'm running 5 fans off the mobo and controlling them all via bios and I like this a lot. In the past, I've read, and have run across some mobo's that had problems when you had too much draw on the mobo fan headers. Not seeing any of that now, and uGuru/bios is really great at handling it so far with this mobo. Silent until loaded.
Click to expand...


thanks slammin, you wouldnt have a pic would you? 40mm fans dont move enough air and its too expensive to replace the northbridge and southbridge together
 
I think the multiplier issue can probably be fixed in the BIOS (not holding my breath) and I'm betting the FSB issue is a physical problem with the boards. We still aren't seeing many games that utilize multithreading so I'd have to think with these limitations that a dualcore with a higher overclock would have to be faster, unless you want many different apps running at one time. I think the mobo makers better get their head out of the sand or Evga is going to eat them for lunch when it comes to 680i. Thanks for all the info officermartinez, do you have any screenshots of the temps and voltages in Uguru?
 
aldy402 said:
thanks slammin, you wouldnt have a pic would you? 40mm fans dont move enough air and its too expensive to replace the northbridge and southbridge together
Click to expand...


I don't currently have access to a camera but basically, just place the fan on top and secure it to the spring clips with twisty ties.
 
officermartinez said:
Yeah.. Only TWO cores are FULLY being stressed. Do you know of a way for Orthos to stress ALL FOUR CORES at the same time.. If so, lemme know!
Click to expand...

mlambert890
"Yep, you have to run Orthosx4 basically, setting the "CPU" setting to which proc you want that instance on. Also have to create four Orthos directories. If you look at my screenshot and JethroXPs, you'll see the four Orthose boxes. Im not running a wide screen res, so they're a little mashed in there, but I tried to make sure that the run times show for each one."
http://www.hardforum.com/showpost.php?p=1030617998&postcount=146
 
Tnoggin said:
I think the multiplier issue can probably be fixed in the BIOS (not holding my breath) and I'm betting the FSB issue is a physical problem with the boards. We still aren't seeing many games that utilize multithreading so I'd have to think with these limitations that a dualcore with a higher overclock would have to be faster, unless you want many different apps running at one time. I think the mobo makers better get their head out of the sand or Evga is going to eat them for lunch when it comes to 680i. Thanks for all the info officermartinez, do you have any screenshots of the temps and voltages in Uguru?
Click to expand...


Sure.. I just took this screen shot about 5 minutes ago..


TEMPS3.jpg
 
Tnoggin said:
mlambert890
"Yep, you have to run Orthosx4 basically, setting the "CPU" setting to which proc you want that instance on. Also have to create four Orthos directories. If you look at my screenshot and JethroXPs, you'll see the four Orthose boxes. Im not running a wide screen res, so they're a little mashed in there, but I tried to make sure that the run times show for each one."
http://www.hardforum.com/showpost.php?p=1030617998&postcount=146
Click to expand...

Slammin said:
Here ya go:

ftp://mersenne.org/gimps/p95v251.zip
Click to expand...


Thanks guys!!!!!
 
No prob.

By the way, your pwm temps at idle read what mine read at load! Mine used to read 70'ish at idle too. I fixed it by using AS5 on the NB and SB and folding the heat tape in half on the PWM's so that it was tighter. Dropped load PWN (120amps) to around 80c and idle pwm sits around 40c. I'm not sure how much I gained by adding the 60mm fan and I have it set to 1300rpm at idle, 3600 at load but I'm sure it doesn't hurt.

I know it's a pain tearing everything down to do it, but once you get the mobo out, removing the heat pipes takes less than 60 seconds! Hopefully the heat tape under your pwm is stable enough to fold. My rig had been running for a few days before I did it, but the tape was still in pretty good shape but I'd imagine had I gone on longer, the tape might have fallen apart, or got stuck or something.

Once abit fixes the bios for the multi, I plan on going to 3.6GHz!
 
Still love this board for C2D. Hopefully, when the QC drops in Qtr 3, most of the issues running 4 cores will be fixed by BIOS updates, and I'll swap then....unless those rumors about initial penryn's running on 680i's via drop in are true;)
 
Slammin said:
No prob.

By the way, your pwm temps at idle read what mine read at load! Mine used to read 70'ish at idle too. I fixed it by using AS5 on the NB and SB and folding the heat tape in half on the PWM's so that it was tighter. Dropped load PWN (120amps) to around 80c and idle pwm sits around 40c. I'm not sure how much I gained by adding the 60mm fan and I have it set to 1300rpm at idle, 3600 at load but I'm sure it doesn't hurt.

I know it's a pain tearing everything down to do it, but once you get the mobo out, removing the heat pipes takes less than 60 seconds! Hopefully the heat tape under your pwm is stable enough to fold. My rig had been running for a few days before I did it, but the tape was still in pretty good shape but I'd imagine had I gone on longer, the tape might have fallen apart, or got stuck or something.

Once abit fixes the bios for the multi, I plan on going to 3.6GHz!
Click to expand...


Yeah.. I'm not about to pull the heatsink off.. I've RMA'd one board.. Who's to say, I won't have to RMA this one.. LOL. Abit's little disclaimer states something to the effect that if ANY part of the board is physically damaged or "altered", that they reserve the right to not refund the deposit for their RMA service. WTH! I should have RMA'd it thru eWiz, where I originally purchased it.. I am a moron.. LOL.
 
I bought an unopened QX6700 from a [H]ard|Forum member yesterday for $730 shipped. Once
that arrives, all I will need is a PSU and a graphics card. I am going to wait until April 17th to
buy the graphics card because Nvidia is supposed to launch its new flagship that day. I hope
they do!

I will make sure to come back here and post some pics and benchmarks for y'all.
 
Has anyone gotten the SPDIF input to work? I have 5.1 analog out speakers, and I can't get anything to come out of them. I'm trying to do this with an xbox 360. The only options I can find in the realtek audio settings seem to be for pass-through (spdif in to spdif out) which doesn't really make much sense.
 
You must log in or register to reply here.
Back
Top