8RDA6+ Pro Romsip Mod

[MietmaN]

Originally Posted by Wolf2000me

Hey,

This afternoon i got my goods The booster and 2x512 OCZ VX EL Gold. I'm currently running at 2-2-2-11 @ 231x11 @ 3.3v
So far it's stable. I got 3579/3352 in sisoft memory, 38sec in superpi and a considerable increase in aquamark3. I'm going to test for absolute stability then continue to go upwards

Hiya, what a coincidence!
Today I got my G.Skill DDR550 LE puppies, but I've been testing the Merlin's bios until now, with some strange results, which I'll probably describe tomorrow (12x is the best multi, then come 10.5x and 11x very close to each other). FSB231 2-2-2-11 is something really really outstanding! Still using the optimal interface? I'm gonna stick to the aggressive setting, since the difference is quite large in my case. Nevermind - great timings! Great memory sticks - let'em fry

I'm currently modding my BeQuiet PSU to be more quiet by changing fans. I'm listening to different fans I've got in order to check, which one will best suit my needs and my mental disease named "silnece" + I've just removed the L12 hardmod and inserted the G.Skill sticks - at FSB200 2-2-2-11 so far stable with just 2.65v. Gonna test them for stability during the night and then up with the FSB - I'm curious how high they can go at this voltage?

edit:
When using mbm5 I notice my +12v rail to be at 13.85v I don't really know how to explain it... The board cannot read that voltage line and gives some BS but this is weird

I don't pay any attention to +12v line anymore - have checked once with multimeter under quite a heavy load and turned on the "ignore function" to all those readings.

cheers!

[Wolf2000me]

Hey,

I only used the optimal interface for the test with the multipliers. For the rest I allways use aggressive. It makes quite the difference on my rig as well.

I appear to be rock stable at this point so i'm going to try a bit further and then test it for stability overnight. What i do notice is the vdimm unstressed and stressed. In windows unstressed when i set 3.4v on the booster i get over 3.5v. While i'm stressing it goes to 3.4v. Also the 3.5v and 3.4v are displayed on the booster. When i'm at startup or memtest this does not happen.

Btw, elhzi... I had to bend two capacitors, one was easy but the other one was a real pain, i almost had to lap the heatsink on the booster for it to fit

[Wolf2000me]

Damn,

I let it prime for 6 hours, no errors at 232 mhz 2 2 2 11. Then i tried 235 mhz at the same voltage while all of the sudden i get errors immediately. Always at the same 2 places. First i figured i might need more voltage or my northbridge was failing me but that's not it. I immediately set the settings back to 232mhz but it kept erroring on me. Again at the same places. So I tried both modules in single channel. My conclusion is that one of them has gone bad on me
Even at 200 mhz memtest immediatly finds error on those 2 places. So i'm going to have to do the RMA dance and work in single channel with 512mb for a while. I never put more than 3.5v on the modules while the lifetime warranty clearly states that the warranty will not expire till the usage of 3.5v vdimm.

[Nate-X]

I had to RMA my 8rda3+ 3.1 over a week ago. Since newegg isn't offering that board anymore they are sending me an 8rda3+ Pro.. Should be getting it today. Can't wait to try out this bios!

My 2x 512mb twinmos w/winbonds are wanting some powah since I couldn't oc much on my 8rda3+

[MietmaN]

I didn't write anything about my tests of different multipliers, as I had promised before, because my mobo passed away 3 days ago. I was trying out some different case fans and after I plugged in a no-name 0.13A 80mm fan to one of the two fan conectors near the socket, something burnt. It's an electronic path/track (whatever) that leads form the middle pin (+12v) of those two connecotrs to a PLL chip near the CPU mosfeds.

@#$%^&* - I have plugged in and unplugged dozens of fans, hundreds of times an nothing like that has ever happened. Got a new one - rev.1.4. (some voltage tweaks, if I should believe in the USDM readings, nothing changed in the layout, the same capacitors and so on).
More news on Merlin's bios and my new G.Skill mem sticks later.

cheers guys!

The new bios - no news from Merlin - I've written something about it here :

http://www.aoaforums.com/forum/showt...095#post298095

[Daniel ~]

Sorry to hear you lost one MietmaN!

[MietmaN]

Originally Posted by Daniel ~

Sorry to hear you lost one MietmaN!

Thank you, Daniel - it was one of the most silly ways, in which you can kill your motherboard. I had never heard about anything like that before it happened.

[Daniel ~]

Death often results from trivial causes... Silly and unimportant take their moral toll! Trust me on this one. LOL

[Wolf2000me]

Originally Posted by MietmaN

I didn't write anything about my tests of different multipliers, as I had promised before, because my mobo passed away 3 days ago. I was trying out some different case fans and after I plugged in a no-name 0.13A 80mm fan to one of the two fan conectors near the socket, something burnt. It's an electronic path/track (whatever) that leads form the middle pin (+12v) of those two connecotrs to a PLL chip near the CPU mosfeds.

@#$%^&* - I have plugged in and unplugged dozens of fans, hundreds of times an nothing like that has ever happened. Got a new one - rev.1.4. (some voltage tweaks, if I should believe in the USDM readings, nothing changed in the layout, the same capacitors and so on).
More news on Merlin's bios and my new G.Skill mem sticks later.

cheers guys!

The new bios - no news from Merlin - I've written something about it here :

http://www.aoaforums.com/forum/showt...095#post298095

Man, that's some bad luck! I'm still waiting on RMA for both my VX modules

Btw, a side question.

In the pic provided (it's the one from Elzhi) There are 2 caps encircled. I fear I may have damaged the one in the circle closest to the cpu socket by placing the ddr booster. So the question is, what is its purpose? I seem to be getting vdimm voltage fluctuations with the booster. It didn't really occur before but now i get fluctuations from 2.7v to 3.6v. It could be my PSU too because it happens during OS load where all components pull their share of the juice.

[MietmaN]

@ Woolf2000me - sorry to hear about that. Such large voltage fluctuations shouldn't happen, even while loading the OS - well, at least I've never heard of anything similar. Seems some bad times have come for overclocking freaks - if we had sitted patiently on our @#$%%, everything would be fine now, I assume.

Merlin's bios and multipliers :
Well, I've done whole lot of tests, I've got whole lot of notes I've taken, but I don't have enough time now to describe everything and make such clear comparison as Naibot or Woolf2000me have done. Apart from that I'm going to Germany again and a lot of work is awaiting me.

I've tested at FSB200 with the latest Sandra 2005 SP1 and Everest Home Edition 1.51 - here is the list of mulitpliers from the best to the worst one (the least efficient one) :

1) 12x (rocks for low FSB speeds)
2) 10.5x (rocks for high FSB speeds)
3) 11x (well optimized, though a little worse than 10.5x)
4) 12.5x (for those with worse memory sticks)
5) 11.5x (this is what I call "a bad multi" - sudden perfromance drop!)

So, I've tested everything form 10.5 to 12.5, every test I repeated at least twice to make sure the results were similar. Allright - seems to be almost identical to what Woolf2000me had tested before, but the differences in my case were more significant and the sudden efficiency drop with 11.5 multi is something weird.

The rest of my problems I'll descirbe in a separte post - they are strange indeed.

cheers

[MietmaN]

Hi guys,

First, I've got exactly the same "cold boot issue" as everyone here with Merlin's bios - strange I hadn't marked this before I went to Germany. My PC must have been running all the time at those days.
This isn't anything annoying, at least for me, but S3 state's fix would be welcomed, since not only me but some other people use /would like to use this way of power saving.

Second, I've got some serious stability issues at higher FSB speeds. To fasten everything shortly and clearly :

a) No problems at all at FSB200, no matter which multi I use - fast & stable!

b) I've tested for max. stable FSB : memtest86+ 1.50 doesn't show any single error at 10.5x210 2-2-2-11 nor at 11x220 2.5-3-3-11. The first errors appear at FSB230 (in Test No 6), but I can easily count SuperPI 16M at FSB235 - not bad for 2x512MB dual channel on nForce2.

c) Raising V-dimm doesn't help : FSB200 - 2.63v (bios), FSB210 - 2.77v (bios), higher FSB - 2.77v (2.90v doesn't give any improvement in terms of stability).

d) So, the threshold of rock stability shoud be somewhere between FSB220 and FSB230, if to trust memtest86+ and Prime95 (Large FFTs, Blend).

e) I would wish it was like that, but .....

I've got serious instabilities in all 3D benchmarks and games!
The last stable setting is 10.5x210, 2-2-2-11. How wierd - SuperPI 16M is as a rule a much better test of stability than 3DMark2001SE and despite the fact, that I can run SuperPI at FSB235, I can't run any 3D benchmark, even at FSB220 (memtest86+ rock stable!).

I've treid everything - differnt V-core, different V-dimm and Vcc, various timings and some other memory settings - nothing has helped.

Last night I flashed back en Epox stock bios - everything runs fine, but the results are much worse - the worst I've ever seen with an nFroce2 motherboard, whole lot of bioses and my current system. To make it cear : 10.5x210, 2-2-2-11 (stock) = 11x200, 2-3-3-11 (Merlin). Nevertheless, I can run all those slilly benchamrks at FSB230 without any problems at all!

The questions are :
- Has Merlin done some AGP tweaks in his bios, that improved efficiency a lot, but stability suffers from them significantly?
- Anyone here with similar problems? Naibot? Runnel? Wolf2000me? Could you run 3D2001SE or Aquamark 3 a few times at high FSB's frequencies like 220MHz or 230MHz????

My video card is fine and can handle much more, my PSU keeps all voltages rock stably and it would be the last thing to blame for those issues.

Thanx in advance.

PS. I'm gonna flash CPC Off version right now and see, what happens.

[Wolf2000me]

MietmaN,

I have indeed seen similar results on my rig. Of course the fact that one memory module gave up on me made it kind of hard to make a good judgement of what's the cause of my instability. I do think there are other instability variables at work on my rig. Right now i'm suspecting my PSU and in less severity the Merlin CPC1 Bios. Reason is: I have been running dual channel 2x232 at 2-2-2-11 and primed it for more than 6 hours. No errors there. Memtest86 v.3.2 didn't detect any errors nor did the windows based memtest that i use sometimes. I then clocked somewhat higher to test and i did not find any real problems. Until i saw some voltage fluctuations and to find one of my modules to be fubar. I have seen instabilities in games as well. I ran Aquamark at 232 and at 222 mhz before the troubles began. No problems noted, just performance increase.
So MietmaN, unlike your findings, i do really suspect my PSU. Since my troubles began, the vcore has been dropping sometimes. Before it would never drop below the bios set value of 1.875v. Now all voltages are still the same with exception of the vcore to have been noted as lowest value to be 1.84v. So far it has not been unstable at settings from before my VX attempts (see sig). I also have my memory at the same settings but the booster remains.
MietmaN, do you use logfiles to be able to read lowest and highest voltage and temps? At any rate, i've bought an OCZ Modstream 520Watt Psu. If it's my current Psu that killed my VX stick i really don't want it to happen again. So i'd probably be able to test your findings for certain when i have received the VX RMA'd modules and the OCZ Psu. Of course other ppl could clarify some question marks

[MietmaN]

Originally Posted by Wolf2000me

I have indeed seen similar results on my rig. Of course the fact that one memory module gave up on me made it kind of hard to make a good judgement of what's the cause of my instability. I do think there are other instability variables at work on my rig. Right now i'm suspecting my PSU and in less severity the Merlin CPC1 Bios. Reason is: I have been running dual channel 2x232 at 2-2-2-11 and primed it for more than 6 hours. No errors there. Memtest86 v.3.2 didn't detect any errors nor did the windows based memtest that i use sometimes. I then clocked somewhat higher to test and i did not find any real problems. Until i saw some voltage fluctuations and to find one of my modules to be fubar. I have seen instabilities in games as well. I ran Aquamark at 232 and at 222 mhz before the troubles began. No problems noted, just performance increase.

First, thanx a lot for such a quick answer.

Second - as I wrote above, I tested at FSB220 with SuperPI 16M (manages at FSB235), Prime95 Blend and Large FFTs and Memtest86+ 1.50 (corrected "classic" Memtest86) - all of them for hours and everything seemed fine. But I couldn't and can't run any single 3D bechmark at 11x220 (moderate CPU's OC). What more - 3D05 seems to be more sensitive to OC than the rest of them - it even crashes at 10.5x210.
Only FSB200 is stable for all the possible benchmarks I can run.

Third - if it was the PSU causing such problems, I wouldn't have been able to make such results :

http://service.futuremark.com/compare?2k1=8276082

Abit NF7 v2.0, D26 Black Mantaray, 2x512MB of Twinmos Twister PC4000, CPU at 11.5x230 - indeed really stressful for the PSU.

http://service.futuremark.com/compare?2k3=3416361

The same as above, but at 11.5x225.

http://arc.aquamark3.com/arc/arc_view.php?run=486264764

The previous Epox 8RDA3+ PRO rev.1.2, stock bios, old mem sticks, CPU at 12x220.

So - is something wrong with the memory? Do my new G.Skill sticks just don't like Merlin's bios or what? Or is this particular board one of the worst I could have bought? I'm already sick and tired of all of this - it's not fun anymore for me.

Originally Posted by Wolf2000me

So MietmaN, unlike your findings, i do really suspect my PSU. Since my troubles began, the vcore has been dropping sometimes. Before it would never drop below the bios set value of 1.875v. Now all voltages are still the same with exception of the vcore to have been noted as lowest value to be 1.84v. So far it has not been unstable at settings from before my VX attempts (see sig). I also have my memory at the same settings but the booster remains.

Could be both your PSU or the bent capacitor you mentioned just a few posts above - but you get both unstable V-core and V-dimm readings, don't you?

Originally Posted by Wolf2000me

MietmaN, do you use logfiles to be able to read lowest and highest voltage and temps? At any rate, i've bought an OCZ Modstream 520Watt Psu. If it's my current Psu that killed my VX stick i really don't want it to happen again. So i'd probably be able to test your findings for certain when i have received the VX RMA'd modules and the OCZ Psu. Of course other ppl could clarify some question marks

I use logfiles from Epox USDM and manually set Motherboard Monitor (can read my fans properly) - voltages seems stable. I've also measured them with a multimeter under some decent load.

I've already almost given up - I've tried CPU Interface both Optimal and Agressive, Memory Auto Precharge Enabled /Auto /Disabled, AGP Auto and 66MHz, AGP x4 and x8 - different timings and various voltages configurations.

Hope we will both solve our problems, somehow ...

[MietmaN]

Well, I was thinking about staying with the stock bios, but the same strange 3D problems started to occur, when I did more benchmarks this night.

I reassmebled almost all of my hardware, including the reinstallation of my VGA Silencer and changing the CPU cooler (that Thermalright SI-97 is broken, heatpipes are cold, temperatures are high - has happend already to a few friends of mine ). After that I reinstalled my favourite Catalyst 4.12 (DriverCleaner in the meantime) and 3DMark2001SE.

Now - everything seems fine apart from that damned benchmark. Aquamark 3, GunMetal Benchmark, 3DMark03 and 05 and those very few games I've got. PC is both Prime and Memtest86+ stable - I've given up.

The only thing that seems to lighten that darkness, that has crept over my PC, is the fact, that Merlin's bios is fine.

[Wolf2000me]

It's indeed quite obvious your problems are not PSU problems.

With the booster still working but not applying any voltage, my system was unstable after 3 hours. I tested without the booster and it does not seem to be unstable anymore. Also the vcore drop doesn't happen anymore. So looking at the info i'd definatly not rule out your thoughts. Probably my Psu not being up to its task. I haven't been able to test that psu out on another rig in serious stress though. For me to make a good judgement i'll have to wait for my Rma and the psu i ordered. This time i ordered an OCZ modstream 520W. I'd say that should help take care of my more pressing juice problems

Have you had any luck with the CPC 2 bios or in worse an official Bios? Do you have a good way making sure the temperatures on the northbridge are okay with your custom cooling solution?

Edit: This being something of a simultanious post makes some of the info provided in this reply not to be in line with the context

[MietmaN]

Originally Posted by Wolf2000me

It's indeed quite obvious your problems are not PSU problems.

So I thought, but asking never hurts.
It's a quality PSU, already well known in Europe - the producer is German, but everything is made in China, as usual.

Originally Posted by Wolf2000me

With the booster still working but not applying any voltage, my system was unstable after 3 hours. I tested without the booster and it does not seem to be unstable anymore. Also the vcore drop doesn't happen anymore. So looking at the info i'd definatly not rule out your thoughts. Probably my Psu not being up to its task. I haven't been able to test that psu out on another rig in serious stress though. For me to make a good judgement i'll have to wait for my Rma and the psu i ordered. This time i ordered an OCZ modstream 520W. I'd say that should help take care of my more pressing juice problems

I'm glad to hear you regain stability of your system. With that PSU beast you've ordered you will have no problem for sure - will be hungry for more all the time.

Originally Posted by Wolf2000me

Have you had any luck with the CPC 2 bios or in worse an official Bios? Do you have a good way making sure the temperatures on the northbridge are okay with your custom cooling solution?

I unzipped and flashed the CPC Off (2T) bios two nights ago, but it seemed to be exactly the same bios - no 2T timing, the same performance and same issues (now I know Merlin's bios isn't to blame). Or I might have done something wrong, since I was very sleepy. The official one - I've already described my experiences with it - shortly : "a real slow cow".

The only thing I didn't reasammble yesterday was the Zalman passive NB heatsink - I had already had enough of all of this and I would have to take the mobo