I think most ppl hit the wall at 260.
Beyond that its 1m superpi only.
Mine still has very few errors at 260 but i can run superpi 1meg a little over 270.
Printable View
I think most ppl hit the wall at 260.
Beyond that its 1m superpi only.
Mine still has very few errors at 260 but i can run superpi 1meg a little over 270.
I managed to run 32mb at 265 but was slower than at 260??
When was the last somebody received TwinMos SP 3200 512mb from Newegg? Looking to pick up another stick since price drop. Mine is 44D just wondering what Newegg has in stock now.
its still the good stuff there, aa4t 44d's
My Twinmos SP w/ AA4T 44D from overclockers.de do 268 2-2-2-6-7-15-2-2-1-1 1T memtest 20 passes without errors at 3.6v.
They also do 270 with 3.7v and same timmings, but I have 2 errors by pass in memtest #5, but it is prime estable anyway...
:toast:
TEDY, nice of you to share how much is CFM ;)
lol that was just conversion so that i know for my fans :P
Got my pair today through the post!!!!! :toast:
So i've got
1 Gig of Twinmos SP 3200 AA4T 44D
1 Gig of Twinmos SP 3500 1A4T 44B
Still waiting for a psu & my hd to be delivered!! :mad:
TEDY i've just noticed on your chips it doesnt say BH5, but on mine they do :confused:
At least that will add $$$ to the value, if i sell them
I still don't think there's going to be a damn bit of difference in the chips at high volts with 2x512 sticks... ch-5 vs bh-5 I mean, someone prove me wrong, I would like to see better results from bh-5...
But the BH chips will do 2-2-2 at lower voltage than the CH since CH really likes 2-3-2 and not 2-2-2.
so basically they perform the same at high voltage but bh has the edge at lower voltage? My AA4T sticks will do 250 1.5-2-2-6 with 3.2v. This is just as good as my bh-5 (Kingston HyperX).
Are you referring to the old BH-5 and CH-5 or the new UTT's?Quote:
Originally Posted by cartmanea
That may have been true with the old dies, but there's lots of UTT-CH that's hitting 250+/3.3v/2-2-2. :confused:
When I say UTT does better at higher voltage I mean over 3v. UTT is untested CH chips, there is also new BH in production. So, UTT still performs like CH-5 did, at lower voltage it can only do 2-3-2, but give it 3.2v+ and it will do 2-2-2. Granted, they may have made improvements since we are seeing 250+ fsb, but A64 is definitely helping there. I had some SP UTT and it topped out around 245 at 3.3v on my NF7-S.
Mine does 255MHz, 2-2-2-x, with 3.2v with ease.Quote:
Originally Posted by cartmanea
ALONSO said they made some mistake in China...3200 should be CH-5 UTT but mine are BH-5 for sure....Quote:
Originally Posted by Owen1978
or else they wouldn't be working 200 2-2-2-5 at 2.5v :D
It maybe this has already been posted here, but:
Duonger - a Mushkin rep - shed some light to this whole UTT-chip case in this post @ OCForums.
I have had four sticks of the new SpeedPremium 3500 w/BH5 for two days now.
Max OC in single channel config (stick no.):
1 262MHz
2 253MHz
3 250MHz
4 246MHz
------------------------------
In dual mode (128bit)
approx. 235MHz (doesnt matter which stiks is used)
------------------------------
All the tests where run at 3,5V@2 2 2 8 1T
4 full rounds in memtest86 w/no errors = "test ok!"
DFI nF4 Ultra-D w/3000+ winnie
Personally, I think your Winnie memory controller is a POS. You should be seeing better dc results than that. Maybe a Venice in your near future?Quote:
Originally Posted by ArcticOC
Dual channel and memory controller issues aside, seems like the individual test are no better (maybe worse?) than the average Twinmos SP 3200 512mb (CH) stick results. I was expecting better from BH. This is not encouraging. :( 3.5v seems like plenty of vitamin V.Quote:
Originally Posted by ArcticOC
Yup, agree with the poster above. I have a TwinMos SP UTT stick supposively with ch die and does 260htt easily with 3.5v. Haven't really burned it in yet but not sure if the ch burns in like bh die. Should have a few more sticks of AA4T from Newegg in a few days. Will post my results.
well they did say originally that ch-5 was supposed to better bh-5 but the original process didnt go to good. Now they perfected the process and ch is really better than bh. Now we just have to wait for DH moduals.
Just burning my Speed Premium BH5 in at 245Mhz 2-2-2-5 3.5V, at which it is 20 passes of memtest 5 stable. It wouldn't do 250 out of the box at any voltage, and I get errors using 3.6V or more, hopefully burning in will sort that.
Looks like the CH5 might be the better bet.
I got my two sticks of 1A4T as well today, and it definitely says BH-5 on the chips. I decided to put them both in at stock timings, stock fsb and stock voltages (2.5-3-3-8 , 216Mhz , 2.7v) just to make sure that they work at all and I get lots of errors in memtest #5. I even upped the voltages and relaxed the timings but still errors :( Surely it doesn't need burning in to work at SPD settings right?
Did anyone else bother to test at stock first before overclocking?
Note, on the same PC I have just taken out some TCCD that ran 100% stable at a much higher FSB!
Ps, Minsta, glad to see you've taken a brief timeout from your dissertation to get us some results, yours are looking good - unlike mine :( Also, I've rated your review over at Komplett "useful" :)
Quote:
Originally Posted by [timko]
Same problem here. My sticks need 2.9v to run 216Mhz SPD. Really disappointing. Shame on Twinmos :mad:
The legend is back? Yeah.. right :rolleyes:
timco cool.i been getn flack from thge hardcore girls here.I get errors at 200 fsb.I should burn it like most do then send it back,but i dont do that.I take whats given me,and live with it.This is the slowest mem i have ever bought,but some peeps claim to get 250+fsb with 2-2-2-5 timings.It could be they sell this mem and hand picked the modules.I myself would love to get a pair of 512 that do this.If u buy these sticks from twinmos get the cheaper ones as u most likely will find they "suk".my offer tp prove this still stands .Send me ur 255 to 270 fsb and i will send u mine.Im lucky to get 200 fsb 8-4-4-2,and still error.pfff twinmos suks hard!!!!!!!Good fer ya
It's good to hear that's it not just me so far then.
Well, I'm going to leave mine "burning in" at 215Mhz (yes that's two one five) with 2-2-2 timings overnight at 3.2v and see if things get any better :(
With no sig's I have no idea what you are running them on or at what settings...complete settings.
If they are normal un-redesigned BH5 then they will act like BH5. So in NForce2 boards they will suk azz.
Chester did you get 1A4T "new BH5" sticks? 2x512M? Ill trade you my 1G CH5 that does 235Mhz 11,2,2,2.0 2T for them?
Hey mate.Quote:
Originally Posted by [timko]
Yeah posted up a few results while I was having a break from this stupid dissertation. Would have posted "back home" but my GF won't give me my password back (I knew it was a bad idea).
Will try my modules at stock tonight, contemplating giving their CH5 a try too if these don't burn in much.
Sorry about that.Quote:
Originally Posted by uwackme
Im running them on a 3500+ CAAZC good for 280 tight with my old pair of BH5 or 300+ with TCCD. MOBO is a DFI nF4.
Ok, here's my setup BEFORE I put in the Twinmos...Quote:
Originally Posted by uwackme
NF7-s v2.0 with NO wire mods of any kind.
Manta Rays D26 XT CPC On (1T)
XP-m 2500 IQYHA + SLK900a
2 x 512Mb Patriot XBLK (TCCD)@ 2.9v and 2-3-3-11
...and here's a pic of it running rock solid at 8mb SuperPi 238Mhz FSB (can also do 16Mb and 32Mb and also hundreds of rounds of memtest #5 error free but I didn't get a shot of those)...
http://img200.exs.cx/img200/5615/timko8mb7272io.jpg
When I switched out the TCCD and replaced it with the two sticks of Twinmos 1A4T BH-5, I set the timings to SPD as advertised -> 2.5-4-4-9, I set the FSB to stock for the RAM which is 216Mhz and put the vdimm to 2.9v... and these are the only settings I changed. This results in memtest #5 multiple errors :(
So as you can see, the rest of my rig can do a higher FSB with absolutely no issues with the TCCD RAM so why can't this BH-5 operate error free at a lower FSB?
RMA Time?
I'm guessing the deadline for your work will be sometime during the first two weeks of May yes? (At least mine was when I did it :) )Quote:
Originally Posted by Minstadave
Plus, you should have got your gf to change your ISP password instead! :D
bwahahahaha.sry:-]I tried these sticks in 3 nforce2s and am a64 3000.I wont even bother putting them in my old p4 423 cuz thats my second worst ram rdram grrrr.U know they ?
may work in my old timex bejuseus outta them on that yea.maybie just bad stiks,but i usually get what i pay for.lemme look through my closet thiss ram is good fer somn<;-(
Timko, I've tested 5G's of BH5 and 1G BH6 and none of it will run for shieeet in MY NF7-S....with Vdimm Vtt Vdd mod's!!! Stock NF7 is even worse. But the UTT does 245Mhz 1T 8,2,2,2.0 on the SAME board.
I have to burn the CPCoff bios back into the board when I want to switch back to CH5/BH5 sticks. Using D26 MantaXT cpcon and cpcoff versions.
For you....
First, it NEEDS the CPCoff bios. For some reason BH5 almost never ever runs 1T on the NForce2 chipset. Second its gonna need voltage. CPCoff it will do 200Mhz+ 11,2,2,2.0 2T @ 2.9Vbios. But to shine on NForce2 its gonna need TheVolt and it will never shine 1T on NForce2.
On an A64 board the results will be much better, 1T, lower Vdimm, etc.
These UTT/VX/SP sticks are a different breed and shine 1T high FSB.... but again NOT on a stock board, you gotta get to 3.2+Vdimm to make them work on a NForce2.
So confirm 200Mhz 5,2,2,2.0 with CPCoff 2.9V first, if it still screws up consider RMA.
Not entirely true. I had 2x512 Mushkin Promo Build BH-5 that did 225, 2.9v, 2-2-2 on my NF7-S with both Abit bios and modded bios, all with CPC on. The loss in performance from going CPC off needs an extra 30 MHz or so to make up for it, not worth it just to say you have higher fsb.
The only real BH-5 I had was Kingston HyperX 2x512MB they did 245 2-2-2-x 1T DC 3.3V on DFI nF2 UI.
@chester: There is no such manufacturer whose every single product they sell works without fault.
You happened to get a faulty pair/stick. No point in wailing about them. RMA them.
My 1GB of TMSP CH-UTT did 250 2-2-2-11 1T DC on DFI nF2 UI @ 3.3V. On current A64 they they freak out in Memtest: 20 rounds of #5 errorless then BAM! 30 errors in single pass. Mostly even #2 & #3 do ~kazillion errors. (BAAaad A64 3k+ CBBID0505) Edit: Hah! Got them stable! tREF is a magic timing.
Got 4 sticks of Twinmos SP 3500.. 1A4T , 44B.
1 pair does 250-2-2-2-5 @ 3.4v blend,memtest,32M stab
Just startet playing with the other 2 stick. So far 32M stab, 245-2-2-2-5 @ 3.4v, will be testing more soon.
This on DFI NF4 ultra-d with Winnie 3000+
How can you run with 3 sticks?Quote:
Originally Posted by ArcticOC
:confused: A4AT :confused:Quote:
Originally Posted by Fosco
Maybe 1A4T or AA4T? Have I missed something?
Keep us informed, I'm getting almost identical results with my SP3500, doing some burning in to see if it helps.Quote:
Originally Posted by Fosco
After testing Twinmos SP UTT on intel ( FSB 210 max 2/2/2/5 @ 3.4V ...)
I just receive 4*512Mo of SP 3500 1A4T BH5
FSB 238 2/2/2/5 @ 3.0~3.1v still on burnin
Stable is more than 10 passes memtest #5 flawless
Tested on Asus P4C800E-Dlx
P4C 3.0 SL6WU
1A4T sorry :rolleyes:Quote:
Originally Posted by largon
Guys, just received 2 more TwinMos SP 3200 512mb week 0506 44d. I must say I truely impressed with these $55 dollar sticks. So far one stick burning in nicely at 260htt with 3.35v, passed Superpi 2m with first boot.
PCB is still KO-6633. These are more consistent from stick to stick compared to Mushkin blue it seems.
Now lets see what 3.5v gets us. :banana:
RacerX, where did you get your TwinMos SP 3200 at?
Quote:
Originally Posted by Toro 45
Newegg I'm sure. :)
yup, Newegg. Seems both sticks maxing out around 263htt no matter what volts. Noticed I need to increase vdimm to atleast 3.45v as it decreases while stressing or buring in the ram, go as low at 3.3v which isn't good. Need to replace the pot on my booster or go with OCZ Powerstream PS for more stable volts.
Tell ya the truth if my cpu didn't have such a POS memory controller then I would be more happy with 300htt slightly looser timings then 260htt tight.
So your buying the individual speed premium sticks, have you tried the Twinmos dual packs from Newegg? My dual pack doesn't perform nearly as well as your speed premiums(237max) I was thinking about picking up the Mushkins but guys are really getting inconsistant results with them. I think I may have to try a couple of the 512mb speed premiums, what timmings are you running with them on your nF3?
I just noticed your using a DDR Booster, did you mod it so you can use 2 sticks are are your results from always running one stick and the DDR booster?
Got my week 48 Speed Premiums stable.
Has anyone overclocked these on an Athlon 64 3500+ Clawhammer? How well will it perform if i install 4 of them in a DFI Ultra-D? I guess they will default to DDR333 speeds but am i able to run them any higher? How safe is it to run 3.2v-3.4v for a long period of time without added cooling for the RAM?
Just received an unlucky set of AA4T/44D. Neither stick will post at 200FSB 2.5/3/3/9/2T @ 2.9V or 200FSB 2/2/2/5/2T @ 3.2V. Individually they will post at 3.3V but they are not stable. Both sticks will produces thousands of errors in memtest86.
Looks like I need to RMA the two sticks.
Anyone have any suggestions?
BTW: I swapped out my NeoPower 480W with a Fortron Blue Storm 500W because everyone was praising the Fortron so much and disparaging the NeoPower as over-rated. However, I seem to get more memory errors in Memtest86 with the Blue Storm. Do power supplies need to be burned in also? Did I get a bad PS from Newegg? Strange. :confused:
cjoe :mad:
Quote:
Originally Posted by cjoe
What platform are you on? Intel? Amd? Give us some info.
They Neo probably had right near perfect rail values since Antec PSU's are very good at regulating voltage. The Fortron might be undervolting slightly on the 3.3v or 5v rail, whichever is used for Dimm voltage on your board.
Yeah, you maybe right. But I still think that the Twinmos is bad irregardless.
I'm a little disappointed by the Fortron PS also.
Need to burn in new Mushkin. Seems okay @ 220FSB 3.2V 2/2/2/6 1T 1:1.
GSkill LE stable up to 290FSB 2.8V 2.5/3/3/6/9 1T 1:1. OCCT & S&M still burning in.
I'm going to swap / test memory & power supplies.
CJoe
3.2 isn't really enuff to run twinmos at 2 2 2, I've gotten errors running 3.2 volts also whether it was 200 fsb or 260, but at 3.3+ they are flawless. after some burn in, they are really nice, (assuming you didn't get crap in the first place). They should be able to do 240 fsb at least for the pair, if your lucky, they can do 260+ with the pair, 512's now I'm talking about, not 2x256, heh.
What is the most STABLE settings for DFI and 512meg BH-5 ?
What values affect the stability most in bios ?
'You sure about that? :cool:Quote:
Originally Posted by fareastgq
Hit it iwth 3.4-3.5V and see how they act. They can be real squirrely below 3.3V.
@uwackme: My memctrl becomes confused before 260MHz. Memtest spits errors from 0.1MB to xxx.xMB. :cheer:
[IMG]http://forum.lowyat.net/index.php?ac...=post&id=53975[/IMG]
i'm buying this pair of RAM's second hand.
2*512mb Twinmos SP (UTT) PC3200 cl2.5
448W45M40570653 / 448W45M40570676
can you guys please check it out and see if it's good?
@benji: Those are week48... So are mine. Have had 3 pairs of these, all did or allmost did 250 @ DFI nF2 Ultra Infinity. On A64 they do it with less voltage.
To clarify, you may have cherry sticks, but for the most part, within the mean of supply, they cannot run 2 2 2 with only 3.2, or everyone would be reporting that they can run 3.2 volts at 2 2 2 flawlessly. ie: 3dmark/memtest/prime stable. Within the whole range, there have been sticks that simply cannot even boot with stock stable, and there are sticks like yours, lucky you :p:Quote:
Originally Posted by largon
I can run mine 2-2-2-X with 2.6V at 200FSBQuote:
Originally Posted by fareastgq
Quote:
Originally Posted by benji
goto to love that tvtorrents, i know I do :toast:
I noticed my Speed Premiums very dependant on some timings, for example tREF@ 3120 or 4672 they just couldn't do 230 even 3.4V. Changed tREF to 1168 and BAM! 250MHz (2-2-2-5-7-15-2-2-1-1-1168-1) with just 3.2V - no bandwidth loss with this setting.
That gives 3033MB/s in Memtest 1.51.
Glad I didn't frustrate before figured out tREF's impact on OC.
Okay. I'll bump up the voltage and set the TREF to 1168.
I'm also burning in my Mushkin Blue Value (Fry's). 8hrs @ 2/2/2/6 1T 1:1 215FSB 3.3V. Memtest86 in general is error free but test #8 avgs 4 errors per pass still. How much burn in did you guys do before the memory became good?
144 hours in a row @ 3V + 9 hours with current timings and voltage. :flame:
Largon: U the man!!! You're right. After setting TREF=1168, Memtest is much cleaner. I'm going to increase the FSB and let it burn for a couple of days.
What is the current prevailing concensus about Twinmos Speed Premiums vs. Mushkin Value Blue? Which one seems be be better in general?
I have the AA4T / 44D SP and BP815 Mushkins (I haven't opened them up yet).
Thanks...
CJoe
I'd go with those AA4T/44D Speed Premiums over the mushkinsQuote:
Originally Posted by cjoe
okay,i'm buying them later in the evening.
thx guys.
Can you let me in on it?Quote:
Originally Posted by largon
OOps, I got it. 1168 seems to be the ticket for my mushkin blues(green) too
Or does it?
musk
What were your settings at 3V?Quote:
Originally Posted by largon
Were you able to run 2,2,2,6 with > 3.3V immediately after purchase?
I can post at 2,2,2,6 with > 3.6V but memtest #8 registers thousands of errors.
My Twinmos SP are stable @ 240FSB 2,3,6,3,7,15,2,2,1,1,1168,1,E 1T 1:1 3.4V. Does this seem normal after 10hrs of burning? Will they continue to improve after more burning in?
I would be delighted if I could achieve 260FSB 2,2,2,6 @3.4V.
Thanks..
Cjoe
Bought them (6 sticks, sold 4) few months ago. First month they ran 240-250 2-2-2-11 on DFI nF2 Ultra Infinity with 3.3V. No fan. :cord:
3.3V was top vDimm on that board. Board did actually die just few days before I was supposed to solder the vDimm mod. Had already put it together separately.
Mem fets were just vaporised overnight.
144hour burn:
If I remember correct it was 225 2-2-2-5 @ 3V w/ Memtest86+ 1.51 + 92mm Zalman fan keeping the temps down. Left Memtest86+ running through #1-"8 for 6 days, not a single error.
Guess nobody is able to tell will your sticks improve... who knows, maybe they will, maybe they won't.Quote:
Originally Posted by cjoe
Just make sure they don't get fried.
tref 1168 could be working also for WINBOND BH-5 1A4T ??
I have burned my BH-5 stick for 3 weeks.
It improves alot first 3-4 nights after that you gain 2-3 mhz on week.
Last week nothing has happend.
Burning ~10 hours a night.
*edit* Tref 1168 was real good! (DFI NF3 & 1A4T 1X512)
Could do 265fsb 1.5-2-2-4 Superpi 8meg.
Been messing with that all day and now it did it right away.
LARGON RULES :toast:
@Ubermann: :toast:
yep tref 1168 helped a little :)
memtest worked longer :p
So my TwinMOS 2x512mb pc3200 comes in today, which were like $110 on newegg. How can I tell what kind of chips these are? Also, is it possible they are BH5?
@DamienKC
BH-UTT should not be on vanilla sticks.
CH-UTT: (99.999% propability)
P/N: ends in AA4T
If P/N ends in something else they'r... :yawn2:
i hear UTT is good anyways, right? I can put 3.4v max into these, so I'm hoping for 240ish on 2-2-2...think it's possible?
Sure, this thread wouldn't even exist if UTT couldn't do something like that...
About the Speed Premium sticks:Quote:
Originally Posted by DamienKC
http://www.twinmos.com/dram/dram_p_sp_ddr400.htm CH5 PC3200
http://www.twinmos.com/dram/dram_p_sp_ddr433.htm BH5 PC3500
So no 3200 for me ;)
so bh-5 is better than ch-5(UTT) ??
Greetings! I just read this thread because of great interest in BH-5 memory. I used Mushkin Lvl.II with "real" BH-5 chips until today. I ordered two 512MB TwinMos SpeedPremium DDR433 stones, rated at 2.5-3-3-8. I'm using an Asus P4C800, P4XE 3.2GHz, 3.3VDIMM.
I got 1A4T. The memory chips are labeled "BH-5".
http://www.xin.at/thrawn/pics/hw/twi...mium_433-1.jpg
http://www.xin.at/thrawn/pics/hw/twi...mium_433-2.jpg
http://www.xin.at/thrawn/pics/hw/twi...mium_433-3.jpg
I got them to run at DDR450 2-2-2-5 stable, tried no other voltages. I'm stuck at that speed anyway because of my SATA RAID-0 Array no being very clock-resistant. (SATA's clock is not decoupled from FSB like PCI).
Frankly, i didn't buy the memory for overclocking, i just wanted to have 1GB of DDR400, that could do 2-2-2-5. And that it does, and it seems to be a real bang for the buck.. =)
Hello. New to posting here but have been reading these threads for awhile. I have a question and a comment. First, where are the TwinMOS Speedpremium DDR433's being sold? Does Fry's have 'em cause I have had great difficulty locating them online. Newegg looks to be the only place for the SP DDR400's and they do not sell matched pairs of them.
Second, the comment. I've wondwered what a "lifetime warranty" meant in the memory business. Here is a direct quote from TwinMOS' website, http://www.twinmos.com/dram_warranty.htm
Lifetime Warranty
This warranty applies to all “TwinMOS” labeled modules products (EXCEPT Rambus, Elixir, and any original modules) based on the condition that the product(s) is still manufactured. Any product that has reached the end of its product lifetime (End-of-Life) or discontinued shall not be covered this warranty. The end of a product line will be announced by TwinMOS on its website six (6) months prior to the discontinuation of any product(s).
So, looks like it is a product cycle lifetime and not the product. Not that many here use a piece of equipment for multiple years, (more like weeks or months), but the information was news to me.
Hmm, interesting.. Didn't know that either.
I got my SpeedPremium 433's from cheeep.de, which is a german shop (I'm from austria). Cheeep still has those modules in stock, and so does oc-wear.de. Both Shops sell it as "BH-5 memory".
The new BH and CH ics are both sold as UnTesTed. Hence the name UTT. The "new" bh-5 can usually achieve the low latencies at less volts. Still over 3 but usually don't require the 3.6v the CH UTT usually do. I'd say the UTT in general is a steal right now, especially for $110.Quote:
Originally Posted by Th3__W00T_GuY__
Sounds like the SP DDR433 may not be sold in the US. Can anyone dispute this? I'm theorising that maybe the remaining "old stock" of BH-5's is only distributed outside of the US. Guess I'll eventually pickup 2 of the SP DDR400 parts at newegg. Anyone have good results with these sticks in MSI's K8N Neo2 Platinum (NForce3 Ultra)? I just cannot afford to go PCI-e just yet. :(
Not exactly. The current UTT-BH is built with the same 0.175um die screen as the BH-5 of yesteryear so there shouldn't be much of any difference between the two. And sure enough, I have a 512MB stick of UTT-BH and a 512MB stick of "old" BH-5, and they both can hit 200MHz 2-2-2 @ 2.85v.Quote:
Originally Posted by cardnut99668
And much of the current UTT-CH can easily hit 2-2-2 with 3.2 volts, you certainly don't need 3.6v.
I'm one of the Unfortunate ones. Mine only will do 2-2-3 with 3.3 and its still unstable in memtest. I'm still getting 1-2 errors per pass. I'm hoping that with burn-in it will improve. If I change the settings to 2-2-2-x, it will produce thousands of errors.Quote:
Originally Posted by xenolith
However, 230-240Mhz @ 2-2-3 3.3V for $110 is still a helluva value.
Cjoe
http://forum.lowyat.net/index.php?ac...=post&id=54225
Does this look good to you??
Tested to run at
2-2-2-5-1T
3.3v ~ 245MHz (DDR 490)
3.4v ~ 255MHz (DDR 510)
3.5v ~ 260MHz (DDR 520)
Looks good benj, is that Speed Premium BH5?
Quote:
Originally Posted by cjoe
Did you try to do what largon did and adjust your Tref value to 1168 or do you not have that option in your BIOS? My MSI K8N Neo2 Plat has little in the way of mem timings in BIOS so I do not think I can adjust it unless one of the esteemed members here have a tool to at least adjust the value in windows. I'm trying to remember if A64 Tweaker had it. I thouoght so but the values were not so high like 1168. I think that value is in microseconds and the A64 tweaker values are in clock cycles. That statement begs the question, "How many microseconds in a clock cycle?"
@jcollett69: tREF 1168 in A64 Tweaker is "166MHz 7.8µs".
Yeah, I used 1168 TREF. It helped alot but it didn't solve my problems. I can't run 2-2-2 clean with any settings. I spent a lot of time on my DFI NF4 playing the settings. Largon said he spent 144hrs burning in his ram. So, I decided to burn in the memory at 2-2-3 @ 200Mhz 3.3V to see if I can improve the performance. I put the memory into my DFI NF2 LP B and letting it just burn in with memtest.Quote:
Originally Posted by jcollett69
Most people seem to hit 2-2-2-x pretty easily and scale to 240FSB with 3.4V in DC from all the posts. I probably just got an unlucky set of memory.
cjoe
Is there a max voltage limit for this memory? AA4T?
I'm "stuck" at 256 mhz (2x512, DUAL, PAT disabled). And If I try voltage like 3.9 there are still errors.
It's on ABIT AS8 with intel 3.2
IMO I think those sticks will top out around 3.6v Any more than that and the increase will be very minimal or actual cause errors.
Thanks largon. Saw that in a screenshot of yours featuring your A64 Tweaker settings. So I assume setting it to 166MHz 1.95 microsecond causes issues?Quote:
Originally Posted by largon
Quote:
Originally Posted by b|lly
Seems like you have it running very fast for an Intel setup.
I think i've hit the wall here. I'll try to burn it a little bit and maybe get a few more mhz...
btw, hipro5 has been running it at 260 mhz ...
I myself have some TwinMOS ValueRAM AA4T Winbond CH-UTT, whatever you guys wanna call them, and after some minor negotiation and swapping DIMM slots about 6 times, I was able to successfuly run them Prime95/memtest stable at 260Mhz 1.5-2-2-7-1T with 3.4vdimm. 260MHz is good, damn good for what I paid for the RAM ($108 :banana: ), but, being an overclocker, I'm never satisfied, so I decided to give thise whole "burning them in" thing a go. So I set it to 265FSB, which gave about ~80 errors per pass on memtest #5 and let it loop for the better part of the afternoon and all night. To my dismay, I came back the next morning to find the number of errors had increase to nearly 800 per pass :slap: I'd say my cooling over the DIMMs is sufficient, hard to think of anything better than 125CFM off of my XP-120 and the ICs are barely lukewarm to the touch so I'm lost :confused: I'm pretty determined to :horse: cause 260 just isn't good enough :D
EDIT: Oh, and I'm new here :D But I've been following this thread from pretty much day one (bought the RAM last week though).