UPDATED INFO, EFFECTIVE AS OF JUNE 9 2008:
We will no longer be using Micron ICs in our PC8500 parts due to lack of availablility from Micron.
Additional DDR2 IC information can be found here:
http://www.houseofhelp.com/forums/sh...ad.php?t=57334
Printable View
UPDATED INFO, EFFECTIVE AS OF JUNE 9 2008:
We will no longer be using Micron ICs in our PC8500 parts due to lack of availablility from Micron.
Additional DDR2 IC information can be found here:
http://www.houseofhelp.com/forums/sh...ad.php?t=57334
Hey guys!
This is the part of our announcement where I will inject some of the data from our lab. As already mentioned, the new 8500C5D is built with Qimonda DRAM chips. How to tell them from its Micron-based predecessor, you'd ask? Look at the label on a module pictured below, please. Two pieces of information need to be acknowledged, which indicate 8500C5D to be Qimonda-based:
2.1V nominal voltage and ver2.2 revision
http://www.clockmehigh.com/TestingAt..._21V_ver22.jpg
I randomly selected one of the validated production kits and put it through extensive testing, then compiled some good data for easy viewing. For the testing platform, I chose Asus P5K Deluxe, based on Intel P35 chipset.
Test setup:
Asus P5K Deluxe, BIOS 0501
Intel C2D E6850, 3.00GHz, 1333FSB, 4MB L2
Corsair TWIN2X2048-8500C5D 2.1V ver2.2
eVGA 8800 GTS 640MB Superclocked
Corsair 620 HX PSU
Let's start off with BIOS configuration. Below you can see how the motherboard has been configured for the purpose of our lab testing. Those should be considered a guideline only, as some of the settings (e.g. CPU multiplier or FSB Frequency) might differ depending on a CPU end-user chooses to use.
BIOS --> Advanced -->JumperFree Configuration:
Ai Overclocking - [Manual]
CPU Ratio Control - [Manual]
Ratio CMOS Setting - [9]
FSB Strap to North Bridge - [Auto]
FSB Frequency - [333]
PCIE Frequency - [100]
DRAM Frequency - [DDR2-1067MHz]
DRAM Command Rate - [Auto]
DRAM Timing Control - [Manual]
CAS# Latency - [5 DRAM Clocks]
RAS# to CAS# Delay - [5 DRAM Clocks]
RAS# Precharge - [5 DRAM Clocks]
RAS# Activate to Precharge - [15 DRAM Clocks]
tWR - [Auto]
tRFC - [Auto]
tWTR - [Auto]
tRRD - [Auto]
tRTP - [Auto]
DRAM Static Read Control - [Auto]
Transaction Booster - [Auto]
Clock Over-Charging Mode - [Auto]
CPU Spread Spectrum - [Disabed]
PCIE Spread Spectrum - [Disabled]
CPU Voltage - [1.3500v]
CPU Voltage Ref - [Auto]
CPU Voltage Dumper - [Auto]
CPU PLL Voltage - [1.50v]
DRAM Voltage - [2.10v]
FSB Termination Voltage - [1.30v]
North Bridge Voltage - [1.55v]
North Bridge Voltage Reference - [Auto]
South Bridge Voltage - [Auto]
Default frequency DDR2-1066 5-5-5-15 @ 2.10v
Now, let's take a look at some stability tests performed at default 2.1v voltage. After our standard production test, I tested our setup with 2 as well as 4 DIMM sockets populated, 2GB and 4GB respectively, in OS environment. Stress test of choice is ORTHOS, which is a 3-rd party GUI for well respected Prime95. Additionally, I have performed some 3D testing and gaming at those settings, and system worked perfectly as well.
ORTHOS 21hrs - 2x1GB DDR2-1066 5-5-5 @ 2.10v ----- ORTHOS 24hrs - 4x1GB DDR2-1066 5-5-5 @ 2.10v
http://www.clockmehigh.com/TestingAt..._21hrs_4x3.gif http://www.clockmehigh.com/TestingAt..._24hrs_4x3.gif
Default frequency DDR2-1066 5-5-5-15 @ 1.90v
Important, and interesting at the same time is the fact that the new TWIN2X2048-8500C5D actually requires only 1.90v to operate properly on some motherboards, including eVGA 680i and Asus P5K Deluxe. Keep in mind however 2.1v is still this kit's official specification. Therefore consider ability to perform at 1.9v a "neat feature" that may be exclusive to selected few motherboards.
ORTHOS 48hrs - 2x1GB DDR2-1066 5-5-5 @ 1.90v ----- ORTHOS 96hrs - 4x1GB DDR2-1066 5-5-5 @ 1.90v
http://www.clockmehigh.com/TestingAt..._48hrs_4x3.gif http://www.clockmehigh.com/TestingAt..._96hrs_4x3.gif
----- Overclocking @ CL5 -----
In this paragraph, I have compiled results of my overclocking attempts at default timings 5-5-5-15.
ORTHOS 24hrs - DDR2-1110 5-5-5 @ 2.10v ------- Dual SPi32M - DDR2-1120 5-5-5 @ 2.10v
http://www.clockmehigh.com/TestingAt..._24hrs_4x3.gif http://www.clockmehigh.com/TestingAt...SPi32M_4x3.gif
SPi1M + SPi32M - DDR2-1120 5-5-5 @ 2.10v ------- Sandra XI - DDR2-1120 5-5-5 @ 2.10v
http://www.clockmehigh.com/TestingAt...i1M32M_4x3.gif http://www.clockmehigh.com/TestingAt...MemLat_4x3.gif
PCMark05 Memory - DDR2-1120 5-5-5 @ 2.10v ------- WinRAR - DDR2-1120 5-5-5 @ 2.10v
http://www.clockmehigh.com/TestingAt...05_Mem_4x3.gif http://www.clockmehigh.com/TestingAt...WinRAR_4x3.gif
3DMark2001 - DDR2-1120 5-5-5 @ 2.10v ------- 3DMark03 - DDR2-1120 5-5-5 @ 2.10v
http://www.clockmehigh.com/TestingAt..._3DM01_4x3.gif http://www.clockmehigh.com/TestingAt..._3DM03_4x3.gif
3DMark05 - DDR2-1120 5-5-5 @ 2.10v ------- 3DMark06 - DDR2-1120 5-5-5 @ 2.10v
http://www.clockmehigh.com/TestingAt..._3DM05_4x3.gif http://www.clockmehigh.com/TestingAt..._3DM06_4x3.gif
----- Overclocking @ CL4 -----
I've also decided to take a look at these modules' overclocking potential when timings are tightened to 4-4-4-12. As shown, at around DDR2-900 system remained fully stable.
ORTHOS 24hrs - DDR2-910 4-4-4 @ 2.10v ------- Dual SPi32M - DDR2-910 4-4-4 @ 2.10v
http://www.clockmehigh.com/TestingAt..._20hrs_4x3.gif http://www.clockmehigh.com/TestingAt...SPi32M_4x3.gif
Thank you for your attention!
Thanks for the heads up :)
I just picked up some C5D ver 1.2s, and am curious about the new revisions. I am actually testing running the C5Ds with a pair of 8500C%s to see how they overclock in 4 dimms. The new version seems very interesting.
Has anyone purchased these in the market place yet? Am very interested in these due to the ability to finally run 4GB in the 680i chipset at rated speeds.
can it run at 1T CR ??
:confused: CR?Quote:
Originally Posted by idiot1234556
Command Rate, he asks if you can run it at 1T.
AH...LOL....do I feel dumb, I realized the 1T of course but the "CR" after it threw me off. If the MOBO in question supports 1T, then yes this memory will do it.Quote:
Originally Posted by .devastatioN
^^^^ ROFL quick edit the post :D:D:D j/k
Nah, it's cool. :rofl:Quote:
Originally Posted by dinos22
It's a good reminder to pay better attention and DRINK MORE COFFEE :up:
so whichs one are better the v1.2 or the v2.2???
That depends on what your idea of "better" is. See post #1 and #2.
which one should i buy if i have both of them in front of me? which one will perform better, and be more stable?
Both are equally stable at the rating timings and speed. But, "better" is still relative here. If you mean running 4 up or if you want to run at a lower voltage, then the Quimonda has better characteristics. If you want pure OCing, the Micron version responds to voltage a bit better and may be able to hit a higher speed although nothing over 1066 is guaranteed.
1.2v is the Micron version?
Yes, see this list for info: http://www.houseofhelp.com/forums/sh...ad.php?t=57334
Any 1.* i s Micron.
Any 2.* is Quimonda.
Some questions> I have a problem with booting on my mobo since it tends to revert to default voltage *1.8 or such, on my "old" Micron domis 8500 c5d. As I understand Quimonda runs at lower voltage, so maybe they could boot easier than the Microns`?, Or do I still have to keep a booting crap pair of ram around?
Are there any known problems with "mixing" these two types of ICs in the same system 2*1024 8500 c5d ver 1.x and 2*1024 8500 c5d ver 2.2? Compatibility wise as well as performace wise, without factoring in the generally known overclocking "drawbacks" in regards to 4 gig system memory setups.
The newer revision have ICs similar to what you would call "crap pair of ram" in terms of voltage requirements so even if the BIOS kicked back to 1.8v those sticks would still boot at the timings BIOS defaults back on.Quote:
Originally Posted by ZilverPhish
And you should have no problems mixing the 2 different revisions of the 8500C5D the only differences you will see between the 2 revisions now is their "theoretical" overclockability, they still might overclock well, but Micron is tried and true when it comes to overclocking and how well the IC's respond to voltage increases, and how well they handle lower timings and higher speeds and whatnot :)
Quote:
Originally Posted by Yellowbeard
ROFL! What is next to get downgraded from Micron to Promos or Quimonda? 6400C3? 9136C5D? 8888C4D?
REST IN PEACE(started as Micron, Died as garbage)=
CM2X512A-4300C3
CM2X512A-4300C3PRO
CM2X256A-5400C4
CM2X512-5400C4
CM2X512A-6400
CM2X1024-6400
CM2X2048-6400C5DHX
CM2X1024-6400PRO
CM2XP1024-6400C4
CM2X1024-6400C4/C4DHX
CM2X1024-6400C4D/C4DHX
CM2X1024-6400C4PRO
CM2X1024-8500C5
CM2X1024-8500C5D
I remember when 8500C5 was expensive, top of the line stuff... and now it is garbage :(
Your product numbers and versions only serve to confuse people and screw them should they ever have to RMA.
For how long? A week?Quote:
EDIT: Oh, just so you know, the 8888C4, 6400C3, 9136C5, and 10000C5 are all still Micron-based.
If you have an issue with any of your Corsair memory meeting its specification, please let us know and we'll promptly get it replaced.
How high do the new 70nm ICs oc compared to the d9s (I see you guys did a bit of ocing, but no voltage increases etc)? I would imagine that if they run cooler and use less voltage it would be smart to put them on the higher grade kits too?
8500C5D was top of the line till 8888C4DF,9136C5D,6400C3DF and 10000C5D came and smacked it in the face :)Quote:
Originally Posted by fhpchris
However I was told by Corsair that the 6400C3DF was discontinued recently? So seems just the 3 top end DDR2 kits.
I've been testing these modules for a little over a week on the P5K deluxe and although the results are not as impressive as bachus_anonym I am impressed.
They do what they are rated at and up to about 1088 stable @ stock voltages and latency.
anything above that so far hasn't been achievable and seem unresponsive to voltage increases.
the difference on my part is likely due to the fact that I do not have a fresh clean install and Q6600 G0 instead of the E6850.
http://i201.photobucket.com/albums/a...7_18-09-22.png
http://i201.photobucket.com/albums/a...7_22-32-43.png
http://i201.photobucket.com/albums/a...7_22-47-19.png
How far have you pushed these in your labs? :D
2.2V 1200 4-4-4-4 PL5 1M stable?
See post number 2. Bachus has done most of the testing with this memory.Quote:
Originally Posted by KTE
I've read it all, he didn't really push it at all, it was all stock volts or below. I mean he must have added volts to at least one pair and checked where it goes as part of personal testing?
I have a set of the new 2.1 version. They don't respond to voltage very well, or really overclock much, but they will do between 1050-1100 with only 1.9v. I only took them to 2.35v, but I couldn't get much more than 1100 @ 5-5-5-15. I have been running them with a Q6600 8x450 5:6 divider and only 1.9v. I have had many D9 sticks die on me with 2.2v or less,(including my ver1.1 of these) so I am pretty happy with what I have now.
Yeah, they don't have the same scaling with voltage characteristics as Microns but, they do have their own strengths. And, anyone can easily use them with a CPU at 1:1 up to a 550 FSB. Not too many people going way past 550.Quote:
Originally Posted by [XC] 4X4N
And thats the reason I would buy them again :up:Quote:
Originally Posted by Yellowbeard
Quote:
Originally Posted by ZilverPhish
I'm not sure they are compatible. Not YetQuote:
Originally Posted by ChaosMinionX
I have both version mixed for the first time.
TEST BED
EVGA 680i A1 mobo P31 bios|Q6600|8800 GTS 640 sli|OCZ700GXS
CM2X1024-8500C5D 2-2GB kits v1.2 & v2.2
no OS yet booting to iso UBCD v4.1.1 http://www.ultimatebootcd.com/index.html
using Memtest86+ & Mersenne Prime Test to establish stability.
I was able to run both pairs individually dual channel mode 2GB config
@ 400x9 Link/sync 1600 A. FSB, SLI-Ready Mem.=Disabled
all spread spectrum & CPU features Disabled
Auto volts except vdim man set @ 2.15v
Mem. time= man set Expert 4-4-4-15-1T all else auto
Both pairs passed 6 hrs. prime stable & 6 hrs zero errors memtset
Now I did not expect that they would run together @ 1T in a 4GB config
but I did hope for 2T 4GB config at the low latencies.
NO JOY
Now I'm not saying this is a flaw I'm hoping I've made a minnor miss calulation
and someone can help me find the love.
Qusetions:
1) Am I wrong to go the low latency route or would you recommend higher FSB for best performance?
2) Is this an unrealistic expectation to pair these two revisions in the same bed @ low latencies or otherwise?
3) What volt would best suit this pairing?
4) What man sets to use? or all auto see where it goes....
mmmmmm I'm gonna try that but I digress
5) I've noticed the PCB HS seems to be somewhat different in that
it appear to be a different color or possibly grade. Could this account for the better thermal performance.
Please Advise!
1. I would go for the highest FSB you can get. Don't let latencies hold your CPU back. CPU cycles are good.Quote:
Originally Posted by owcraftsman
2-3. No, they should run fine together. 2.2v or 2.25v is fine when mixing them. See #1 for latency.
4. Manually set 5,5,5,15 and 2T, all else auto or defaults.
5. No, there has been no change in the PCBs that would affect the thermal performance.
Qimonda versus Promos chips
C4DHX XMS640X v2.x...PROMOS........64MX8...90nm
C4DHX XMS640X v4.x...QIMONDA......64Mx8...Rev.C
Which are preferable, the Promos or the Qimondas IC's?
Purpose: for 2 x overclocked Q6600 systems with P5K-E and GA-P35-DS3P M/B's.
Thanks
Both of those are going to perform about the same and they'll easily run spec. They are not going to respond to voltage increases and OC quite like a Micron D9 set will.
That's understood and there was never a question about that issue, ie. a comparison with the D9's liking for voltage.Quote:
Originally Posted by Yellowbeard
The question was if you had to pick between DHX's with Qimonda or Promos IC's, which kit would you choose based on whatever anecdotal feedback has come in on these different revisions of the same product line.
I honestly have not compared the 2 so I can't make an educated guess based on any anecdotal feedback as I really have none. We don't get much hardcore enthusiast feedback on our non-Micron memory these days.
If I was forced to chose, I would take the Promos. IIRC, we have been using the Promos ICs longer than we have the Quimonda. The failure rates are VERY low so I know it is a very reliable IC. The OCing is limited somewhat due to the voltage regulation on the IC. But, I have no other evidence or experience to offer.
I've had the C4 PC2-6400 v2.1 XMS2 (rated 2.1V) which are ProMos IC's. They dislike and do not scale with voltage above 2.35V but they run very cool, far cooler than D9's so dying at stock (overheating) is very unlikley unlike with Crucial RAMs.
Once you hit 2.6V on them or a little regular 2.4V the IC's will be dead and start throwing thousands of errors. This is regardless of temperature, doesn't matter if you have them at 1℃ all the way. However, this is a 2.35V run on them: http://valid.x86-secret.com/show_oc.php?id=268251
Because of the difference in voltage requirements between the two versions... can they be mixed?
I currently have a 1.* and I'm adding a pair... can they be mixed with the voltage differences?
If so... what to set the voltage to? 2.2 or 2.1?
thanks
hagatha
Set them to 2.2v as the revision 1.* is tested at 2.2v. The extra .1v will not damage the newer RAM. If you want, you can always test at 2.1v to see if it is stable there.
I had a TwinX kit of those rev. 2.1 with ProMOS chips and i couldn' get them higher than 1000MHz or perhaps 1066MHz at some point, but they did rune really nice on speeds below that with very tight latencies. In my signature there are images of what i achieved with them, most impressive was SPi stability at 960MHz 4-4-3-5-2T.:cool:Quote:
Originally Posted by KTE
Not so bad chips if you haven't got D9's.
Hey, i just bought 4 gigs of this dominator 8500 ram at the same time from newegg... and i got 2 different versions... version 3.1 and version 4.1 rated at 2.2v on thier label... but in cpu-z it says the 3.1 runs at 2.1v and the 4.1 runs at 2.2 is it wise to run these 2 different RAM together? will i have any problems from mixing them?
You shouldn't but you may need to set the RAM timings and voltage manually. If they are not stable at 2.1v, 2.2v is safe for both versions.Quote:
Originally Posted by GunterFalstaff
wait you bought it as a kit? or do you mean you bough 2 4gb kits? you imo should not have recieved 2 different version in a matched kit..
no no... i bought 2 2gig kits (1gigx2)
If you have the option to exchange them, a 2 x 2gb kit is a much better solution. But it is up to you.Quote:
Originally Posted by GunterFalstaff
Ver 4.2 realy :banana::banana::banana::banana: ?
I just received Ver 4.2 also from Newegg.
(1) What chips do they use?
(2) What is the technology (e.g., 65 nm)?
(3) Will they boot at 1.8 volts?
I don't have a motherboard yet, so what timings can I expect at 1.8 volts?
they are single sided powerchips, and if they are like the double sided kits they should do 1000 5-5-5-15 with 1.8Quote:
XMS850X v4.X...PSC..............128Mx8 (2.1 Volts)
Thanks. I missed that. That works for me.
what should tRC be set at?
in CpuZ is says EPP #1 setting should be at 22 but when its on auto in bios is shows 31
http://farm3.static.flickr.com/2169/...a422a0e5_o.jpg
i have them set at 2.1V in bios, will they do tighter timings? say 4-4-4-15-2T?
also any tips for OC these with the phenom 9850???
The only way to know about the tighter timings is going to be to experiement with them yourself. I am seeing VERY few people using Phenoms these days so there is little info available. The only attempt I have made was with a B2 stepping 9500 and an ABIT AX-78 (ATI chipset).
You are already far past what I was able to do so I'm afraid you are on your own unless you find some other Phenom users. Is your CPU a B3 stepping?
thanks...
its a phenom 9850 Black Edition and it is the B3 stepping
right now i have it at 2.9ghz (200x14.5)
it ran OCCT 2.0.0a stable for the hour test.
i had set the memory timings at 5-5-5-15-25-2T, at the ddr2 1066 speed, ganged.
seems to be doing fine.
how far can i set the voltage? probaly nothing higher than 2.2V im assuming.
my phenom doesnt like fsb changes so ill be sticking with the multiplier
I just bought the 2GB (2 x 1gb) of these modules, ver 2.2 and they are giving be BSODs on my P5B Deluxe. Any specific settings i need to make? They are running at ddr2 650mhz but if i change the strap on the board to 800+Mhz, it doesnt boot.
If you are at 650mhz, then you are WAY past the tested settings and are on your own as far as OCing. Make the modules are well cooled. Also, that speed on a 965 MOBO, you are most likely going to have to run the 1066 strap (does it even have a faster strap???) as the chipset latencies will be too tight at the 800 strap.Quote:
Originally Posted by Sanad
What voltage are you using?
650mhz as in 2 x 225Mhz. My FSB is 225mhz! That isnt way past the tested settings as the ram is tested to run at 1066Mhz! Im almost half that! It's running underclocked. I tried 1066 strap at 2.1v but it wont boot. Now its at 1.9v and 225mhz with 5-5-5-15. Latencies wont tighten either.
Actually, they are tested to run at 533mhz, not 1066. What memory divider are you running now?Quote:
Originally Posted by Sanad
Also, you need to tune your memory controller voltages and raise them a bit to run that speed.
Keep in mind you will need run the 1:2 memory divider to get anywhere close to 1066 on the memory if your FSB is low. What CPU you are you using?
Im using a quad core Q6600. Its running at FSB 266 x 9. My voltages are as follows:
CPU-1.28
Ram: 1.9v
NB: 1.55v
SB: 1.5v
UPDATED INFO, EFFECTIVE AS OF JUNE 9 2008:
We will no longer be using Micron ICs in our PC8500 parts due to lack of availablility from Micron.
Additional DDR2 IC information can be found here:
http://www.houseofhelp.com/forums/sh...ad.php?t=57334