Super Talent DDR3-1866 2GB Kit Testing

Printable View

Show 100 post(s) from this thread on one page

09-17-2007, 11:34 PM
metro.cl

Quote:

Originally Posted by v_parrello

I am a little bit confused because I do not understand if it is a trouble related to kit in my hands or it is a general trouble with these kits.

It seems very strange that Supertalent was able to build a kit with default specs DDR3-1600 7-7-7-21 1.8v that is able to work stable at 1866 8-8-8-24 1.9v, and after they build a new kit with default specs DDR3-1866 8-8-8-24 1.9v that are not able to respect default specs.

Then I have to verify if it is a trouble with sample in my hands or it is a generalized trouble with these memories.

I think is a general problem i tried testing at 1.8v and seemed to work better than 1.9v but i have had stability issues with the kit.
09-18-2007, 01:06 AM
saaya

Quote:

Originally Posted by jugeen

you can blame me :) ddr3 is an X-file. i am really sry if it sent back a good kit.

hey, its fine with me, i just hope your not sad that you swapped the kit since i think your replacement is probabaly not as great as the one you sent back ^^
i mean 1950 876 with 1.85v 100% stable, thats pretty damn good :D

Quote:

Originally Posted by v_parrello

I finished all my tests for this kit just yesterday.

My final verdict it is the same of BenchZowner. These memory kit is not superPI 32M stable at rated specs.
No way I tried a lot of bios settings but nothing to do.

They are stable at 1866 8-8-8-24 with 2.05v (a little bit better of the BenchZowner kit) but this is too much voltage.

I am very confused because DDR3-1600 7-7-7-21 performs better than this kit. DDR3-1600 are superPI 32M stable at 1866 8-8-8-24 with 1.9v :confused:.

With same setings I get out from ram slots DDR3-1866 and get into ram slots DDR3-1600, and the DDR3-1600 performs better than DDR3-1600 with every bios settings, timings and voltages. Very strange :confused:.

I think that in Supertalent made a big mistake with chips selection it is not possible DDR3-1600 performs better than DDR3-1866.

i dont think its a problem of the chips, i checked several batches again and i made a mistake before, i tested on the crappy old board that only does 1825 max ^^

anyways, the newer chips are the same as the older chips...
i think supertalent has a problem with their binning procedure...

vincenzo, did you try the sticks in single channel?

Quote:

Originally Posted by metro.cl

I think mine work stable at 1866mhz but is weird i can boot much higher but not stable

sounds like you have a good and a bad stick paired together, try them in single channel.

did you guys contact supertalent?
what did they say?
09-18-2007, 01:19 AM
hipro5

To whoever has problems....... :D

IF your motherboard has the "DDR3_REF" voltage features then put ALL THREE of them at +30mV.....and feed your Northbridge with ~1.85 - 1.91Volts AND (0.67x).....I believe that you'll get more out of them.... :)
09-18-2007, 03:19 AM
saaya

sounds like the blue kit finally arived? :D
george, i suspect the vdimm and vtt to be unstable on the p5k3, i hope we will soon see some nice clean ripple free vdimm and vtt from a maximizer :D

EDIT: btw, whats this x0.67 thing about?
i really wonder why asus adds options like theese and then doesnt explain what they actually do :confused:
09-18-2007, 03:19 AM
v_parrello

Quote:

Originally Posted by saaya

hey, its fine with me, i just hope your not sad that you swapped the kit since i think your replacement is probabaly not as great as the one you sent back ^^
i mean 1950 876 with 1.85v 100% stable, thats pretty damn good :D

i dont think its a problem of the chips, i checked several batches again and i made a mistake before, i tested on the crappy old board that only does 1825 max ^^

anyways, the newer chips are the same as the older chips...
i think supertalent has a problem with their binning procedure...

vincenzo, did you try the sticks in single channel?

sounds like you have a good and a bad stick paired together, try them in single channel.

did you guys contact supertalent?
what did they say?

Sascha,
I think you made a right question.
I already made these tests and the modules work well in single channel (1866 8-8-8-24 1.9v) but they are not able to work in dual channel.
Edit: Therefore I do not know if you are right when you saw that sounds like we have a good and a bad stick paired together.

I am waiting some reply from my contact in Supertalent.
09-18-2007, 03:28 AM
hipro5

Quote:

Originally Posted by saaya

sounds like the blue kit finally arived? :D
george, i suspect the vdimm and vtt to be unstable on the p5k3, i hope we will soon see some nice clean ripple free vdimm and vtt from a maximizer :D

Yeap they finally came....and already fed them via a DDR Maximizer......:D

http://www.xtremesystems.org/forums/....php?p=2438496

.
09-18-2007, 06:10 AM
saaya

:eek:
09-23-2007, 02:42 PM
BenchZowner

Started testing the kit for the review today.
Just finished finding the best timings for fully stable operation at DDR3-1333.
Un-tweaked SuperPi 1M run screenshot follows...

http://i14.tinypic.com/4xvo57c.jpg

Edit:
A Everest Cache & Memory Benchmark:

http://i4.tinypic.com/66t02z8.png

Used Performance Level 4 instead of 5 this time.
It's stable :D

Edit #2:

Oh well! Performance Level 3 as well :D

http://i4.tinypic.com/4zeicgi.png
09-24-2007, 06:01 AM
giorgioprimo

waithing for some more testing at higer clocks :)
09-25-2007, 05:59 AM
saaya

read the rest of the thread, i dont think there will be much higher speeds :D
maybe single channel... it seems one of the sticks is bad.
09-25-2007, 06:05 AM
dinos22

Quote:

Originally Posted by BenchZowner

Started testing the kit for the review today.
Just finished finding the best timings for fully stable operation at DDR3-1333.
Un-tweaked SuperPi 1M run screenshot follows...

http://i14.tinypic.com/4xvo57c.jpg

can you post something we can't already do with DDR2 :p: :D wink wink :D
09-25-2007, 06:56 AM
saaya

go easy on him, hes handicapped ^^
come on then bill, show us some nice SC then if DC doesnt work well with this kit.
09-25-2007, 08:16 AM
VictorWang

Quote:

Originally Posted by hipro5

To whoever has problems....... :D

IF your motherboard has the "DDR3_REF" voltage features then put ALL THREE of them at +30mV.....and feed your Northbridge with ~1.85 - 1.91Volts AND (0.67x).....I believe that you'll get more out of them.... :)

sometimes maybe not.

leave dimm_vref to auto seems better in most situation.
set mch_vref = +30mv when u wanna stable @ 1900mhz+
09-25-2007, 08:28 AM
stone_cold_Jimi

Quote:

Originally Posted by saaya

sounds like the blue kit finally arived? :D
george, i suspect the vdimm and vtt to be unstable on the p5k3, i hope we will soon see some nice clean ripple free vdimm and vtt from a maximizer :D

EDIT: btw, whats this x0.67 thing about?
i really wonder why asus adds options like theese and then doesnt explain what they actually do :confused:

Any more info on these 3 things? I haven't found the thread with the P5K voltage points again, if anyone knows where it is post it up plz, want to meter those. :up:

There was some suspicion that P5K3 Vdimm was actually higher than set in BIOS, no? And what are these "reference" multipliers about? Like you saaya, I'm bemused by them.
09-26-2007, 12:24 AM
saaya

about the voltage being stable or not, i have no idea... but the weird results some are seeing on the orange slots on the p5k3 dont seem to be voltage related after all.

George, how much improvement did you notice from switching to the maximizer on the blitz? would you say the blitz has good vdimm or is it rather unstable?
how much of 10 points would you give it, 0=can kill mem, 10=maximizer :D

and yeah, what exactly do the x0.67 options do?
09-26-2007, 12:52 AM
SOLDNER-MOFO64

DDR3 must still be in the "doesn't quite work correctly" stage since I can almost pull that with standard DDR 400 and cold temps. DDR2 makes those numbers look positively pathetic.

Perhaps they've mixed up IC's or summin' 'cos those bandwidth numbers are very low indeed. Not to mention it's using the same vdimm as DDR2.
09-26-2007, 01:16 AM
hipro5

Quote:

Originally Posted by SOLDNER-MOFO64

DDR3 must still be in the "doesn't quite work correctly" stage since I can almost pull that with standard DDR 400 and cold temps. DDR2 makes those numbers look positively pathetic.

Perhaps they've mixed up IC's or summin' 'cos those bandwidth numbers are very low indeed. Not to mention it's using the same vdimm as DDR2.

DDR3 ARE FAST.....Depents on who's working them correct.....;)

Quote:

Originally Posted by saaya

about the voltage being stable or not, i have no idea... but the weird results some are seeing on the orange slots on the p5k3 dont seem to be voltage related after all.

George, how much improvement did you notice from switching to the maximizer on the blitz? would you say the blitz has good vdimm or is it rather unstable?
how much of 10 points would you give it, 0=can kill mem, 10=maximizer :D

and yeah, what exactly do the x0.67 options do?

ASUS Vdimm is good enough - 2 phase circuit - .....I haven't placed it onto the osciloscope though to "investigate" it a bit...... :)
BUT I trust my Maximizer much more....Even at cold boot it helps.....;)
The 0.67Volts option - in simple words - it's like GTL_REF Voltage of a Core 2 CPU but it's the Northbridge's......:)
09-26-2007, 02:07 AM
saaya

Quote:

Originally Posted by SOLDNER-MOFO64

DDR3 must still be in the "doesn't quite work correctly" stage since I can almost pull that with standard DDR 400 and cold temps. DDR2 makes those numbers look positively pathetic.

Perhaps they've mixed up IC's or summin' 'cos those bandwidth numbers are very low indeed. Not to mention it's using the same vdimm as DDR2.

what ddr1 numbers are you talking about?
maybe a64+ddr1 can get similar bandwidth results, but even that would surprise me. a big problem with ddr3 is the fsb limiting the performance... we need csi to see the full bandwidth :(

and about the voltage, dont know what your talking about, ddr1 was used with almost 5v, ddr2 is being used with almost 4v, ddr3 is only used up to 2.3v. and in the end who the hell cares about the volts? what matters is the power consumption, and thats where ddr3 is a lot better than ddr2, the difference is so big that you can feel it with your own hands, you dont even need to meassure the temperatures. getting ddr2 to the point where you burn your fingers touching it is easy, good luck trying to get ddr3 to those temps... :D

Quote:

Originally Posted by hipro5

[B]
The 0.67Volts option - in simple words - it's like GTL_REF Voltage of a Core 2 CPU but it's the Northbridge's......:)

thats what raju said as well, but i dont quite understand it.
if i select that option the the chipset option will be 76% of another voltage, right? but which one?
09-26-2007, 02:48 AM
v_parrello

Quote:

Originally Posted by SOLDNER-MOFO64

DDR3 must still be in the "doesn't quite work correctly" stage since I can almost pull that with standard DDR 400 and cold temps. DDR2 makes those numbers look positively pathetic.

Perhaps they've mixed up IC's or summin' 'cos those bandwidth numbers are very low indeed. Not to mention it's using the same vdimm as DDR2.

Hey man,
what are you talking about? :confused:
Looking at DDR3 benchmark results and technology argumentations about DDR, DDR2 and DDR3 RAMs it is very hard to understand what you want to mean. :shrug:
10-17-2007, 03:59 AM
BenchZowner

Let me revive this thread :D

Got another P5K3 Deluxe some days ago, just tried the sticks on this one, and they seem to be running fully stable at DDR3-1868 8-8-8-24 1T @ 1.8V

Gotta test them more thoroughly, later on...this is getting weird...the rest DDR3 kits that I've got there work fine on the first P5K3 Deluxe, and reach very similar ( almost the same ) frequencies on both P5K3 Deluxe's :o
12-12-2007, 07:04 PM
metro.cl

Quote:

Originally Posted by hipro5

To whoever has problems....... :D

IF your motherboard has the "DDR3_REF" voltage features then put ALL THREE of them at +30mV.....and feed your Northbridge with ~1.85 - 1.91Volts AND (0.67x).....I believe that you'll get more out of them.... :)

Sorry to ask Hipro but isnt 1.85v way to much for X38 chipset?

I also tested upping all DDR3_REF to +30mV but didnt help, im on maximus extreme.

and 0,67x didnt help either :(
12-12-2007, 07:19 PM
BenchZowner

Quote:

Originally Posted by metro.cl

Sorry to ask Hipro but isnt 1.85v way to much for X38 chipset?

I also tested upping all DDR3_REF to +30mV but didnt help, im on maximus extreme.

and 0,67x didnt help either :(

You're asking the man who used 2.00V on the NB & 1.85V ( or so I'm not 100% sure on the SB voltage ) on 680i :D :rofl:

1.80V+ should be ok for benching with P35...but you gotta keep the chipset cool.

I know somebody who has even pushed 2.4V on 975X :D
Mobo still alive after all :D
12-12-2007, 07:47 PM
metro.cl

Quote:

Originally Posted by BenchZowner

You're asking the man who used 2.00V on the NB & 1.85V ( or so I'm not 100% sure on the SB voltage ) on 680i :D :rofl:

1.80V+ should be ok for benching with P35...but you gotta keep the chipset cool.

I know somebody who has even pushed 2.4V on 975X :D
Mobo still alive after all :D

Yes i know that hipro is crazy :D

But im on X38 now, and 1.8v on the NB is well in the "red" zone
12-12-2007, 08:53 PM
eva2000

Hipro is just stating what is needed to reach certain clocks.. whether it's safe of not - well that is another matter LOL
12-12-2007, 09:12 PM
metro.cl

Quote:

Originally Posted by eva2000

Hipro is just stating what is needed to reach certain clocks.. whether it's safe of not - well that is another matter LOL

what clocks? i think he just gave a broad recomendation. I've got more out of my DDR3 that anyone here in this thread and way lower volts than hipros recomendation.

Show 100 post(s) from this thread on one page