ASUS Maximus Formula SE

[sdumper]

@GAR

qx9650 that I borrowed from work doesn't have the prime bug (4.0ghz with stock voltage on water)

seems to be only with Q6600 even with bios 0903

What prime bug can you explain?

[jonny_ftm]

Temp is def my limiting factor. Jan is when I will take apart my WC loop and then I plan to use a custom backing plate and use a fuzion nozzle setup for the cpu. Lapping a cpu, right now for me, is like the black arts. Not sure what's involved or what tools to use to go about it. I am very pleased with my speed and dont think even with all the addons ill make it stable over 4GHz, but it's worth a try, if not just once.

You're right. Before lapping, you also can check if the IHS is really concave or not. Take a razor blade and put it on the IHS surface. Direct a spot of light towards it and see how much light can pass between the IHS and blade.

Also, the GTX is a good block, You won't gain much migrating to the Fuzion. In any case, the few degrees won't give more OC headroom.

The backplate is definately worth it. Mounting the GTX without it is kinda a lottery. You have really to make many mounts before finding the sweet spot. Also, turning the mounting screws too much will bend your motherboard loosing some contact between the WB and IHS. You have to try, while Priming on small FFT, untightening the screws alternatively till you find a better contact. This is really vital with the GT/GTX blocks.

If you use the bowed version of the WB, this will be even more true. You also have to try horizontal and vertical mounting on the socket.

It can look troublesome to do all these testings, but really worth it. Was also well discussed in the forums here january 2007 then mid june if I remember when Swiftech included the new backplate with the WB.

The easiest option: contact michelle@Swiftech. She will ship you for some 6USD a backplate for your GTX and socket 775. The biggest advantage, is that the mounting will be reproductible and won't depend on the screws tightening strength

[jonny_ftm]

@GAR

qx9650 that I borrowed from work doesn't have the prime bug (4.0ghz with stock voltage on water)

seems to be only with Q6600 even with bios 0903

Of course you can't see the bug. It was only described at a voltage above 1.50-1.55v if I do remember. Running the QX9650 at stock vcore won't rule-out the bug. That bug is also seen only on the Maximus as I know, so not really related to the Q6600

[Aldy402]

What prime bug can you explain?

When I prime with my q6600 (whatever test) It just restarts out of nowhere but completely stable under SuperPI, OCCT, and Everest Stability.A lot of users have been reporting this problem but I can prime at 400 fsb but anything over 420, it just restarts.

I haven't tried running 2 instances of Orthos though. It doesn't seem to happen with C2D users though

[Aldy402]

Of course you can't see the bug. It was only described at a voltage above 1.50-1.55v if I do remember. Running the QX9650 at stock vcore won't rule-out the bug. That bug is also seen only on the Maximus as I know, so not really related to the Q6600

interesting, ill pump in 1.6v to the QX9650 and report in.. i'll also do the same with my E6300 to see if I can narrow the issue down to either voltage or FSB.

If your right, that sucks for my q6600, I cant stabilize 3.8ghz with anything under 1.5v. Do you remember what page the bug was described on?

[fadetoblack]

10-12°C, difference between cores is too huge. The thermal compound resting time won't help you, this is really huge and abnormal.

Check the loading program is all cores are used the same way
Then, reseat your heatsink
And finally lap it, as a last solution

loading prog was prime95 with 4 worker threads. ill try reseating the heatsink today. btw, can someone tell me which quadrants on the cpu (given that the edge with the triangle is on the bottom right) roughly correspond to which core? or are they not laid out like that?

[leoftw]

So I woke up and my system was still stable.

3.3ghz @ 1.32V

I guess thats fine for now on 4x1GB 2.0V @ 900MHZ

[ElEctric_EyE]

loading prog was prime95 with 4 worker threads. ill try reseating the heatsink today. btw, can someone tell me which quadrants on the cpu (given that the edge with the triangle is on the bottom right) roughly correspond to which core? or are they not laid out like that?

There's two dual cores in there. Look towards the bottom middle of the link for a pic of what's underneath the heat spreader.
http://www.arcticsilver.com/pdf/appi..._quad_wcap.pdf

[ElEctric_EyE]

You're right. Before lapping, you also can check if the IHS is really concave or not. Take a razor blade and put it on the IHS surface. Direct a spot of light towards it and see how much light can pass between the IHS and blade.

Also, the GTX is a good block, You won't gain much migrating to the Fuzion. In any case, the few degrees won't give more OC headroom.

The backplate is definately worth it. Mounting the GTX without it is kinda a lottery. You have really to make many mounts before finding the sweet spot. Also, turning the mounting screws too much will bend your motherboard loosing some contact between the WB and IHS. You have to try, while Priming on small FFT, untightening the screws alternatively till you find a better contact. This is really vital with the GT/GTX blocks.

If you use the bowed version of the WB, this will be even more true. You also have to try horizontal and vertical mounting on the socket.

It can look troublesome to do all these testings, but really worth it. Was also well discussed in the forums here january 2007 then mid june if I remember when Swiftech included the new backplate with the WB.

The easiest option: contact michelle@Swiftech. She will ship you for some 6USD a backplate for your GTX and socket 775. The biggest advantage, is that the mounting will be reproductible and won't depend on the screws tightening strength

I am using the backing plate that came with the GTX, but it is still bowing the board. That's why I gotta go custom with a 1/4" steel plate... I'll check for flatness with the razor blade idea. I've premounted 4 times before the final mount, so it will be interesting to see if it really is still flat. Also I'll go with AS5 instead of ACeramique.

[iTravis]

So I woke up and my system was still stable.

3.3ghz @ 1.32V

I guess thats fine for now on 4x1GB 2.0V @ 900MHZ

Nice job leo, mind if I see your settings?

[leoftw]

everything auto except for

fsb : 366
Dram : 2.0

[jonny_ftm]

When I prime with my q6600 (whatever test) It just restarts out of nowhere but completely stable under SuperPI, OCCT, and Everest Stability.A lot of users have been reporting this problem but I can prime at 400 fsb but anything over 420, it just restarts.

I haven't tried running 2 instances of Orthos though. It doesn't seem to happen with C2D users though

No, it's not that. If Prime hangs, and closes, than your PC is unstable. The prime95 bug will show in task manager: the CPU utilization is not 100%, rather 70% what ever you do

So launch prime95, and check CPU usage for the Prime95.exe process in task manager (not the graph which would show a wrong 100%). If the Prime95 process is too far than 100%, that's the Prime95 bug mentioned by GAR and confirmed by many

It appeared on 3.6GHz if I remember and especially above 1.5 or 1.55 vcore if I remember

[jonny_ftm]

task manager shows you 100% but when you look at prime95, prime is only using about 80-90%....this is with 1.5v and over, i guess only vista users can see this.

OK, so after testing, i figured out, any voltage over 1.5 and prime 95 doesnt load all the way, when you look in task manager you will see.

It's not a Vista issue..it's the board. confirmed behavior on XP32, didn't have to load up Vista to see it Vcore over 1.52 idle results in less than 100% prime load. Quad's only tho, not dualies.

And here, for the lazy new comers, and any one who missed it :-) all the discussion about that bug. If anyone can verify it with last bioses...

[NLight95]

Think most quads are limited to 470 FSB mate extreme or not to be primestable... more CPU FSB limited then then mobo limited

I managed 480 FSB on my G0 Q6600 (L737B) using my IP35 Pro. I think it had more to go but the IP35 Pro just couldn't handle it. I just picked up a Maximus so hopefully it can do at least the same if not better.

[NLight95]

I was intimidated by lapping, but there are some great tutorials online, and as long as you are meticulous and careful, it should go fine and I noticed a drop of about 4-5 degrees Celsius. (My heatsink base was really uneven)

I personally like this lapping kit. Includes everything you need.

http://www.svc.com/lapkit-p.html

[Tergon]

The problem that I was having before downgrading to 0602 is that all the BIOS past 0602 gave incorrect readings from the core didn't matter which program you used the CPU at idle would read 50C or so, under load it was at 80C or so, under 0602 I get 22-25 C Idle and under load 35-45 C depending on which progam loading with. I hope this helps.

Tergon, what exactly are the temp issues you are having? im also having some problems. im gonna paste in the problem from my other post here:

i installed a qx9650 today and having an issue with the temp. temps are normal on all cores except the 2nd one. in core temp (0.95.4) they are : 22,41,19,20. bios reports cpu temp as 41, and so does speed fan. the asus probe software also shows 41 for cpu temp. i have not seen the 41 change at all. even when i run prime, the 41 remains constant. the other cores fluctuate normally.

btw first i had the problem in 0901. then i updated to 0902 and still having the same issue.

[NLight95]

Has anyone replaced the nb heatsink with a Thermalright HR-05?

Or how about a Thermaltake ES II?

(http://www.thermaltake.com/product/C...4/cl-c0034.asp)

If so, then what are your results?

I'm wondering if the aforementioned coolers would would be more effective than the stock Formula heatsink with a 40m fan. I used these coolers on my old ICFX3200 and still have them hanging around.

Also, are there any issues with mounting a Tuniq Tower on the non-SE version of this board?

[Zucker2k]

Who here has done 8x500 fsb on 266 strap 4:5 ratio? I'm stuck at 8x495, run out of nb volts

[kup]

You've got to 495x8 with a Q6600? What volts/settings?

[Grnfinger]

Could somebody tell me the secret to the 9 series bios

902 was a complete disaster
903 will not post unless I use STOCK settings

Flashing with EZ Flash, load my oc profile and the bugger does is sit there on det ram, power off then back on and hit f2 to load default and it works.
Reboot and load my settings ( auto or manual ) and it wont boot, same thing det ram...
besides a very large hammer, anybody got some other ideas?

[kup]

I think are still issues with the dividers. Set them to Auto with both settings and see how it goes?

[Grnfinger]

I'm not running a divider, but I will try it.

Is 903 worth the hassle? gotta be getting close for Asus to spit out another

[giorgos th.]

i`m starting to getting tired of the memory management of the maximus....
i think i`m going to dump it because of it....
cannot play 4:5 with 9 multi and a lot of problems with 0901,0902,0903....
the guys in Asus should seriously consider of their bios work..

[Grnfinger]

i`m starting to getting tired of the memory management of the maximus....
i think i`m going to dump it because of it....
cannot play 4:5 with 9 multi and a lot of problems with 0901,0902,0903....
the guys in Asus should seriously consider of their bios work..

would you suggest NOT to flash from 802 to 903?

running 802 currently and have finally found success on 3.4ghz 3.6ghz and 3.8 ghz ( currently testing )

[giorgos th.]

my opinion is to stay with 0802...i might go back to it or to 0602..