Too bad. Would be nice to test these double sided D9PFJ on cold and 2 volts :) Ballistix elite 8Gb kit is not easy to get in sweden.
Printable View
Too bad. Would be nice to test these double sided D9PFJ on cold and 2 volts :) Ballistix elite 8Gb kit is not easy to get in sweden.
I must admit, I've never messed with DDR VTT settings. I assumed it was automatically set to a certain % of VDIMM and would increase as you increased memory voltage. I never considered that manually tweaking it could increase stability since it never seemed to come up in overclocking guides.
VTTDDR: Supplies power to the VTT input pin on DRAM memory modules. In most cases this setting can be left on Auto. At high DRAM clocks (in excess of DDR3-2400) increasing this voltage may help improve stability. Start with 0.85V and work up. Traditionally this setting should be at 50% of VDIMM, however in our testing we have found 0.85V a good starting point for improving stability in Super Pi 32M. A one or two step change above or below that can help 32M pass where it would otherwise fail.
http://rog.asus.com/forum/showthread...C-Tuning-Guide
EDIT ; thanks I.nfraR.ed
There are also DDR3-1866 C9 1.5v D9PFJ based modules with the Elite spreaders, the all black ones, not black & yellow ones.Quote:
Don't buy modules with Elite heatspreaders, but yellow stickers. All these have the old chips.
Yes, that's what I wanted to say. The modules with high heatspreaders (like those Elite uses), but with yellow stickers does not have the new Micron D9 1.5V chips.
Those with blackish sticker are the ones we're interested in.
Yeah, might be. I said that in the thread context, because we're discussing new 1.5V stuff, presumably using latest Micron D9 chips.
CDON has those. http://cdon.se/hemelektronik/crucial...66mhz-19062487
Edit: 8-8-8-24 2400MHz 1.92v. Not too far from 9-9-9 2600MHz, failed after round 17. :S
http://i4.aijaa.com/b/00679/11922624.jpg
Edit 2: And here is 9-9-9-27 2600MHz 1.92v.
http://i5.aijaa.com/b/00731/11922837.jpg
Awesome,2x2 sticks run better it seems :) Try twcl 7,it will make a big impact on bandwith,also please make an aida screen with 5 ghz processor and 2600 9-9-9,it would be great.Great testing!
Too bad the 2x2GB kit costs 35€ in Germany ( it's out of stock anyway ) while the 2x4GB kit is only 15€ more...
I don't see 2x2 as an alternative for benching, if it really is single sided. Great clocks though :D
I might get lot of laughs, but is there big difference on performance single vs dual sided?:confused:
Looks like twcl 7 is working fine. Currently testing on W7 cause i also have to some other testing! :D
Edit: Well, almost working. :D
http://i8.aijaa.com/b/00774/11926676.jpg
0.9 VTT DDR didn't help me any unfortunately.
These are Ballistix Elite 1866 CL9 2GB (2 pcs) 1.5v BLE2G3D1869DE1TX0 8FMR (16FMR is 4GB stick).
Buyd those separately, 7,50€/unit. Here dusty picture of mem's.
http://i11.aijaa.com/b/00976/11927477.jpg
2x4GB Micron aren't an alternative for benching either... :p:
I just meant buying them for playing with them, but it's hard to justify when the 2x4GB kit is just a bit more considering the latter will be of more use in a 24/7 system ( and better resale value of course ).
I don't remember anyone testing it, but I think the difference won't be that big but noteworthy when benching competitively. :)
However since the 2GB sticks can do higher clocks than the 4GB sticks they might still be the better choice for benching.
In Spi32M single sided is very slow, several seconds slower than double sided at same timings/clocks. For 1M/pifast and such probably not much difference at all. On the other hand, what do we want super fast RAM for if not for 32m ;)
Why not, if 2400 8-8-8 is stable for 32m. Seems pretty fast, no slow timings anywhere, even trrsr 4? Compare to my (pretty bad) kit of blue ripjaws 2133 cl8 which does 2400 8-10-7 @ 1.8 volt.
.."several seconds slower than double sided".. i don't believe is "several" ,as far as i know 8gb vs. 4Gb is 0.5~1 sec ,correct me if i'm wrong (screene please)
EDIT; have this coming my way also http://www.ebay.com/itm/120980714297...84.m1497.l2649
Well, that's really a bad kit! :D
On a more serious note, I'd like to see a comparison between 2x2GB Micron and 2x2GB PSC or BBSE at the highest clocks on air, same OS, same tweaks. Then we could draw conclusions. I tried to compare the 32M to some others in the 5GHz challenge, but Win7 and most likely no tweaks and no waza make it pointless.
Don, what do you think about this Crucials? http://www.alfacomp.cz/php/product.p...00040000001UBS
I want it to order...
If you compare the results of 1600CL8 and 1866CL9 kits it seems like the latter are better, so I'd go for a 1866CL9 kit if I were you. :)
The picture or the specification from FlanK3r is incorrect there.
They say Tactical LP but the picture is from standard Tactical. Also 1600CL8 1.35v are for LP, not standard which is at 1.5v.
LP are D9QMT exceptional vor daily use. 2000CL8 1.6v, 2133CL9 1.5v, 2200CL9 1.55v may be even 2300CL9 1.6v and 2400 CL10-10-10-24 1.6v Memtest.
32M, 2150-2200CL8, 2450-2500CL9, 2600CL10...validation 2780CL11. Depends on how good the kit is.
D9QMT is almost as good as D9PFJ even better having higher frecuency stability....The only thing, don't scale above 1.72v but at 1.72 can do almost as good as D9PFJ at 1.92v.
I have the LP kit 1.35v, my result were on AMD, FX 8320.
I would buy it especialy for AMD.
received 8Gb set today 16FED2
https://www.dropbox.com/s/9kkpm24fzuz4hd8/CIMG1090.JPG
http://img543.imageshack.us/img543/8226/phpilc1gsam.jpg
Uploaded with ImageShack.us
And how they do? FED2, different marking on Tactical than on Elite, or other chips than D9PFJ ?