nice...
300, 2.5-3-3 is the holy grail for GIG kits of tccd.
C
Printable View
nice...
300, 2.5-3-3 is the holy grail for GIG kits of tccd.
C
UCCC is a 64x8 density ICQuote:
Originally Posted by TEDY
So:
UCCC is 64MB IC. With 8 ICs on the DIMM you get 512MB / stick. (8 x 64 = 512MB)
TCCC, TCC5 and TCCD are 32MB ICs. (8 ICs x 32MB = 256MB)
We can't give an ETA, due to the fact we are still testing chips and binning procedures. We hope to make an announcement soon. PC4000 2gb with tighter then 3-4-4 timings ain't easy! We are just as anxious as you are and will do our best. :)Quote:
Originally Posted by brandinb
http://www.bleedinedge.com/forum/showthread.php?t=9833
This thread has info for you. Posts 9, 10 and 11 are for TCCX memory. The 310 BIOS is fine unless you have an X2 CPU.
heh, just talk to micron. my crucial value that has micron 5b D chips is currently memtesting at 265mhz, 3-4-3.. i know some real binning will fix the problem, and allow for tighter timings.. i'm sure you remember how popular the EB line used to be... though, speaking of that, i remember it coming in 2x1gb form. what chips did it use back then?Quote:
Originally Posted by andyOCZ
Quote:
Originally Posted by ozzimark
Never came in 2x1Gb afaik. The EB line used Micron -5B C chips, which were supposedly better than the -5B D chips, hence why OCZ stopped producing the EB line once the C rev ran out.
blah i hate micron chips i have had so much trouble with my 512x2 kit of ballistix pc4000 i hate it. and i cannot get it stable with dividers at all and in the end i want a better chip. it requires really slack timings besides the usuall cas trp trcd timings.
ocz i belive says thier 2x1gb kit isnt using any micron chip so im happy for that. plus that crucial w/micron chips is damn expensive and im not paying some stupid 400 bux on 2x1gb of ram when i can get some overclockable 3200 sets for 230-250ish. im just hoping ocz can keep the price down on then new set of 2048mb pc4000 kit
________
Michigan Medical Marijuana Dispensary
http://www.ocztechnology.com/product..._bandwidth-eolQuote:
Originally Posted by EnJoY
it just makes me sooooo curious :pQuote:
Available in 512MB (2x256MB), 1GB (2x512MB) and 2GB (2x1GB) Dual Channel Optimized Kits
also.. the C definitly didn't run out, my 128mb stick that i got back in march or april is 5b C, but i digress :D
brandinb.. while i can understand your sentiments, i disagree with them. i have had nothing but awesome luck with micron chips, owning a 128mb stick of 5b C that does 265mb at 2.5-2-2, 2x512mb ballistix that'll do ~255mhz at 2.5-2-2 and now this 1gb stick that can break 250mhz at 3-3-3
well ozzimark i notice you are having problems keeping prime stable and that is what im talking about i just cannot get them stable! they will run "ok" but crash every now and then and prime is a no go and same with super pi thats why i did not like my crucial/microns
________
M104 ENGINE
Hello, the truth is that the Micron C die chips had a good week or 2 that did 2.5-2-2 timings, then "progressed to 2.5-3-2 timings, and then yields for those timings became terrible. We refused to raise timings for the EB and just decided to drop the product.Quote:
Originally Posted by ozzimark
I also have a 2 x 512mb set of the earliest EB PC3200 that does 2.5-2-2 at 260mhz + , it's just too bad the yields became poor and inconsistent.
that's good stuff to know, and explains why EB was phased out, thanks andy :toast:Quote:
Originally Posted by andyOCZ
(but then i guess it means that my luck was that much more astounding, or else micron has since fixed yeild issues)
heh.. i know what you mean there... i managed to bring it within 1mhz of my memtest stable speed.. it just takes a bit of work. imo, that's the best part about ram.. if it was simple to overclock, i wouldn't enjoy it nearly as muchQuote:
Originally Posted by brandinb
The 2 gig kits of EB were hynix based ;), we hardly sold any and the price of the IC's kinda killed it anyway, maybe we should go back and revisit it now ;)
Rest assured we are trying to bring a 4000 rated part to the masses with good timings.
I'm fortunate to have two sets of TCCD that run SuperPI 32M at 300, 2.5-3-3-7 and I can tell you, there are many different definitions of that. Things get very squirrelly here, same as on P4 boards.
I've seen sticks that do it Memtest clean (at 2-2-1-2 secondary timings at that), but are not 3D stable till 290 or so. Sandra memory test passes 50% of the time. No way SuperPI 1M, not even at 290.
I've seen sticks that do 300 Memtest clean and will pass Sandra Memory and Cache-and-Mem, but passing SuperPI 1M is an adventure (it will pass if you try enough times). Forget 32M. But 3D stable.
Sticks that will run 300, 2.5-3-3-7 stable 99% of the time in SuperPI 1M, and can also clear 32M are very rare (unless you work for a memory manufacturer and can pick the cream of the crop). The one set of FFs I've tried will do it, but you're paying through the nose for it. I wouldn't be surprised if many sets fall a tad short as the sticks are only guaranteed at 275 here (nor at 32M).
My PQI Turbo DC kit (rated PC3200, 2-2-2-5 heh), has run 300, 2.5-3-3-7 on a 3000+, 3200+, 3500+ (all Winnies), and a 3700+ SD. I just submitted a 32M run on Bacchus's thread. Maybe it's the board? Not bad I guess, for a used DFI Ultra-D I got for $60 U.S. in Japan :shrug:. This ram BTW, also did 290, 2.5-3-3-6 on an AI7 (865P board) @ 3.05 VDIMM :eek:. I quit trying for more as it was Memtest clean and 2D stable, but not 3D. I did not try 32M.
I tried at FX-55 once. The mem controller wasn't any better than the CPUs above.
All I can say is each set of TCCD has it's own peculiar characteristics and idiosyncracies; getting Memtest clean here is the first step, but in no way guarantees functionality. I still say it's all in the sticks. I will be installing my backup DFI SLI board soon to test this theory.
and even more useful info! thanks for your knowledge guys :woot:
i'm also going to be keeping a close eye on how these 1gb sticks do, even if i suspect they will be out of my normal price range for parts
can you mention the chip used in the new pc4000 2x1gb part BIGTOE?? im hoping for hynix chips as ive had a lot of fun overclockin hynix in 512x2 config.
________
Paxil lawsuit information
Can't say what we're testing bud, hope you understand ;)Quote:
Originally Posted by brandinb
FX CPU's have the best memory controllers and usually pretty sweet cores, if they did not why would we all buy them, cause they are unlocked up the way?...i think not ;)
I have a brand new venice 3000+, nice CPU...2.8GHZ stable, highest 300+ though is ONLY with the 8 multi and not the 9.
Cores clock well, memory controllers do not.
which is very sad.. but oh well. perhaps amd will make strides with M2 + ddr2Quote:
Originally Posted by bigtoe
not quite. as you can see from pdf you provided, 'u' instead of 't' just means that ic packag is lead-free.Quote:
Originally Posted by largon
to high5:
Examine the chart further...
http://img331.imageshack.us/img331/5...ungdram9zf.jpg
from here you can clearly see that 'u' or 't' (among others) tell only packaging type and not density. from picture you can also see that #4 determines density.
to high5:
YES.
U means lead-free.
But what does xxxxxxxxxx-UCCC mean? Let's examine the next page, shall we:
http://www.helsinki.fi/~llounent/UCCCs.png
you're either reading the sheet completely wrong or you're lacking some basic knowledge.
-uccc rating has nothing to do with the first part of the part number, if you're looking for 64Mx8 density ic, p/n of desired ic would be k4h5108xxx-xxxx. it doesn't necessarily needs to be uccc and not all uccc are 64x8 (which you can see easily from pic you posted), also there are 32Mx8 ic's rated uccc, as you can see in the attachment.
i'm just saying your statement that uccc is 64Mx8 density ic is completely wrong and is based on wrong facts.
Quote:
Originally Posted by Clevor
Also, if you take all the same hardware otherwise, and the same exact pair of ram and place it in 5 different DFI MF4's. you will get 5 different top HTT #'s, and some quite a bit apart. It has been done, and that is exactly what happened.
What does it matter in this particular case, does UCCC exist in other densities than 64x8?
The question was:
Can a 1GB stick be built using Samsung UCCC? If so, should the chip be 64x8?
According to the quoted data sheet, the answer is:
Yes.
Don't you agree?
Yes, I realise my statement "UCCC is a 64x8 density IC" is unaccurate. But it is also true. There is a UCC chip whose density is 64x8. Let's not argue about things that are irrelevant.
This isn't "The XtremeGrammarSystems Forum". ;)