i never let my sticks hit 30c;)
even if i was runing stock i'd still put a fan
Printable View
Were those A-Datas D9 sticks too?
I can only comment on 2.4V and higher, I never test D9 at low voltages and I don't run them as my 24/7 memory, they get quite warm when fed with enough voltage, but I'm using a big fan pointed on the memory sticks when doing so. Never took the exact temperature of the memory stick, sorry.
A-datas were probably based on ProMOS, even with 2.7v they werent as hot as d9. :(
It looks like none of my kits can touch the Cellshock 8000 C4s (not even my newest 9200 C5s), 2.49V for 650MHz C5 MemTest and SuperPi 32M stable with my DK X48 and new E8500.
http://img7.imageshack.us/img7/6287/...5pl8cs2221.jpg
Damn, this ram is weird.
Yesterday it works working fine at 600 2.2v, today it gives random errors in memtest on exacly same settings. :(
It's not fit for using 24/7 at 2.4V.:)
Ambient temp is rather high, around 28*. But as I already said I bring a big room fan just for testing which is throwing air into case right on the sticks and mch.
Yesterday I could accomplish 30 passes of #5th test in memtest and 50 loops in linx, today I couldn't accomplish 30 passes in memtest on exacly same settings.
Weird memory.
Anyway I will try to run again my old a-datas and see what voltages I'll have to set to get stable 500fsb:600mem. That will assure me what mch voltage is required for such mhz.
vdimm was 2.675v bios set. Didin't run everest at those exact settings, but it would have been just a bit lower than the Everest SS I posted here
They look to be a good kit, how high will they go in 32M? I haven't really pushed mine either. How does the e8600 go on the X48?
Just wondering about the max FSB.
Second kit of TeamXtreem 8500's.
SPI32M @ 542 FSB / 650 MHz @ 2.33v / PL8 / 1.58 vNB
http://i720.photobucket.com/albums/w...33vPL8Kit2.jpg
2.33v for 650MHz C5 32M stable is insane. What chips are under the hood, GKX or GMH? Do you have the newer Team kits with black PCBs, those seem to really fly.
Got today other kit of double sided tracers but there are rated at 800 and they don't want to do 600 cl5 even at 2.5v, damn. :(
Hmm, are they used? Sounds like they've degraded, Ballistix 6400 16FD3 should do 600MHz C5 with 2.40~2.45V.
Yes, they are used.
Anyway from what I read on i4memory ram list in this can be found good gmh, gkx, gct and/or bad dqt.
I'm assuming that I might get dqt but I'm not sure, testing so far.
Oh and btw these pair is fd3 with clips while 8500 are fd5 without clips.
It's doing 575mhz 2.2v - what doesn't look like this sticks would be degraded, maybe overclocking d9 requires adjusting others voltages/alphas than overclocking promos.
I think the methodology I employed for benching my RAM was flawed. I'm now going to forgo MemTest since it's useless for high clocks and go straight to booting & benching. It's not like I'm trying to get something like 650MHz 24/7 stable. With that said, my new Cellshock PC2-9200 kit only needed 2.40v BIOS set for 32M stability.
http://img12.imageshack.us/img12/434...5pl8cs2221.jpg
The 9200 kit's sibling, the 8500 C5 modules with D9GMH chips, needed 0.10v more for 32M.
http://img32.imageshack.us/img32/628...5pl8cs2221.jpg
Shiranui Gen-An - 32m spi stable doesn't make them 100% stable, if you are using memtest for testing memory don't do all tests, to get a basic conception will they work fine at desired clocks you can run few times 5th test of memset - if memory is unstable this test will lets you find out this in the shortest amount of time.
zsamz_ - maybe it's just me who can't oc it stable, maybe it's temperatures outside and inside - now it's really hot in here, too hot.
As I already said I'm using room fan to cool them.
DQT/DQW is not bad, it's just old fatbody D9 and was discontinued when the RAM list was made by Ronald, that's why it's red. It was once the stuff to get. :yepp:
As far as I know fatbody IC's weren't used on FD3 sticks, my bet is on D9GCT too, G.E.I.L. once had a set of 9200s I think it was, you might want to look for reviews and compare them to your results.
575MHz CL5 isn't what fatbodies could reach with 2.2V anyway.
I never found a kit that used D9GCT, which revision of Tracers is it?
I think it maybe something other limiting my sticks except heat ofc. :p
Another test with the X48-T2RSB Plus and my naked TG kit:
500 FSB / 600 MHz @ CAS 4 / PL10
http://i720.photobucket.com/albums/w...600MHzCAS4.jpg
Everest @ 500 FSB / 600 MHz @ CAS 4 / PL10
http://i720.photobucket.com/albums/w...600MHzCAS4.jpg
Cryptik what volts?
with the multplier at 6x you lose the fast memory access
use an higher multiplier can help you to have an higher bandwidth and lower latency
BTW you have really good tg kit :yepp:
Cryptik, what are yours subtiming and pl?
i can bench more of you at 1200 5 5 5 15, 5:6 divisor, 333 strap and pl7 on p35...
2.6v
Yes, I had low multi on purpose, and very loose chipset timings. I didn't care about performance, I just wanted to hit 600 MHz CAS4.
That was by no means an attempt at performance. It was PL10, with very loose chipset & sub timings, and as mafio pointed out, 6x multi and very low cpu speed which reduces performance even more.
Guys does any of you overclock your d9 on rampage formula?
I'm experiencing some real weird behavoiurs and I'm not sure is it ram fault or mobo.
I can get working 575mhz on mem with 1.25v on mch and 2.2v on mem - no problems, 100% stable.
BUT
590mhz on mem with 1.45v on mch and 2.45v on mem gives me plenty of errors in memtest.
My conclusion is that such overvolting doesn't look like it would be memory problem, giving 0.25v more and can't get 15mhz seems weird. What you guys think?
Played around a bit and found 668 MHz to be around the max SPI32M stable frequency attainable with this particular kit. At 670 MHz I got to loop 23 and it gave the out of round error. more voltage on the ram and NB didn't help and I tried a few combo's of drive strength/skew model but couldn't make it pass. Seems to like 1/2/7/7 for the Team Groups.
Also managed to get 700 MHz CAS5 screen. None of my other boards have ever let my ram go that high.
SPI32M @ 445 FSB / 2:3 / 668 MHz / 5-5-5-15 / PL8 / Team Group 8500
http://i720.photobucket.com/albums/w...B668MHzPL8.jpg
700 MHz / 5-5-5-15 / 2:3 / Team Group 8500
http://i720.photobucket.com/albums/w...700MHzCAS5.jpg
So Hello everybody :) . Im kinda new in this thread , so ill just say that im from estonia and iv been readin Xs for a long time now . And i was think to test abit my rams . So here it goes :D
Test Rig:
Intel Pentium Dual-Core E5200
Team Xtreem 2x1gb DDR 800 PC2-6400 TXDD2048M800HC4DC
Gigabyte EP45-UD3P
Inbox cooling :D
Noname 600W PSU
Club3D HD3850 256mb DDR3
And the screen :
http://www.upload.ee/thumb/121047/13.578.JPG
In the test it was 2.3V . But i managed to get 660mhz . But thats not the "redline" . So ill try some more tommorow . Sorry for my english , its not that good :) . So lets see what happens :p: . :wave:
1200W,very nice OC there.will be waiting to
see what you get tomorrow.P.S you english is fine
So i had some more testin . And found the best bios for "air" :D . Well ok heres the screen :
http://www.upload.ee/thumb/121282/680mhz_superpi.JPG
Ram - 2.380V
I just tested it 1 mhz at the time . :p: Well , I am really happy for the result . And still its not "redline" . I think , i have to change my inbox cooling for some better one :yepp::ROTF:
Nice work 1200W, that board is well known for its ram clocking abilities. It would be interesting to see what you can pass SPI32M at.
What is usually voltage required by x48 mch to achieve 500fsb, 600mem and pl8?
Just like in Germany, it's been >30 degrees for the last 2 weeks.... not good for overclocking... :(
Well, there are some other things you can do during summer.... :D
I still have to test my OCZ Blade LV, and I found a kit of OCZ Titanium Alpha VX2 on ebay, should be quite good, specs are 500MHz CL4, 2.3V for those who don't know these kits.
could some please explain how efficiency
works in Super Pi? i see people talking
about great efficiency,but i dont understand
how it calculated,(how does someone tell)
if its getting good efficiency.
thanks
Probably around 1.40v, depends on a few factors though.
For full stability my DFI X48 needs 1.42v for 500 FSB, 600 MHz, PL 7, chipset optimizations set for max performance.
http://i720.photobucket.com/albums/w...vPL7MoreAg.jpg
As far as I know, Seconds x MHZ = efficiency
So say you take 7.8 seconds at 6000 MHz for SPI 1M. That's 7.8 x 6000 = 46800 (or 46.8K you may see it expressed as).
From what I have noticed, 45 - 46K seems to be considered good efficiency. 48K is not so good.
I'm by no means an expert on this though so if I am wrong, please correct me.
1200w and cryptik your TG looks very good :yepp:
thanks CryptiK,now i get it.
so this run here is getting 9.3x5004=46537(46.5 efficiency)
http://i42.tinypic.com/14aext.jpg
I got hold on some Kingston HyperX KHX9600D2K2/2G, they should be Micron D9GKX, I know for sure when I take those heatspeader off.
http://i282.photobucket.com/albums/k...y/KH1200_1.jpg
http://i282.photobucket.com/albums/k...y/KH1200_2.jpg
http://i282.photobucket.com/albums/k...y/KH1200_3.jpg
Hope they go well. Be good to see some results when you have taken them for a spin.
Those Kingston HyperX was crap. :down:
Can not boot in to Windows XP @ rated speed (1200) @ rated voltage (2,30 -2,35).
Canīt even pass Memtest86+ in single channel with up till 2,40v, 3 sticks freezes at start and the fourth within 30 sec.
Have tried voltage from 2,00 - 2,40v.
How about that guys, are they crap or what.
RMA, here I come. :mad:
Really disipointed, I have 2 other set of the same kind and they are really good.
That's disappointing. They could have been good kits. Hope the replacements perform better.
There was a review at a domain that is no longer available that tested the newer HyperX 9600 kits with Elpida ICs and they couldn't hit 600MHz either.
Okey guys, I took a sneak peak under those HyperX heatspreaders and I can confirm that they are Elpida E5108AJBG.
Itīs the same kind as my 4x1 Kingston HyperX PC-8500 have, I took a look in my log book and it said that I could boot into Windows @ 1200C5 with 2,24v (didnīt go higher) and SKEW advance 300ps but it was it.
The best the could do was 1128 4-5-5-15 Moderate tRD6, NBv 1,31 and DRAMv 2,12 @ bus 450 or 1066 5-5-5-15 Stronger tRD6, NBv 1,31 and DRAMv 1,86 @ bus 400, all the above was with my Rampage Formula.
Run Elpida @ 2,30 - 2,35v seems to be to much, according to the data sheet 2,30 are the absolute maximum voltage.
They gonna retur to sender :yepp:
Thank you Shiranui Gen-An for the link, much apprecieted, it was to some comfort. :)
http://www.trustedreviews.com/cpu-me...-Memory-Kit/p1
Frankly, I'm wondering how Kingston can market that crap as DDR2-1200.
Somebody should scatter a email to Kingston telling them how :banana::banana::banana::banana: those sticks are and give all the technical details why. On a brighter note, I installed a Q8400 in my rig today :D not really tried OCing yet just finding its FSB wall, going @ 460MHz so far and counting :up:
Ket so you got yourself a quad,nice
looking forward to some results
Workin on it, limited by how far I can push the FSB because of the 8x multi. Took a small gamble on I would get a quad that can do 470-480 FSB. So far I'm testing 450*8. That in itself isn't shabby for a quad, but I would like a little more so I can run a faster DRAM frequency than whats (to me) a rather mundane 1125MHz. I'll be satisfied with 1150MHz+
i hear you on that,thats why i sold my Q9650
couldnt get the high fsb memory clocks that i was used
to with my dualcore,the highest i could get stable was
518fsb.and after experiencing 550fsb w/the dual core
and my cellshocks running 5:6,well the quad just did
not want any of it.
Pushed my new 8600 little bit to get a better idea of what this board can do.
SPI32M @ 4703.5 MHz (523 x 9) / 627.1 MHz / 5-5-5-15 / PL 7
Time = 9m 58.953s
http://i720.photobucket.com/albums/w...6271MHzPL7.jpg
Everest @ 4703.5 MHz (523 x 9) / 627.1 MHz / 5-5-5-15 / PL 7
http://i720.photobucket.com/albums/w...271MHzPL72.jpg
nice one CryptiK,thats a nice setup you got now.
E8600and the DFI X48-T2RSB plus.i think im going to get
an E8600 too.what the vid on that one,and where did you
buy it at?
Yeah I bought it for benching with so I don't have to worry about hurting my nice low voltage 8400 which now stays in the M2F for my 24/7 system.
The VID is 1.1125v. I don't live in the US, I just bought it from a local e-tailer.
wow,nice low vid.
Yeah, but it's uncontrollable on air. I have a TRUE copper with a 2000 rpm Scythe Ultra Kaze + another 120mm fan blowing on socket/PWM and it hits 60 - 70*C so fast under load. It wont even boot with 1.50v BIOS set, not sure why. Think I my need to get another one and hope for a VID around 1.16 - 1.18v, might be easier to cool.
Yeah its a real pain in the arse. I'm loving this quad for all the extra power it gives but I'm really not used to seeing sub 9.5GB\s memory bandwidth results :down: I think I'll shoot for using the 266 BSEL strap and hope I can keep my current settings, that'll boost things enough for me to be relatively settled.
I did a little Everest bandwidth testing. Seems 565 FSB on the 5:6 divider using PL 8 is the max I can stabilize enough to bench, and that took 1.78 vNB. Think this is about as far as X48 and DDR2 can go on air on this setup.
555 FSB / 666 MHz (DDR2-1333 MHz) / PL 8
http://i720.photobucket.com/albums/w...B666MHzPL8.jpg
560 FSB / 672 MHz (DDR2-1344 MHz) / PL 8
http://i720.photobucket.com/albums/w...B672MHzPL8.jpg
565 FSB / 678 MHz (DDR2-1356 MHz) / PL 8
http://i720.photobucket.com/albums/w...B678MHzPL8.jpg
So close to 12K across the board!
this any good?:D
Thanks, I do want to get 12K across the board with DDR2, not sure if I will get there on air though.
Very nice, a few more FSB and you will have 13K across the board. I want to get a REX to play with, think its time for me to go DDR3 + X48 before I start abusing i7.
you need ut p45 or ut p45 for 12k
i coulda had 13k across the board it if i just tweaked it but was past 3 am lol
i gotta say the rex is an amazin board and for the price now its hard to resist
but if you get the rex get a tri kit so you can pass it allong to i5-i7
i think i5 gonna be better;)
Without cranking the CPU up to ~5.08GHz to get the read up that last 203 MB/s, I don't see this getting any higher.
570 FSB / 684 MHz (DDR2-1366 MHz) / PL 8
http://i720.photobucket.com/albums/w...b684MHzPL8.jpg
So I've here my Cellshock DDR2 800 with Micron D9GMH
I didn't got the best but ok for a beginner like me
http://img35.imageshack.us/img35/2719/cellshock1320.jpg
Validation at CL5-6-6, 2.58V with 1333MHZ (666,5)
http://valid.canardpc.com/show_oc.php?id=601968
At CL3-3-3-9 with 900
http://valid.canardpc.com/show_oc.php?id=601968
And here my AData vitesta Extreme with Micron D9GCT
http://img33.imageshack.us/img33/491/adata1170cl5.jpg
Validation at only 1140
Will get Cellshock DDR2 1000 CL5 with D9GKX, i hope they will do better
not to forget, all on air cooling--------------:welcome:
aslanwas flash f13 or f14 bios
BTW good work
No I didn't try using them actually, I tried PL 7 but it didn't POST with 1.78 vNB. I'm so used to my Asus board with pullins that dont work at all I just forget about them. Might be worth a try, but I don't see it making it into windows let alone giving me 200 MB/s read.
Some of my results on A-Data 2x1 800 cl4, dunno what ics:
Lower PL would be probably possible with other divider than 1:1.
500 5-4-4-4 @ 2.02v
600 5-5-4-5 @ 2.4v
620 5-5-5-15 @ 2.4v+ don't remember exact voltage
Anyway from what I have heard it's ProMOS not D9.
http://img30.imageshack.us/img30/5876/clipboard02c.jpg
Max valid: http://valid.canardpc.com/show_oc.php?id=586151
pretty good fsb on rampage formula :yepp:
I picked up another kit of ram, and I am quite surprised with it's performance. It happily runs through loop after loop of memtest Test #5 at 620 MHz @ 1.98v! Basic timings set to 5-5-5-15, trfc 30, trrd 3, rest auto. With this kit I also managed a low voltage SPI32M personal best of 650 MHz @ 2.19v (idle) / 2.16v (load).
SPI32M @ 542 FSB / 650 MHz @ 2.19v / PL 8
http://i720.photobucket.com/albums/w...dimmTG6400.jpg
CryptiK where did you get those?
*stares through the window looking in* I'm not a part of this anymore :(
No, but I have a quad. Limited by quad FSB :( Most i'll probably be able to get is 500FSB.
Further testing the new kit. It only needs 2.25v (idle) / 2.22v (load) for 660 MHz SPI32M stable.
SPI32M @ 550 FSB / 660 MHz @ 2.25v / PL 8
http://i720.photobucket.com/albums/w...225vTGCAS3.jpg
Everest @ 550 FSB / 660 MHz @ 2.25v / PL 8 (same settings as SPI32M run)
http://i720.photobucket.com/albums/w...225vTGCAS3.jpg
Around 2.0v depending on the kit.
SPI32m isn't good for testing OC, my crucials 1066 are 600mhz SPI32M stable at only 2.12v but 5th memtest test gives errors after few seconds of benching on same voltage.