OCZ TCCD lowest vdimm test

[ahmad]

I am interested in what others can do at 2.50v :D . If you have the time, contribute!

My sticks are actively cooled (heatspreader still on) and I cannot squeeze a single MHz out of them without slackening trp and tcl some. If you have any suggestions on how I can get a few more MHz at 2.5-4-3, please do tell :)

(OCZ Platinum Rev. 2 v1.0 2x512MB).

http://img492.imageshack.us/img492/5246/des4xh.jpg (http://imageshack.us)

Updated.

[ahmad]

Woah... thats a messed up tweaker. So I have to do some more testing.

[Tony]

Woah... thats a messed up tweaker. So I have to do some more testing.
Just open cpu-z first, its a known driver bug

[ahmad]

Yeah I thought it was, thanks :)

Turns out I am only stable at 295MHz. 32m wouldn't run at 300MHz (at 2.5v).

http://img492.imageshack.us/img492/1588/a646lr.jpg (http://imageshack.us)

Now you can suggest timing changes :p:

[Bamboo]

This was done a while back, 3000+ with Geil One SE.

http://members.shaw.ca/tsopin/0708052.jpg

[.sentinel]

Where is voltage though.

[ahmad]

Hey bamboo. 2.57v.. thats cheating :p:

If you had an A64 I could compare, but very nice.

I can do 305ish with 2.6v I think but it requires tweaking.

Where is voltage though.

In mine? MBM reports it under DRAM.

[Bamboo]

OOPS :p: Anyway, yeah, it is a bit high. Following is the A64 setting.

http://members.shaw.ca/tsopin/070853.jpg

[Zeus]

A bit off-topic but what do you guys think is the max voltage one should put through TCCD for 24/7 usage?
I'm asking this cause i had a wonderfull Gskill kit that i ran at 2,5-3-3-7 300Mhz all day long with 2.86V

After 2 weeks one of the sticks broke and started to give BSOD's and errored in OCCT, not in memtest though.

The sticker on it says 2.6-2.9V.

I even had a 80mm fan blowing over them.

Ontopic: great results for low voltage! :toast:

[ahmad]

OCZ warrants to 2.9v, but I think just like your case Zeus, its a bit much. I had a few pair die on me at 2.8v as well, and I don't feel like any more RMAs.

Personally think 2.65v should be enough. Don't know how much they can withstand from a technical perspective. Maybe some gurus can help.

Thanks for the a64tweaker shot Bamboo.

[ahmad]

Damn... Just checked Samsung's data sheet:

http://www.samsung.com/Products/Semiconductor/DRAM/DDRSDRAM/DDRSDRAMcomponent/256Mbit/K4H560838F/DDR400_256Mb_F-die_x8_x16_11.pdf

http://img442.imageshack.us/img442/487/data2tv.jpg (http://imageshack.us)

2.7v is the max is what they say.

Check out this:

http://img442.imageshack.us/img442/5194/data13tx.jpg (http://imageshack.us)

What the heck does that mean?

:confused:

[ahmad]

Who knew a data sheet could be so interesting. After some more reading, I came across this:

http://img432.imageshack.us/img432/2959/op9po.jpg (http://imageshack.us)


Best operation is 2.7v @ 0*C? :slobber:

2.6v @ 25*C? I can't keep mine below 35...

Also, I managed to interpret the timings they have on that same data sheet:


cl: 2.5 (6ns per cycle)
trc: 9 (9.16)
trfc: 12 (11.66667)
tras: 7 (6.666)
trcd: 3 (2.5)
trp: 3 (2.5)
trrd: 2 (1.666)
twr: 3 (2.5)
twtr: 2 (2)
Read Preamble (tRPRE): 5.4ns
tref: 7.8us

I had to convert them from ns to clock cycles. According to the data sheet, its 6ns per cycle at a cl of 2.5. I am gonna give these a go at 2.6v soon.

Update: Timings gave me a few suggestions and for the most part worked. However I noticed tras must be strictly less than trc. Also a trfc of 12 is far too tight and nothing below 14 would post (above is unstable too). 2.6v made it get hotter and slower, so I kept it at 2.5v.

[Zeus]

Nice find Ahmad.

2.7V at zero degrees is a bit odd.

Most people give around 2.8-2.9V without a fan.
The average casetemp is about 35C

Hmmmm....

[ahmad]

Woohoo! Managed 300Mhz at 2.5v:

http://img496.imageshack.us/img496/1381/3004fb.jpg (http://imageshack.us)

I almost fell asleep during that run. No way I am going to run 32m! I am going to go for 310 and see what happens.

[bachus_anonym]

I've never tried "lowest-voltage racing" on any TCCD I had but this is what i dug out for now... I'll get back here later when I put them back in and see what more they can do at 2.5vdimm (BIOS). But keep in mind, that it's not really 2.5v, though. DFI NF4 as many, many boards overvolts by about 0.04v ;)

2x512MB G.Skill PC4400FC 275x11 6-3-3-2.5-1T, 2.50v (with DMM it's 2.54v) ;)

http://www.clockmehigh.com/ddr/GSkillPC4400FC/275x11_3025Mhz_SP2004_96mins_8x6.jpg (http://www.clockmehigh.com/ddr/GSkillPC4400FC/275x11_3025Mhz_SP2004_96mins_16x12.png).

EDIT: Damn! I just noticed - OCZ... Do other brands apply?

[ahmad]

Yes any TCCD sticks count. Change the title please, don't want to cause more confusion :). I am looking forward to what you can do.

Yeah the BIOS reading for my setting is 2.53v (2.50v), so I guess I'll trust the BIOS now. Can't go any lower than that!

I don't think its possible to pull off 300Mhz at 2.5-3-3 at 2.50v. Anyone care to prove me wrong :D (hint: challange).

[ahmad]

Anymore contributers?

[Digg]

small test with my ocz el pc4200 platinum v1.0 on abit an8 ultra:

270 2.5-3-3-7 1T @2.49v

http://mitglied.lycos.de/meinnamegehtnicht/OCZ/270@2.50.JPG

:cool:

[ahmad]

Nice timings. I see you dropped Read Bypass to 8x and you were still able to use 7ns for async latency! Very nice. I think I am going to go for a 2.5-3-3 run and see what I can manage.