Elpida Hyper Graveyard

[gandalfone]

I JUST LOST one of my Corsair 2000C7 1GB stick.....
BLUE SCREEN during loading windows.....

I was OK on my latest benches and yesterday I booted to install new Vistas.....NO GO....

I then checked the rig with ONE - BY - ONE stick.....
ONE stick didn't booted at 1.65Volts and at 1333MHz.....

I used to work them with 1.72Volts with a DELTA 8 x 8cm fun over them...

I have some problem with OCZ Blade 7.8.7.21
3 weeks of life, max 1.75v bench, not daily...LOL.
Best problem with classi....post code 68 and , now I try one for one on p6t with memtest and no error but in tri channel I got error (All default...but if i set 1.65v one stick don't run)...mmm

[Leeghoofd]

So when you set 1.65 in bios, yet eg it overvolts to 1.76 (measured via DMM) in some cases : it's not the board, but still the ram ? Take the Gene II , preset 1800Mhz ram speed it overvolts to 1.8, set 2000Mhz it sets 2volts. Still the ram then ? I know most here are experienced users that hardly trust bios settings and always have a DMM nearby. Some users just trust AUTO settings or presets. I'm talking in the name of the latter, the plug and play users.

Look I'm just wild guessing here as the first Elpida that died on me (kingston Dimm) was the one closest to the IOH. Just trying to remain logic... not trying to talk nonsense as you so kindly write down time after time. The 2 OCZ kits are still running strong. The Kingston kit died on me the next day, during sub stock volt/speed testing.

I know these Ram IC's are fragile from the many posts here , but why does one dimm die and the other two still go on ? It's the same PCB/IC combo in the same kit ? Also explain me why the 2 Blade kits get very warm, yet the Kingstons are barely warm ? I think only Elpida can answer the real question. Did they maybe reduce quality control to speed up production ?

Sorry to be repetitive but it seems I'm being misunderstood (probably due to my lack of your mother tongue)

[gandalfone]

Hipro did it work on classi??

Now test blade on p6t one for one...first one module don't work at 1.65 but work at default voltage (and not work in tri channel, not post)...now run memtest in tri channel at 1.65v 1600Mhz...I don't understand

[drnip]

It is the RAM, more specifically the IC, why is the question we are trying to answer. I am sorry, but this argument about it being the board is nonsense. We have had failures on five different boards (ASUS, MSI, EVGA, Gigabyte, ASRock) at various voltage settings ranging from 1.50V to 1.80V VDimm and with VTT at stock 1.15V settings up to 1.375V.

It made no difference as to the supplier of memory either, OCZ, Corsair, GSkill, Kingston, Patriot, and now Mushkin. The argument about IOH temperatures being a factor is nonsense also. The memory has failed with IOH temps ranging from 47C to 79C and in various slots.

So, it simply comes down to the fact that either there is a problem with the IC (maybe PCB combo used) or a problem with the memory manufacturers recommending voltages higher than the 1.50V limit suggested by Elpida. On a side note, as far as the possibility of a bad batch, our failures have occurred with first lot parts all the way through to the latest lot.

6 boards as a set failed on my DFI LanParty UT X58-T3eH8 as well as my Classified. Now mind you my Blades failed on the 2nd day with only 1.65v fed to them.

[kiwi]

Any other deaths on AMD except cdawall's non-hypers?

Also just curious anyone tried burn-in? Like in the old winbond days

[leotonch]

Followed by 4 pack RMA to Kingston, 1 more pack to go....
Very light usage even at 1.5v for ram at DDR3-1066 VTT is 1.2v, 1 stick die ! RIP Hyper Chips & Kingston HyperX.....
Btw, all are the same batch ! Hope next batch will be better

[Leeghoofd]

Would it be possible to take a piccie of that batch mate.

@Kiwi, I performed the same procedure I took with the Blades, yet the Kingstons didn't cut it

[ToyTen]

I bought a 2000Mhz cas7 OCZ Blade in early may this year and I never got a single boot at their specs. One module was dead at the reception...

1.) After how much time did your memory fail?
at first boot
2.) Did you stress it a lot?
could'nt boot at all
3.) How did the memory die?
/
4.) What exactly happened?
no boot
5.) What board, pcb rev and bios version did you use?
P6T WS PRO / rev 1.01G / bios 0603
6.) What speed and timings did you run the memory at when it died?
/
7.) what speed and timings did you run the memory at most of the time?
2000 cas 7-8-7-20(1T) @ 1,66v(BIOS)
6.) What vtt did you run when the memory died?
1.35v
8.) what vtt did you run the memory at most of the time?
/

I still wait for those to come back from RMA.
Today i ordered OCZ Blade 12800 cas6-6-6-24 @ 1,65v. I wish ... I do not know what chips are on it.

[Gautam]

Did anyone kill elpidas yet WITHOUT ever having run SuperPi 32M, memtest or linpack?

[drnip]

Did anyone kill elpidas yet WITHOUT ever having run SuperPi 32M, memtest or linpack?

Both of mine killed over without these utilities. True I had run the utilities prior to the deaths but not during or even within hours of it. Simply cruisin the web, turned off computer just to turn it on the next day for them both to be dead. Separate mobo's too.

[ToyTen]

Did anyone kill elpidas yet WITHOUT ever having run SuperPi 32M, memtest or linpack?

I couldn't even start any those applications
sad story

[kiwi]

I wonder what the failure rate of these chips are

And why some modules still work fine

[jagt]

I talked to a guy at Corsair and he claimed the failure rates on the GT's were not any higher than on their other high-end kits. In fact, he claimed they've had less trouble with the GT's than they expected for a high-end part. Sounds highly suspicious to me.

[Dami3n]

Strange answer
Mine are still alive, but could die tomorrow. With all that dead sticks, you cant never know if your ram will be ok the next day

[LardArse]

1.66Volts in bios, it's 1.74Volts REAL...

3x1G? 1066? 1600?
As we all know about load lines, bigger the loads the lesser the volts and vice versa.
you will get a bigger droop @ 6x2GB 1600 vs 3x1GB 1066 for example etc etc

[ZillaPct]

Kingston KHX16000D3ULT1K3/3GX elpida hyper

1.) After how much time did your memory fail?
three monts
2.) Did you stress it a lot?
testing for the review, total 20 hours of prime&linpak and benching, rest of life daily use and little benching, max voltage 1,75V for little session of bench
3.) How did the memory die?
bios don't recognize one of three modules
4.) What exactly happened?
at boot
5.) What board, pcb rev and bios version did you use?
Foxconn bloodrage G25 & Asus gene 0809
6.) What speed and timings did you run the memory at when it died?
1600 mhz cas 6-6-6-18 1T
7.) what speed and timings did you run the memory at most of the time?
1600 cas 7-7-7-20(1T) @ 1,60v(real voltage)
6.) What vtt did you run when the memory died?
1.35v
8.) what vtt did you run the memory at most of the time?
1,30V

[hiwa]

Wlc zilla to the Hyper Graveyard

[SAE]

3x1G? 1066? 1600?
As we all know about load lines, bigger the loads the lesser the volts and vice versa.
you will get a bigger droop @ 6x2GB 1600 vs 3x1GB 1066 for example etc etc

Hey Peter,

I have to add that it heavily depends on used psu as well. I swapped my old backup psu OCZ Powerstream 520 (Enermax Revo85+ did not work) with my new Corsair HX1000 and overvolting decreased 0.05V (0.07V->0.02V).

Greetz,
Stefan.

[hipro5]

3x1G? 1066? 1600?
As we all know about load lines, bigger the loads the lesser the volts and vice versa.
you will get a bigger droop @ 6x2GB 1600 vs 3x1GB 1066 for example etc etc

I'm at 1066 strap with 3 (well now 2) x 1GB modules.....

I think that it's time to put my DDR Maximizer on to be sure.....

[bingo13]

Any other deaths on AMD except cdawall's non-hypers?

Also just curious anyone tried burn-in? Like in the old winbond days

I had two modules fail on the MSI GD70, this is simply an IC quality problem, plus a couple of bonehead design decisions by Elpida we are getting verified.

[bingo13]

Did anyone kill elpidas yet WITHOUT ever having run SuperPi 32M, memtest or linpack?

Yes, in fact had one module go on first POST with them in the board at all stock settings. It is quality issue with Elpida.

[bingo13]

So when you set 1.65 in bios, yet eg it overvolts to 1.76 (measured via DMM) in some cases : it's not the board, but still the ram ? Take the Gene II , preset 1800Mhz ram speed it overvolts to 1.8, set 2000Mhz it sets 2volts. Still the ram then ? I know most here are experienced users that hardly trust bios settings and always have a DMM nearby. Some users just trust AUTO settings or presets. I'm talking in the name of the latter, the plug and play users.

Look I'm just wild guessing here as the first Elpida that died on me (kingston Dimm) was the one closest to the IOH. Just trying to remain logic... not trying to talk nonsense as you so kindly write down time after time. The 2 OCZ kits are still running strong. The Kingston kit died on me the next day, during sub stock volt/speed testing.

I know these Ram IC's are fragile from the many posts here , but why does one dimm die and the other two still go on ? It's the same PCB/IC combo in the same kit ? Also explain me why the 2 Blade kits get very warm, yet the Kingstons are barely warm ? I think only Elpida can answer the real question. Did they maybe reduce quality control to speed up production ?

Sorry to be repetitive but it seems I'm being misunderstood (probably due to my lack of your mother tongue)

Leave my mother's tongue out of this.

I have no idea what is going on with your GENE board, I set mine to 2000 and VDimm to 1.66, get 1.672 at the slot. Anyway, we have had two failures with the Blades and others already, it is not manufacturer specific. This is simply a quality issue with the ICs, plus another side issue that could affect it. The memory suppliers are working with Elpida and all stand behind their warranties, thankfully.

[Raja@ASUS]

3x1G? 1066? 1600?
As we all know about load lines, bigger the loads the lesser the volts and vice versa.
you will get a bigger droop @ 6x2GB 1600 vs 3x1GB 1066 for example etc etc

Yeah, and you also get a larger droop within a 64 bit OS too with 6GB. 1.77V set on my E760 gives 1.732V in 3d mark 06 during the first GPU test. During 32m on XP 32 bit with maxmem to 600 or so the droop is negligible.

[Leeghoofd]

I have no idea what is going on with your GENE board, I set mine to 2000 and VDimm to 1.66, get 1.672 at the slot.

@Off Topic :

If you set the memory preset (OC Memory Level up) to eg 1600mhz what happens then at the Vdimm setting ? My board sets automatically 1.8Vdimm, at 2000mhz 2.0Volts in bios... we passed it to the bios engineers, plus ome other little issues we encountered...

I tried a reflash, popped the battery,... prob persists on my review board

[Sepyx]

Hi,

Add me as well. This morning suddenly one my TR3X6G2000C7GTF-DOMINATOR GT, 6GB memory sticks died.

I just booted up my PC and it was gone.

I suppose it has to do with your luck.