2x1GB HZ's in A8N-SLI Prem.

[GatorLCA]

Ok, so I know of the possible compatability issue with these boards and what not but i do not think that it's totally the boards fault, main reason being that I test out both sticks individually.

One stick tops out at 245MHz, go any higher it hangs
the other tops out st 259MHz, once it goes to 260MHz it hangs.

This in turn allows me to only go up to 245MHz when both sticks are in which is not the motherboards fault but actually the one bad stick of RAM. I find it hard to believe that there is a compatability issue when I have seen people with the A8N series mobos and the RAM work perfectly fine. There might have just been a bad batch of RAM chips and some peopel got lucky and some didnt

Obviously the people wanted the most out of the chips are pushing it to it's limits (270MHz+) but there's no way that we can get there if individual sticks cannot get that high or rather some cannot even run at the stock speed.

So, my main question is, am I allowed to RMA the sticks when they aren't even posting stock speeds?

[GSKILL TECH]

yes if it can not do 250
but i think your mobo is a big issue.... a8n is really bad for high FSB...

[GatorLCA]

well there are few people on another board I am on that have OCZ GOld that use UCCC chips and they can run them fine on their A8N board as well as with these sticks as well over.

I know my board can handle over 250Mhz since I got the one stick to 258MHz and ran tests and SuperPi'd it as well, the RAM posted at 260, but it hung right after it did

There are others as well with this new batch of RAM that are using DFI and Abit boards and are having problems with it as well, so I do not think it's neccessarily my board but the RAM itself since one stick only gets to 245MHz

Do I go through you all or Newegg to get this RMA'd

[GSKILL TECH]

well there are few people on another board I am on that have OCZ GOld that use UCCC chips and they can run them fine on their A8N board as well as with these sticks as well over.

I know my board can handle over 250Mhz since I got the one stick to 258MHz and ran tests and SuperPi'd it as well, the RAM posted at 260, but it hung right after it did

There are others as well with this new batch of RAM that are using DFI and Abit boards and are having problems with it as well, so I do not think it's neccessarily my board but the RAM itself since one stick only gets to 245MHz

Do I go through you all or Newegg to get this RMA'd
well you can rma to us or newegg depends which one you like better :)

[DAK1640]

yes if it can not do 250
but i think your mobo is a big issue.... a8n is really bad for high FSB...
???? How can you possibly say that...I have this mobo and when Crucial had their killer Z503's I regulary did 285 Stable...I have read many, many posts by others with the same Mobo with killer OC's. Where is the evidence that these mobo's don't OC well. I currently have the HZ's at 238 x 11 with no problem. I havent tried higherm but 238 runs Rosetta 7x24 on both cores.

[kamakazibond]

I dont have this mobo, or any a8n mobo's for that matter, but the more I read through the forums here I am really starting to believe that G.skill is having issues with product quality control..

I have owned 3 (yes 3) different kits of F1-4000USU2-2GBHZ. I happened to get 2 kits the same day, one delivered by Fedex, the other by UPS (A UPS mixup caused one kit to be delayed, I ordered another hopping to get them sooner, anyway).

The fedex package came 1st, I tossed them into my MSI NF3 mobo and ..... no boot.. seperated the sticks to find one would not POST at all.. Kinda disapointed, but along came the package from UPS.. I tossed those in and was ok to POST, after some testing I found that I couldn't get them Stable at 200MHz let alone 250MHz. Long story short I sent both kits back to newegg.

I Then ordered some GBZX's. They did do their rated timings and speed on the MSI NF3 mobo. I tried to OC them them but couldnt get any past 230MHz. Well I have since bought a new motherboard with nf4 chipset on it and some nice OC options, very similar to any of the DFI lanparty series mobo's.. So I put in the GBZX's and they do a tiny bit better, but only hit about 236MHz.

So now I bought another kit of GBHZ's for this new NF4 mobo. Well they do seem to be stable at 250MHz they wont go much higher..
Check my other post (http://www.xtremesystems.org/forums/showthread.php?t=95755).

Reading around these forums, I have come to believe that there is possibly some bad QC over at G.skill. here are just a few recent threads that people are having problems.

http://www.xtremesystems.org/forums/showthread.php?t=93876
http://www.xtremesystems.org/forums/showthread.php?t=90196
http://www.xtremesystems.org/forums/showthread.php?t=95789
http://www.xtremesystems.org/forums/showthread.php?t=93186

I have a friend who has a kit of GBHZ's and he is running them at 300MHz.. So these sticks have a chance of being quite good, but I have had the worst luck with them. It's been around 2 months since my 1st 2GB "memory upgrade" purchase, and I'm yet to be fully satisfied.

[GSKILL TECH]

???? How can you possibly say that...I have this mobo and when Crucial had their killer Z503's I regulary did 285 Stable...I have read many, many posts by others with the same Mobo with killer OC's. Where is the evidence that these mobo's don't OC well. I currently have the HZ's at 238 x 11 with no problem. I havent tried higherm but 238 runs Rosetta 7x24 on both cores.
from the reports of testers, i had send many handpick DDR580 2GBHZ 24/7 stable and the report are can not do even 250 1T with A8N SLi. there are about 5 cases from different power users / testers. that is why i said that.

[GSKILL TECH]

I dont have this mobo, or any a8n mobo's for that matter, but the more I read through the forums here I am really starting to believe that G.skill is having issues with product quality control..

I have owned 3 (yes 3) different kits of F1-4000USU2-2GBHZ. I happened to get 2 kits the same day, one delivered by Fedex, the other by UPS (A UPS mixup caused one kit to be delayed, I ordered another hopping to get them sooner, anyway).

The fedex package came 1st, I tossed them into my MSI NF3 mobo and ..... no boot.. seperated the sticks to find one would not POST at all.. Kinda disapointed, but along came the package from UPS.. I tossed those in and was ok to POST, after some testing I found that I couldn't get them Stable at 200MHz let alone 250MHz. Long story short I sent both kits back to newegg.

I Then ordered some GBZX's. They did do their rated timings and speed on the MSI NF3 mobo. I tried to OC them them but couldnt get any past 230MHz. Well I have since bought a new motherboard with nf4 chipset on it and some nice OC options, very similar to any of the DFI lanparty series mobo's.. So I put in the GBZX's and they do a tiny bit better, but only hit about 236MHz.

So now I bought another kit of GBHZ's for this new NF4 mobo. Well they do seem to be stable at 250MHz they wont go much higher..
Check my other post (http://www.xtremesystems.org/forums/showthread.php?t=95755).

Reading around these forums, I have come to believe that there is possibly some bad QC over at G.skill. here are just a few recent threads that people are having problems.

http://www.xtremesystems.org/forums/showthread.php?t=93876
http://www.xtremesystems.org/forums/showthread.php?t=90196
http://www.xtremesystems.org/forums/showthread.php?t=95789
http://www.xtremesystems.org/forums/showthread.php?t=93186

I have a friend who has a kit of GBHZ's and he is running them at 300MHz.. So these sticks have a chance of being quite good, but I have had the worst luck with them. It's been around 2 months since my 1st 2GB "memory upgrade" purchase, and I'm yet to be fully satisfied.
i am really sorry about what happened to you but we did 100% tested before we shipped
i do believe UCCC gives us big problem, and seem like only DFi mobo, some of Gigabyte, and Asus without issue with 250 1T UCCC.
90% of users can not run 250 1T is with A8nSLi and 5% from Abit. another 5% for really damage ram.
about 75% of RMA of 2GBHZ is good on DFi mobo. i really hate to said but Asus is like 10 times bigger than DFi... 1 mo's A8N production = 1 years of DFi all production. beliver or not but there is the number.
but why Asus A8NSLi has such problem but Dfi doesn't......

someone is about send me an A8NSLi to test! i will post reslut soon once i receive.
i do think QC no problem and DOA probably is due to shipping

[uOpt]

Count me in as somebody who severely fights with his (new from RMA) A8N-E SLI Premium to make good and fast use of 4 sticks of RAM. This is a joke...

[kamakazibond]

Count me in as somebody who severely fights with his (new from RMA) A8N-E SLI Premium to make good and fast use of 4 sticks of RAM. This is a joke...

I could be mistaken, but I believe that it is the NF4 chipset or the AMD Memory controller that has issues with 4 DIMMs of double sided memory. One of the two.. Try a google for it I'm sure you'll come up with sumthing.

As for the GBHZ's only working correctly on DFI mobo's.. I had a friend come over with a brand new DFI Ultra-D, and his GBHZ (this would be the 4th kit I have tested with) memory to swap around and test with..

I will follow up with the restuls of around 4 hours of testing here (http://www.xtremesystems.org/forums/showthread.php?t=95755).


i am really sorry about what happened to you but we did 100% tested before we shipped

Could you go into detail about what kind of testing is done? Do you test them yourself before shipping them to newegg or do you mean the people in TW test them before they ship them out to you, etc.?


i do believe UCCC gives us big problem, and seem like only DFi mobo, some of Gigabyte, and Asus without issue with 250 1T UCCC.
Maybe I am wrong, but isn't Samsung UCCC rated at 250MHz 1T?? I find it hard to believe that a major manufacturer such as Samsung is making memory rated for 250MHz that will only run on DFI mobo's.

Don't get me wrong I'm not here to bash G.skill in any way, I'm just looking to make some kind of sense out of this situation. Again, more info to be provided in the link below.

Testing the 3rd and 4th kit of GBHZ (http://www.xtremesystems.org/forums/showthread.php?t=95755).

kamakazibond

[GSKILL TECH]

I could be mistaken, but I believe that it is the NF4 chipset or the AMD Memory controller that has issues with 4 DIMMs of double sided memory. One of the two.. Try a google for it I'm sure you'll come up with sumthing.

As for the GBHZ's only working correctly on DFI mobo's.. I had a friend come over with a brand new DFI Ultra-D, and his GBHZ (this would be the 4th kit I have tested with) memory to swap around and test with..

I will follow up with the restuls of around 4 hours of testing here (http://www.xtremesystems.org/forums/showthread.php?t=95755).



Could you go into detail about what kind of testing is done? Do you test them yourself before shipping them to newegg or do you mean the people in TW test them before they ship them out to you, etc.?


Maybe I am wrong, but isn't Samsung UCCC rated at 250MHz 1T?? I find it hard to believe that a major manufacturer such as Samsung is making memory rated for 250MHz that will only run on DFI mobo's.

Don't get me wrong I'm not here to bash G.skill in any way, I'm just looking to make some kind of sense out of this situation. Again, more info to be provided in the link below.

Testing the 3rd and 4th kit of GBHZ (http://www.xtremesystems.org/forums/showthread.php?t=95755).

kamakazibond
no no i understand! i am really welcome any voice from everyone!
the test is in Taiwan and they ship to here. then i ship to newegg, personally bring it to CA wharehouse and UPS it to NJ.
well i think DOA probably is due to shipping also the receiveing of Newegg. god knows what happened during receiving (actually i saw it but don't wanna say anything!)

UCCC is only -5 by samsung! which mean only certify 400 by Samsung. and 1T 2T there no such thing in RAM/IC ...... cuz 1T 2T those "word" is only for AMD. there are not such "word" b4! Intel doesn't even have this term!!!

[GSKILL TECH]

I could be mistaken, but I believe that it is the NF4 chipset or the AMD Memory controller that has issues with 4 DIMMs of double sided memory. One of the two.. Try a google for it I'm sure you'll come up with sumthing.

As for the GBHZ's only working correctly on DFI mobo's.. I had a friend come over with a brand new DFI Ultra-D, and his GBHZ (this would be the 4th kit I have tested with) memory to swap around and test with..

I will follow up with the restuls of around 4 hours of testing here (http://www.xtremesystems.org/forums/showthread.php?t=95755).



Could you go into detail about what kind of testing is done? Do you test them yourself before shipping them to newegg or do you mean the people in TW test them before they ship them out to you, etc.?


Maybe I am wrong, but isn't Samsung UCCC rated at 250MHz 1T?? I find it hard to believe that a major manufacturer such as Samsung is making memory rated for 250MHz that will only run on DFI mobo's.

Don't get me wrong I'm not here to bash G.skill in any way, I'm just looking to make some kind of sense out of this situation. Again, more info to be provided in the link below.

Testing the 3rd and 4th kit of GBHZ (http://www.xtremesystems.org/forums/showthread.php?t=95755).

kamakazibond
btw i am still reviewing your problem. give me few days ok! thx man!
always welcome to shoot me a pm to remind to review it... too busy those days

[uOpt]

I could be mistaken, but I believe that it is the NF4 chipset or the AMD Memory controller that has issues with 4 DIMMs of double sided memory. One of the two.. Try a google for it I'm sure you'll come up with sumthing.


I have two Opteron dual-core CPUs and four 4-GB sets of RAM which all do DDR400 in my SLI-DR. All 4 combos fail in the Asus, all parameters the same. Just today I moved one CPU and 4 sticks from the DFI to the Asus and I have to take back memory frequency by 20% solid (and the DFI wasn't even at max).



As for the GBHZ's only working correctly on DFI mobo's.. I had a friend come over with a brand new DFI Ultra-D, and his GBHZ (this would be the 4th kit I have tested with) memory to swap around and test with..


My G.Skill HZ (2 GB) runs fine in the Asus A8N-E SLI Premium.

[kamakazibond]

btw i am still reviewing your problem. give me few days ok! thx man!
always welcome to shoot me a pm to remind to review it... too busy those days
If your too busy, I'll take a job doing this ;)

No seriously tho, I understand you are busy and me not being able to OC my memory past it's rated speed, the speed I actually bought, shouldn't be one of your priorities. I do thank you for the quick responses, and working with me on this.. It was most likely you who initially told me that the MSI NF3 could have been an incompatability issues with the original GBHZ's, tho that was done thru email, so who knows..

I have about 3 more days before my RMA for the GBZX from new expires, so I better hurry up and finish testing this stuff and make up my mind :)

An answer to the RMA thru G.Skill question posted here (http://www.xtremesystems.org/forums/showthread.php?t=95755), would deffinately help with my decision.

thankx again
kamakazibond

[Micketh]

could this problem also be found when using an asus sli se because my opty 144 wont get past 2.5 and i was wondering if it was the ram or mobo because i have tio run the ram at 266 divder to get it stable in windows.

[uOpt]

I don't think there is an "asus sli se". What exactly do you have?

I can only say that the A8N-E SLI Premium is trash. The A8V-E SE works just fine for me, including 4x 1 GB RAM at full speed and highly overclocked dual-core Opteron, and ECC memory (all at the same time).

[high5]

A8N-SLI SE exists.

but i must say i'm not sure about 'A8N-E SLI Premium' you're mentioning. i know of A8N-E and A8N SLI Premium...

[fhpchris]

I am Sending the A8N SLI ASAP, waiting for mushkin RMA on my end...

It should go out fridayish.

[GatorLCA]

Just an update on my status. I RMA'd the RAM back to Newegg and got the new set back today

...and

I reached 270MHz on a 1:1 @ 3-4-4-8 2T, must have been a bad batch going around but here's the CPU-Z validation

http://valid.x86-secret.com/show_oc?id=91392

[confusedsoul]

Well, I got it stable at 250mhz 3-4-4-8 2T. Haven't tried going 3-4-3-8 or anything like that. Sux my A8N-Sli Deluxe didn't seem to like 1T at those clock rates. :( Didn't try any higher voltage than 2.8v.

[kamakazibond]

well.. it's been a while, and I no longer have Gskill memory.. But a friend of mine still does, and there are some things I am not understanding about this memory.

My bud got his HZ's to 300MHz+ (306 I think) on his DFI NF4 UT DR-SLI with his X2 4800+. Another bud of mine has the same memory and a DFI NF4 UT DR, and his HZ's wont go past 255MHz w/o issues. We tried 2 different kits of HZ in his mobo and neither would go past 255MHz at 1:1. There is a thread I posted with a ton of info about that here somewhere.

Now here's the thing.. the same friend that had his X2 4800+ and HZ's up past 300MHz... Now has a 3400+ (3700+??) San Diego in his same mobo and with the same HZ's.. now he can't budge past 250MHz 1:1...

Could there be simply a compatability issue with some AMD64 chips??!!!? I know the Gskill techs test on DFI mobo's with a Opty 170. Maybe if they switched CPU's they'd find more errors with their memory??!?

We know that the San Diego can easily pass 250 HTT with a divisor set on the mem, and are 90% sure it went to 265MHz 1:1 with a set of 512MB x 2 BH-5 chips.

Just some thoughts.. Any ideas???

[uOpt]

My bud got his HZ's to 300MHz+ (306 I think) on his DFI NF4 UT DR-SLI with his X2 4800+. Another bud of mine has the same memory and a DFI NF4 UT DR, and his HZ's wont go past 255MHz w/o issues. We tried 2 different kits of HZ in his mobo and neither would go past 255MHz at 1:1. There is a thread I posted with a ton of info about that here somewhere.

Now here's the thing.. the same friend that had his X2 4800+ and HZ's up past 300MHz... Now has a 3400+ (3700+??) San Diego in his same mobo and with the same HZ's.. now he can't budge past 250MHz 1:1...

Could there be simply a compatability issue with some AMD64 chips??!!!?

Each AMD64 chip's memory controller has its own personality, all this is normal.

You should put that known-300 set into the latter box.

Any screenshots of the 300 available? Sounds very high.

[fhpchris]

G Skill Rep has had my board for a month or so now...

You guys should have some results eventually??

[kamakazibond]

Each AMD64 chip's memory controller has its own personality, all this is normal.

You should put that known-300 set into the latter box.

Any screenshots of the 300 available? Sounds very high.


I understand no chip is identical, and I can understand that maybe the San Diego can't do over 250, but it will with other memory or on a divisor..

So my question is more of.. are some AMD mem controllers picky in terms of memory and others not so much? Or is it that the memory simply works better on certain chips?

BTW the 300+ memory is in the same rig with the same mobo and only thing different is the CPU.. I think I know what u meant tho.. and unfortunately we live across the country from each other and that isnt gonna happen :)

I guess it's all about what chip you have with what memory, some chips can handle some memory well while others can't.. or some memory is picky as to what chip is running it, i guess...

It was all just food for thought. I'm quite happy now with my OCZ running at 2.5-3-2-5 245MHz on the 166 divisor, I'm working towards 300x9= 2,700MHz/11 = 245MHz... I'm totally fine at 2.6GHz 1.440v (1.475 set in BIOS) so i think 2.7 should be do-able.

[uOpt]

I understand no chip is identical, and I can understand that maybe the San Diego can't do over 250, but it will with other memory or on a divisor..

So my question is more of.. are some AMD mem controllers picky in terms of memory and others not so much? Or is it that the memory simply works better on certain chips?


Absolutely. There are AMD64 CPUs that do great clocks themself but fall short in the memory department.

I think my CABYE Opteron ran memory faster (at least easier) than my CABNE, although the CABNE clearly goes 100 MHz faster itself.