ASUS P5Q3 Deluxe discussion

[Telo]

I am going to be honest with you here, 2GB kits are not giving us huge issues on P45, hence I have spent more time with the 4GB kits hoping to help get some solid answers as to what these boards will actually do.

Now that does not mean I will not test with 42GB and I will try to show the tweaks needed to get the board clocking well. All I ask is for now clock down a little, tighten the timings and enjoy your PC. let me sort out this 4GB mess we are seeing then I can move on.

I tried the lower divider with tighter timings idea and the same crap came up. Then I tried the lower divider with the famous 6-6-6-22 fix and it got into windows. So I got greedy and did 6-6-6-18 and it went in as well. My brother doesn't know jack about FSBs and RAM Mhz and all that so I can buy some time until everything is sorted out I guess. In retrospect, I did try 1066Mhz RAM with a 5-5-5-15 timing before and was no go, so I'm guessing this is the neighborhood my brother will have to live in until ASUS sorts out their RAM issues. Man I tell you, funny how you get more support from bystanders than from the friggin' truck that cause the accident, sort of speak; this just makes me 1000% more confident that I did the right thing by dropping the P5E3 Premium for the BlackOps. I have had the BlackOps for just over 2 weeks and I have seen like 10 BIOS revisions out so far, but, that is for another place and time. Please keep pressing on ASUS to sort out support for us, the small people with the 2GB kits.

[bingo13]

I'm working on it, Gary. I'm running at exactly the same settings with 4GB kit of 1600C7DHXIN v3.1 which is based on Micron Rev.B. So far so good at DDR3-1660; I will post a screenshot once I get back...

Does your retail board behave the same? What Rev. is it - 1.01G or 1.02G?

On 1.02G and it is much better, PM sent about the kits that do not work.

[GAR]

OK, i gave up, even at low settings, 1333mhz i couldnt get prime belnd to stay stable, i have since ordered this kit

http://www.newegg.com/Product/Produc...82E16820145200

Hope these are better, they will be here friday.

[Telo]

OK, i gave up, even at low settings, 1333mhz i couldnt get prime belnd to stay stable, i have since ordered this kit

http://www.newegg.com/Product/Produc...82E16820145200

Hope these are better, they will be here friday.

That kit more than likely will work, but it has a CAS9 off the bat; might be a bit slow.
Can't you just downclock the current RAM you got and turn it back up once this issue gets resolved? That's what I'm doing, but then again, I can't return my RAM since I had it longer than 1 month. I have 2x 2GB kits, so for me to change the RAM will mean a loss for my brother of over $500.

[BenchZowner]

Guys chill out.
All of you, especially the "big heads" ( Bachus, Tony, bingo13 ) you've been in the industry for a while, more than enough to know that sometimes a person will come up with a "trick" that gets the job done ( high clocks with 4GB (2*2GB) Kits in this case )...
This time it was X, maybe the next time it'll be Y, and maybe the next time Z.
You don't have to start beaching at each other, neither you have something to separate.

Gary ( bingo13 ) which revision is the "final" ( retail ) one m8 ?

[GAR]

That kit more than likely will work, but it has a CAS9 off the bat; might be a bit slow.
Can't you just downclock the current RAM you got and turn it back up once this issue gets resolved? That's what I'm doing, but then again, I can't return my RAM since I had it longer than 1 month. I have 2x 2GB kits, so for me to change the RAM will mean a loss for my brother of over $500.

I tried everything, it just refuses to work right.....i was going to downlclock and wait, but it just doesnt want to work.

[Telo]

I tried everything, it just refuses to work right.....i was going to downlclock and wait, but it just doesnt want to work.

Give DDR3-1066 a try, with 5-5-5-15 or 6-6-6-18 on 1.80v to the vDIMM and 1.24 to the vNB.

PS: If you are interested, there is an auction on eBay for 2x 2GB kits of CellShock 1800Mhz. Right now it is at $390 or so plus $15 shipping in the USA. No, I'm neither the seller nor involved in the sale. I was going to bid until the RAM started to work at 1066Mhz. The regular price for a 2GB kit is $420 from tankguys.com and $368 from memoryc.com plus a hefty shipping charge. Here is the link: http://cgi.ebay.com/ws/eBayISAPI.dll...MEWA:IT&ih=002 1HR left on it though.

[bingo13]

Guys chill out.
All of you, especially the "big heads" ( Bachus, Tony, bingo13 ) you've been in the industry for a while, more than enough to know that sometimes a person will come up with a "trick" that gets the job done ( high clocks with 4GB (2*2GB) Kits in this case )...
This time it was X, maybe the next time it'll be Y, and maybe the next time Z.
You don't have to start beaching at each other, neither you have something to separate.

Gary ( bingo13 ) which revision is the "final" ( retail ) one m8 ?

I have tried to stay out of it by explaining we have more problems than just a BIOS fix (which should be here shortly by the way), the chipset is not designed for what some of us expect it to do and that is match X48 at a lower cost. It just was not designed to do it, but it does not mean that it's a bad chipset, far from it.

That said, I am not going to step in the memory land mine field right now as every kit I have from C to Z is not getting to 1600 so it does no good to throw modules at each other. However, obviously there are some kits that work and figuring out how to get the rest working correctly is something ASUS is working on feverishly right now. I am convinced it will be fixed shortly.

Rev 1.02G with the Marvell/SiL controller chips is the final production version and is the one I received in a retail kit today. It is working fine, except for the memory clocking problems. I will say this, the board is very fast with 4GB at DDR3-1333 right now with all kits, it just needs a little fine tuning is all. update - I am getting ready to try a new Micron kit that just arrived on a very late UPS truck, so maybe it will bring success tonight. Gotta love drivers who go the extra mile for you.

[Charles Wirth]

Checked my notes and I think I figured a few things out for getting the memory up there.

Try dropping back to the 333 strap. I will test more with Microns when I switch rigs back.

I had the Samsungs bugging at 1600Mhz.

A fast tool for me is a Vista install CD, I boot to it and use the memory checker, it will fail within 20%, usually around 7%

[BenchZowner]

Thanks for the reply Gary
And yes Asus is really working hard nowadays to get some new BIOSes with lots of fixes for various boards, really nice to see that happening myself

Mine should be a retail version as well.
The first two weeks will be dedicated to bug detection and "hardcore" overclocking ( 2*1GB RAM of course ), and then I'll start working for the review.
First targets... break 600MHz FSB with "the supposed to be magic E8500 Q807" and break DDR3-2000 7-6-6/7-6-5 with the marvelous CellShock 2GB D9GTR kit.

[Tony]

the main issue with most of these high performance boards is that they are tuned with 2x1GB, next to no one uses 2x1's any more going by sales figures I see, as prices on DDR3 come down even more it will get even worse with many option for 4x2GB.

Now my overall point in all of this was this.

Forget I work for OCZ, the board is having issues with Micron memory rated at 1600MHZ, I was trying to help all get to where they needed to be. Along comes C with "our kit does this" which happens to be samsung (which I have here BTW also) and basically spits in the face of all those who have all ready spent hard earned cash.

Now if C wants to help lets see an SPD dump of these micron dimms that do 1660+ as thats all it can be, I just spent 5 hrs looking at SPD for DDR3 and im confident now the board is having an issue reading some of the info correctly on some modules. It is not like DDR2 where you tell the board what to do, this time the board has to calculate what to do based off info within the SPD, if the board has issues reading what is there then it will possibly not run correctly.

So im not being bigheaded, I was p1ssed someone came along and snubbed those who were trying to help. I will gladly post a universal 2GB spd once I sort this crap for all to use and get the best from their P5Q3.
I can't see C having the balls to offer that to the community

[Pyr0]

thinking my 12hr memtest was a pure fluke now
pc rebooted last night when i wanted to try aquamark and has been hard locking just running windows update
today when i came to startup the pc, it froze as it booted into windows


just been trying to swap my raptor 150 with my window x64 install on, onto one of the the marvell sata ports and could not boot at all
disk read error occured
press ctrl+alt+ el to restart (the D was missing)

also tried it with a 2gb kit of ballistix 12800 mem which has been working ok and had the same message
i made sure the ide controller was enabled, the sata cable plugged into the top port, and tried changing the "drive xpert" mode to normal, but nothing

[edit]
Just tried a Seagate 80GB SATAII drive in the same marvell port and it works

[BulldogPO]

I´m using A-Data´s Vitesta Extreme X1600:
http://oc.adata.com.tw/1_product_det...?pid=AD31600XU

I were able to run them fine with this board at 1600MHz at 2.1V and CL 7-7-7-18 when I was usin Xeon E3110 but now with E7200 they do´nt oc, weird.

[GAR]

I´m using A-Data´s Vitesta Extreme X1600:
http://oc.adata.com.tw/1_product_det...?pid=AD31600XU

I were able to run them fine with this board at 1600MHz at 2.1V and CL 7-7-7-18 when I was usin Xeon E3110 but now with E7200 they do´nt oc, weird.

Those are microns as well, i guess this board doesnt like micron ic's much.

[mrcape]

First targets... break 600MHz FSB with "the supposed to be magic E8500 Q807" and break DDR3-2000 7-6-6/7-6-5 with the marvelous CellShock 2GB D9GTR kit.

This is exactly what I'm hoping to see on P45 / DDR3. Best of luck!

[zozo]

(also posted on http://www.ocztechnologyforum.com/fo...d.php?p=278514)

Hello,

I've got an Asus P5Q3 with an OCZ3P16002GK kit and Q9450 processor, and could only install a system and boot under Splashtop/Linux/Windows by setting DDR3-800 in the BIOS. All other settings I tried with DDR3-1333 (voltage, timings, skew) were unstable. I tried all available BIOSes, even 801 one.

Today I decided to do some tests. Now I've got Prime95 running stable for half an hour (self-test 1024K passed) with the following BIOS settings in System Performance Settings :
DRAM Frequency : DDR3-1333MHz
DRAM Static Read Control : Disabled
DRAM Dynamic Write Control : Disabled
DRAM Read Training : Disabled
DRAM Write Training : Disabled
MEM. OC Charger : Enabled
Ai Clock Twister : Lighter
DRAM Voltage : 1.80
NB Voltage : 1.20
All other settings : Auto

The OCZ kit is on orange slots, the others aren't populated.

Some settings may be useless, but now it's stable I will try setting back each parameter one by one to Auto and lowering DRAM Voltage later, and test stability.
At the time of this writing, Self-test 8K passed.
Hope this helps...

Another thing : SPD says this RAM is PC3-10700... shouldn't it be PC3-12800 ? Btw, when I bought this, the vendor said "DDR3-1600".

Thanks,
Olivier

[K404]

I´m using A-Data´s Vitesta Extreme X1600:
http://oc.adata.com.tw/1_product_det...?pid=AD31600XU

I were able to run them fine with this board at 1600MHz at 2.1V and CL 7-7-7-18 when I was usin Xeon E3110 but now with E7200 they do´nt oc, weird.

Do you absolutely need 2.1V for the clocks and timings? What vNB are you using?

If you really need that voltage, thats hinting to me that the board is tuned "too tight"?



K

Its a bit crap that basically RAM manufacturers have to have RAM rated for the chipset it will be used on as they all seem to perform so differently- cross-use looks like it might lead to disappointment?

[Tony]

Today I have 12 hrs on the board, 1600 was a go although unstable and using a custom SPD. Noticed from Bachus's post fine delay settings are totally different on his micron runs to what I am seeing others are getting so i asked Asus for fine delay options in bios so we can see if this makes a difference.

just to show this board is faster at 1333 than 1600 take a peak at this.



Board overvolts like crazy, 1.76V ram =1.83V real, 1.20V NB= 1.26real
0.620 NB GTL and still tweaking.

Nice flat line over 10K for read write and copy, much lower speed than Bachus showed with tighter timings but higher tRD.

Board is weird for sure

[BulldogPO]

0801 bios did help, memory works now fine at 1600MHz 7-7-7-18, 1.95V now.

[pmp]

With 1:1 divider ?

[Tony]

0801 bios did help, memory works now fine at 1600MHz 7-7-7-18, 1.95V now.

if you have 1.95V set in bios you have near 2.1V at the dimm, becareful you do not kill them, the board overvolts like crazy

[Chri$ch]

if you have 1.95V set in bios you have near 2.1V at the dimm, becareful you do not kill them, the board overvolts like crazy

this is from board to board different...

on P5Q-Dlx is same.. some boards overvolt and some undervolt

but it is better if everey check it on his board!

The P5Q3 i had undervolt ~0.01 - 0.03v

[Tony]

Got 2xp5q deluxe and 2xp5q3 deluxe and they all overvolt here m8...i still say becareful

[pmp]

Some 1600+ tries quite successfull though with 2x1 GB setup. 2000 is rather much too far


Haha guess such low timings are working only on P45... chipset made for high FSB and low latency screenshots haha

[BulldogPO]

How to measure Vdimm?