OCZ VX, Best Memory made since BH5!!

[OPPAINTER]

I just started playing with my new OCZ VX memory, looks great so far.

This stuff is the memory to replace BH-5,, finnally:D
Made for Overclockers by Overclockers!!

Great job OCZ!! :toast:

OPP
http://xtremesystems.org/forums/attachment.php?attachmentid=20964&stc=1

[Orker]

Very nice oc.. should have posted this in the official vx results thread though

what voltage was that ? and is it memtest #5 stable?

[OPPAINTER]

Very nice oc.. should have posted this in the official vx results thread though



I'm makeing a stand alone statement here:D

OPP

[bachus_anonym]

glad you like it, OPP ;)

i'll be testing one of those kits very soon, maybe this weekend...
hopefully it's really capable and extends BH-5 as i haven't seen ABSOLUTELY stable e.g. 260x10 6-2-2-2.0 1T from any 2x512MB of BH-5 in Dual Channel and AMD platform (NF2 or NF3).

[Orker]

I'm makeing a stand alone statement here:D

OPP

Its very good ram isnt it :)

Best ram ive ever owned

btw can you please answer my questions from before

[ArcTan]

"Best Memory made since BH5"

so still not as good as BH5 :D

[MaxxxRacer]

arctan.. the antagonist... lol... I would say that opp endorsing it would increase sales, but the stuff is flying off the shelves so fast i bet ocz's heads are spinning..

[blinky]

i havent seen it scale high like bh5 does tho :(

if it gets 270 2-2-2-5 3d benchable ill jump for it

[Alexandrus]

OPP, what voltage are you using for those VX ?

[ojdr2001]

very nice for 1 GB... :)

What board and voltage please?

[halcyon]

Second that. Board and voltage :)

[GomerPyle]

Would like to test those OCZ :slobber:

Great results Opp.

[_Eduard_]

Didn't OCZ buy the rights of winbond to make BH-4? That's what these chips are aren't they?

[DaWaN]

Didn't OCZ buy the rights of winbond to make BH-4? That's what these chips are aren't they?

OCZ is not telling what these chips are, but i don't think these are BH-4, because they aren't as good as BH-5

[musk]

I'm makeing a stand alone statement here:D

OPP

Respect!

[reject]

i still havent seen them in aust :(

[LilGator]

why is everyone saying these aren't as good as BH-5 ?

It seems like every pair of these are smacking 250 MHz 2-2-2, with a 2x512MB kit no prob.

BH-5 never consistently did that. Espescially with 2x512MB.

[alexio]

Well, I think the 2*512mb pairs of OCZ VX might be better overclockable then the BH-5 pairs of 2*512mb

The difference between OCZ VX and BH-5 is more in the fact that some 256mb BH-5 sticks have done 300mhz 2-2-2-5 and there is no way OCZ VX can beat that.
Maybe the OCZ will clock a bit higher with the new ATI chipset, how knows what will happen?

I'm thinking of buying a 2*512mb pair of OCZ VX for my new system but I'm not really sure if that would help allot sind I allready have some Corsair XMS PRO PC4000 that can do 260mhz at 2.5-3-3-7

Well if my new system wil be an FX-53 or FX-55 I think I might just by a pair of OCZ VX but if my new system is going to be a Pentium-M Dothan I'll stick with my XMS.

[OPPAINTER]

The problem these days is if you don't use 1gig and only have 512Mb like BH-5 in most cases, you lose out on performance. The 1gig gives you 2-way interleave, 2 gigs will give you 4-way interleave. This is on S939 motherboards anyway.
So haveing 1 gig or higher is in if you can use aggressive timings,, so now you have VX:D

OPP

[OPPAINTER]

i want some VX.
Any Overclocker should:D
I only played with the stuff for an hour, I think we'll see guys running 3D into the 270s on this mem. I'm sure this will be the next memory we'll see top 3D2001 users useing.

OPP

[alexio]

I can get them for 245 Dollar without shipping, do you guys think thats cheap?
Or do you guys know a store wich sells them for less?

[OPPAINTER]

Very nice oc.. should have posted this in the official vx results thread though

what voltage was that ? and is it memtest #5 stable?
I've never run mem test in my life, no need. If my mem can run 2001 lobby tests I'm a happy camper.

As far as voltage, 3.65V idle, Sandra bench drops it to about 3.45V, 2001 lobby tests drop it around 3.52V.

OPP

[OPPAINTER]

Maybe this wil lift you up to 43K, :toast:
My next scores should do that,,, if I ever get my new PCI-E mobos and a Graphics Card:D

OPP

[HermS]

So haveing 1 gig or higher is in if you can use aggressive timings,, so now you have VX:D

Thats nothing but good news! :banana4: Its great being able to score massivley in 3dmark with just 512Mb or less but we all know that most current games run pretty bad with anything less than a gig of ram regardless of clock speeds etc. It's good to know that the gap between overclocked gaming rigs and purely overclocking systems is being closed finally. Mind you I bet it's not long before 1 gig isnt enough to game with either :eek:

[Dissolved]

that is great new's i hope there will be a bigger supply of them than the last batches of mem..seemed to be limited supply.


Are you sick again or something?



-Opp never ran MEMTEST before? Thats a shock..

[mrlobber]

It seems like every pair of these are smacking 250 MHz 2-2-2, with a 2x512MB kit no prob.

BH-5 never consistently did that. Espescially with 2x512MB.

Does anybody have an explanation, why BH-5's in 2x256 config can run up to 300 fsb with enough volts while in 2x512 config they lose out to the VX?

[OPPAINTER]

that is great new's i hope there will be a bigger supply of them than the last batches of mem..seemed to be limited supply.
Thats a good point. I hope they don't discontinue this stuff. Gunna have to get the inside scoop on that, may have to stock up if nessasary.

OPP

[funkflix]

Singlesided vs. Doublesided! (afaik that is the problem)

If they run with 260 2.0-2-2-5 1T i will get a pair of these, tahts nice for
gaming and when i bench i will stick in my BH5s! :)

[cantankerous]

I hear ya. 260 is my goal too. 260 X 10 would be sick with all timings tightened via A64 tweaker. Would be a HUGE jump over my current Mobile XP system.

[althes]

Not bad at all opp.

[Playful_Buffalo]

Thats a good point. I hope they don't discontinue this stuff. Gunna have to get the inside scoop on that, may have to stock up if nessasary.

OPP
do you think it is possible for OCZ to run a product like this long term?

[OPPAINTER]

do you think it is possible for OCZ to run a product like this long term?
I have no Idea. Maybe they will stop by and comment on this.

OPP

[charlie]

Sweet stuff. OCZ has made some nice products the last few months... the Plat R2 (tccd), the VX and one of the most understated is the DDR2 that runs like 300 at 3-2-2-x

C

[conrad.maranan]

Thats a good point. I hope they don't discontinue this stuff. Gunna have to get the inside scoop on that, may have to stock up if nessasary.

OPP
I plan on buying a few more kits as well. ;) And I'm glad to see that you're enjoying your new memory.
:toast:

[xviz]

ocz said in the VX release thread that the product woulnt be discontinue like the EB

[p4z1f1st]

ok, let's say, someone doesn't care on tight timings, what they can do with 2.5-3-3-10-1T-timings for example?

[blinky]

ok, let's say, someone doesn't care on tight timings, what they can do with 2.5-3-3-10-1T-timings for example?this ram doesnt do any better with loose timings over tight ones.... so u might as well run tight ones

[conrad.maranan]

blinky's right. I tried to loosen things up a bit and all I get are errors, regardless of the frequency I'm running them at. VX was made for 2-2-2 timings on the A64 platform.

[ryanpgroovy]

The reason that any ram OC's higher in 256 x 2 over 512 x 2 is that there are fewer banks to address.

It also the same reason a single unit (usualy) ocs faster that 2 , the fewer banks the less stress on the memory controller. I posted a detailed explionation of this a year ago lol .

:p

[charlie]

I checked the price on VX, frankly I was expecting an outrageous price (like that killer DDR2 3-2-2 stuff @ $440) but was a bit surprised to see it at about $250....
GJ, OCZ!

C

[hollywood]

I checked the price on VX, frankly I was expecting an outrageous price (like that killer DDR2 3-2-2 stuff @ $440) but was a bit surprised to see it at about $250....
GJ, OCZ!

C

Where can I get some?????? :(

[texuspete00]

Wow, they are that cheap? That changes everything. OCZ's flavor of the month is usually more.

[conrad.maranan]

Where can I get some?????? :(
After the thousands of posts and numerous threads on the VX, it puzzles me that you wouldn't know where to purchase a set. Fry's Electronics has 'em. mWave (http://www.mwave.com/mwave/viewspec.hmx?scriteria=BA20923) has them as well. And the list goes on...

[xgman]

Anyone know about what the crossover point is where performance of 2.5-3-3-7 with say tccd chips would equal VX at 2-2-2-7? So in other words, how high an htt would 2.5-3-3-7 ram have to run to equal 250 or 260htt at 2-2-2-7 if cpu speed and all else was equal? This assumes the paltform is A64.

[mrlobber]

The reason that any ram OC's higher in 256 x 2 over 512 x 2 is that there are fewer banks to address.

It also the same reason a single unit (usualy) ocs faster that 2 , the fewer banks the less stress on the memory controller. I posted a detailed explionation of this a year ago lol .

:p

Ryan, I think you missed the point or I didn't explain it properly which seems a more probable reason every time I reread my post :D. Actually, the essential question behind it all I asked was, why 2x512MB of VX oc's higher than BH-5 at tight timings. Is the simple answer that the chips used on VX are superior over BH-5? If yes, does that imply that if OCZ made a 2x256 MB kit of VX, that one would scream up to 300mhz DDR speeds much more easily than 2x256 BH-5 did provided that the memory controllers and mobos would be able to keep up with that?

[mcnbns]

If the difference between 2x256 and 2x512 of VX is the same as 2x256 and 2x512 of BH-5/6, then that would be awesome. Somehow, I don't think that will be the case.

[ryanpgroovy]

gotcha

[mcnbns]

Does that "gotcha" mean 2x256 VX would rock? If so, is there any chance OCZ would release the VX line in 256MB modules?

[ryanpgroovy]

oh I was responding to lobber ,

256mb VX are bound to overclock much better than 512mb VX , and I have no clue if/when we plan to start making them .

[Rabbi_NZ]

Posted these elsewhere but guys are asking "will these do 270 2-2-2 3D", "Why don't these scale much past 260 2-2-2 like BH did", "what's the fuss bout, these are no better than BH"....

OPP, hope ya dont mind me posting a few screenies here to answer those questions ;)

My highest screeny with ~3.35vDimm 1x512 VX (http://webpages.charter.net/conrad-digital/Rabbi_NZ/8x275%202-2-2-10-1T.JPG)
My highest SuperPI 32MB with ~3.35vDimm 2x512mb VX (http://webpages.charter.net/conrad-digital/Rabbi_NZ/7x268%20superpi%2032mb%202x512.jpg)
My highest 3Dmark01 with ~3.35v 1x512mb VX (http://webpages.charter.net/conrad-digital/Rabbi_NZ/9x270%202-2-2-10-1T%203D'O1.JPG)

I have a feeling all our VX modules will run a fair bit higher once AMD release CPUs with better memory controllers too :D

[Tony]

Posted these elsewhere but guys are asking "will these do 270 2-2-2 3D", "Why don't these scale much past 260 2-2-2 like BH did", "what's the fuss bout, these are no better than BH"....

OPP, hope ya dont mind me posting a few screenies here to answer those questions ;)

My highest screeny with ~3.35vDimm 1x512 VX (http://webpages.charter.net/conrad-digital/Rabbi_NZ/8x275%202-2-2-10-1T.JPG)
My highest SuperPI 32MB with ~3.35vDimm 2x512mb VX (http://webpages.charter.net/conrad-digital/Rabbi_NZ/7x268%20superpi%2032mb%202x512.jpg)
My highest 3Dmark01 with ~3.35v 1x512mb VX (http://webpages.charter.net/conrad-digital/Rabbi_NZ/9x270%202-2-2-10-1T%203D'O1.JPG)

I have a feeling all our VX modules will run a fair bit higher once AMD release CPUs with better memory controllers too :D

I have a feeling you are correct m8, a LOT has to do with the memory controller ;)

[DoGMaN]

I gotta get me some :o that is just too sweet!

[NapalmV5]

The problem these days is if you don't use 1gig and only have 512Mb like BH-5 in most cases, you lose out on performance. The 1gig gives you 2-way interleave, 2 gigs will give you 4-way interleave. This is on S939 motherboards anyway.
So haveing 1 gig or higher is in if you can use aggressive timings,, so now you have VX:D

OPP

So 4-way interleave doesn't work on 512/1024 MB ?

Man, there's always something new about A64/Mobos.

[NoX]

Did you guys at OCZ perform some testing on Intel platforms? I can't find anybody with a P4 who tried this mem and I don't understand why. Did I miss something?
Also how does the VX compare to the EBs at the same timings? I.e. both at DDR500 at 3-2-2-5. Would the "Enhanced Bandwidth" make a difference?
In the end, I have to go to Fry's tonight and I want to know if with my rig it's worth to give the VX a try.

Thanks!

[Micutzu]

Intels do not benefit that much from the tight timings, TCCD gives better results. There's really no point in using 260Mhz 2-2-2 when 300Mhz+ 3-4-4 is faster and usually needs under 3.0V to run.

[cuddles]

It's mostly made for A64 systems. Those with Intels don't seem to get as good results, it performs like any other ram from what I've seen on Intel.

[quicksilverXP]

Would it be faster on an A64 to run memory at 300Mhz at 2.5-4-4-8 (or sometimes 3-3-7) or to run these VX on a divider? I'm contemplating on switching from my TCCDs.... but I know for sure these can't hit 300 HTT 1:1... and I only have an Athlon64 3000+ with the 9 multi.

[KuRi]

Quick tests, 100% stable:

http://koti.mbnet.fi/rnm/discipline/3200VX/

[NoX]

There's really no point in using 260Mhz 2-2-2 when 300Mhz+ 3-4-4 is faster and usually needs under 3.0V to run.

I would agree with you if I had a cpu that could run at 300Mhz FSB.

[Grov]

Yeah the price ain't to bad on em, cheaper than the 3200EB TCCD or whatever it is, well here in the UK anyways.

Sold my Corsair C2PT 2x512mb for 110quid, got these for 190quid, so not to bad there. Big upgrade.

Gotta get a booster though. :mad:

[Tony]

I hit 250fsb 2-2-2-5 on the 875P-T at 3.2Vdimm dual channel, wxchaser hit 260+ on a p4c800 so they can clock well on these rigs also.

[ojdr2001]

For some reason I'm not getting more than 235 stable 3D 2-2-2-5 DC 1T :(

Did this @ 3,5V. Went up to 3,75V and no gains...

MSI K8n Neo2 FX55 bios 1.36b

Could it be MOBO Bios or just bad luck on this pair????? :(

And for some reason cannot change CAS in A64 Tweaker. Have to boot with cas 2 from bios...

[ryanpgroovy]

intels run well at 2.5-2-2 but not at 2-2-2

[OPPAINTER]

Just did a 3d2001 comparison of my Mushkin BH5 512Mb and my VX 1Gig

ATI 9800 Pro, at stock clocks
MSI Neo2 and 4000+ at 10x260 2600MHz
Memory at 260Mhz 2-2-5 Cas2 A64 tweaked out the same
Windows XP SP2

You'll see the 1Gig of VX utilizing the 2-way Interleave beating out the 512Mb of BH5 and No-way Interleave:D

I believe that if this CPU was doing say 13x260 at 3380MHz this difference in the score between the 2 memories would be much larger.

OPP

http://www.xtremesystems.org/forums/attachment.php?attachmentid=21012&stc=1

[MaxxxRacer]

interesting info you got there opp.the difference is inperceptible to our senses, but its still there.. going for that last extra fps the vx do help out

[Rabbi_NZ]

OPP, do you know if the Interleave benefits are also seen on Socket 754 or only the DC AMD64s?

[cuddles]

VX wins!! *ding ding ding*

[OPPAINTER]

OPP, do you know if the Interleave benefits are also seen on Socket 754 or only the DC AMD64s?
Sorry, I don't remember seeing it, but then again I don't think I ever used a Gig on 754. May or may not be there:D
Looks like you have a gig on a 754, do a Sandra mem bench and see if it says it's there.

OPP

[G|-|oST]

OK what exactly is the 'interleave' referring to? 1T/2T?

[OPPAINTER]

Nothing to do with 1T. Heres somthing I found on the net, old school.

OPP

-------------------------------------
The chipset can interleave requests to different DRAM banks and gain a massive performance boost from it (relatively speaking). Without interleave, a typical session would be:

Latencies while refreshing DRAM
Send address to bank w
Receive data from bank w
Latencies while refreshing DRAM
Send address to bank x
Receive data from bank x
Latencies while refreshing DRAM
Send address to bank y
Receive data from bank y
Latencies while refreshing DRAM
Send address to bank z
Receive data from bank z

This makes extremely poor use of the bus, since we can refresh the DRAM simply by reading or writing to it.If we have some interleaving going on, we make far better use of the bus.

Latencies while refreshing DRAM
Send address to bank w
Send address to bank x, receive data from bank w
Send address to bank y, receive data from bank x
Send address to bank z, receive data from bank y
Receive data from bank z


All four banks can be accessed in turn, without having to refresh the RAM and wait the killer CAS. This is four-way interleave. Two-way interleave is only two banks at a time.

The only time you will be unable to set 4-way is if you're using a very small amount of RAM. Every DIMM has at least two banks, but 64MB and smaller DIMMs have only two. All larger DIMMs have four. However, if you use two two-bank DIMMs, you have a total of four banks and can use 4-way. Always set this setting, then, to 4-way and don't expect any problems from it. Bank interleaving was not covered in the RAM guide, so I've gone into a bit more detail here.

[DoGMaN]

This is single channel DDR400 2-3-3-7 MSI K8N Neo with a 3200+ CH 4-way interleaving

http://www.thecrucible.ca/~dogman/sanmem.jpg

[hollywood]

After the thousands of posts and numerous threads on the VX, it puzzles me that you wouldn't know where to purchase a set. Fry's Electronics has 'em. mWave (http://www.mwave.com/mwave/viewspec.hmx?scriteria=BA20923) has them as well. And the list goes on...


Thank you Conrad. We don't have Fry's on the East Coast and thats the only place I saw people talking about.

oop's hit edit..my bad ment to quote.

[3NR1C0]

The problem these days is if you don't use 1gig and only have 512Mb like BH-5 in most cases, you lose out on performance. The 1gig gives you 2-way interleave, 2 gigs will give you 4-way interleave. This is on S939 motherboards anyway.
So haveing 1 gig or higher is in if you can use aggressive timings,, so now you have VX:D

OPP

Sorry Opp, I don't understand. Can't I enable 2 or 4-way interleave 2*256 BH-5 with? :confused:

[Sgt_Strider]

I'm curious here, is it the reason why these RAM are so good is because you can pump some serious voltages into them? How does it overclock if you use <3.0 v? Like maybe 2.8v?

[3NR1C0]

skt 939 as well.

[RocKer]

Anybody got some info on how the VX is doining on nF2 boards,i can try it but i got only 1x512mb of VX,wont there be a big differens in running sc on nF2,or is it better to run dc on it.

O hell i'll try it anyway:D.

[mrlobber]

Anybody got some info on how the VX is doining on nF2 boards,i can try it but i got only 1x512mb of VX,wont there be a big differens in running sc on nF2,or is it better to run dc on it.

O hell i'll try it anyway:D.

On nforce2 dual channel with 2x512 sticks won't get you anywhere with 1t anyway.

[cantankerous]

I am running them on my DFI LPB NF2. Stay away. With 3.3v I can't get a lick over 243 stable. Windows anything over 200mhz causes very weird behaviour. It isn't the sticks it's the weak chipset. My BH-6 is doing much better on this board than the VX. I hope with more volts and a better A64 board I can do higher than I am now.

[RocKer]

Ok thks for reply on nF2;

On s754(rig in sig) with VX i get a lot of blue sreens with"0x000000D1 NVATABUS.sys" i have been looking at MS site for info on that but can't find it,dos enybody know what this is about.

I think its got to do with my Maxtor HD 120gb 8mb,but am not sure,don't have an other hd to try out if its any better,only one i have is an old Samsung 4gb,can i try it with this 4gb hd or is it better not.

[Bluebeard]

http://home.arcor.de/twinmos.support/WB_UTT.JPG


new winbond chips (can be found on TwinMOS/M.tec - cheapest TwinMOS DDR400 module!) fired with 3,5V - so i believe thats what is under the VX also....

[KPJ]

Alonso,

Where did you get this????

[H2OGun]

And, what's the part no. please?

[Bluebeard]

http://www.forumdeluxx.de/forum/showthread.php?p=1360650#post1360650

here u find a bad quality pic (the 2 ones down) - these are on sale in germany and already should be in distribution channels...

but not sure if all M.tec use same IC...

[KPJ]

I can't connect. :/

Server is busy I guess????

[Bluebeard]

works fine here...

[Reflex1]

So which one is better, TCCD or VX? i can't decide.

[Bluebeard]

if u have enough vdimm i´d go with the vx (or any other module using new winbond ic ;) )

[KPJ]

Alonso,

Now it works. Looks good. Thanks for the heads up. :)

[georgesod]

So winbond has brand new ic's? Are u sure?

[jlccarv]

Someone quick, grab that cat! :p:

get it, cats outta the bag. Ok bad joke...

[funkflix]

So winbond has brand new ic's? Are u sure?

Infineon bought Winbond and these ICs are produced at the old
winbond-factory for Infineon! So the ICs are Infineons. labeld as M.Tec <<-- Afaik!
Maybe some new original Infineon sticks have the same ICs on it?

[cantankerous]

NVATABUS is the IDE driver from the Nvidia Nforce package. MS won't have anything on that cause the file in question isn't a direct part of windows.

[Zeus]

Ok thks for reply on nF2;

On s754(rig in sig) with VX i get a lot of blue sreens with"0x000000D1 NVATABUS.sys" i have been looking at MS site for info on that but can't find it,dos enybody know what this is about.

I think its got to do with my Maxtor HD 120gb 8mb,but am not sure,don't have an other hd to try out if its any better,only one i have is an old Samsung 4gb,can i try it with this 4gb hd or is it better not.

If you suspect your harddisk, try running memtest86+ from a floppy.

[jinu117]

So which one is better, TCCD or VX? i can't decide.

Depends on how high FSB.multi you will get. For this winchester 3000+ VX probably is much better as I go down on memory divider and end up having to do around 250-260mhz tops, while my 3500+ newcastle is better off around 280-300mhz. It also depends heavily on memory controller of the CPU unfortunately as even with higher multi you might have to go down on divider for max overclock if your memory controller can't keep up.
Choices, choices.

BTW, opp :) I beat you in 1 stick... out of 4. 267mhz 2-2-2-8 running fine @ ocz memtest86 #5 fine for hours now. The other 3 were not as good though :P

[ryanpgroovy]

Alonso ,

As far as I can tell MTEC is normaly PCS die , so if people want to take there chances buying mtec , we really are not bothered.


The funny thing is early today we got a blackmail letter , saying the same thing your posting , that we have found some majic new winbond chips.

The one comment I have is even if people figure out whos making these chips , there still not going to be speed screened , and only 20-30% of the current raw parts pass our internal test spec. If you really want to take your chances on something I can get the raw un- binned parts made into a special part number , which you can then order at a much lower cost ... no guarantees it will do as well as vx in every case though


Let me know :-)


Oh , here is the funny blackmail letter

" Hello,

i know what the chips are on your VX ram. They are Winbond UTT blank chips :> Very cheap chips.

And also know that you, ocz, don’t want that this is going to be puplic, cauz more than 100% win on one 512mb DDR-Ram is very much ;)

I’ve got 50 Mtecs with Winbond UTT…

So I think you’re pleased that I keep my Mouth shut … cauz u would like to make much money as possible, that’s normal I know.…

(mtec oem about 39Eur/512mb)

I think there could maybe a deal… for keeping my mouth closed or do you want that everybody know that this ram have one of the cheapest chips on the market…



Hope for a fast reply.

In Rar archive some screens of my Winbond UTT…


Grats Martin"

[KPJ]

ryan,

LOL!!!! :) What did you offer him? ;)

The best way to get proper VX modules is to go OCZ. No doubt about it. But the Mtec is worth a gamble IMO. I will probably buy the VX (for benching) when it's released in 256MB modules if they reach much higher speeds then 512MB modules. Any info about that?

[kristos]

Martin strikes me as a very smart guy :D

[ryanpgroovy]

But it has me thinking we should release unbinned vx for those crazy guys who like to take chances .. We could eliminate the cost of speed binning and hand testing . We could sell them cheaper.


Call it VX value series ;-)

[KPJ]

Might be a good idea with the VX value.

On Brainpower PCB.... :)

[Torin]

What exactly does your speed binning process determine? Whether they can run 2-2-2 @ a specific voltage? How do the chips that don't pass speed binning for VX run? Do they just not go as high at 2-2-2, or do they not hit 2-2-2 at all, or...? What does a chip have to test out at before you bin it as VX gold?

[enzoR]

what does PCS die mean?

[rapra]

Hello!
I have few comments:
1. Infineon CE-5 have the same specific - they have a frequency, over which relaxing of timings do nothing. But I don't tested they over 2.8-2.9V anytime, because voltage do nothing for them.
2. I saw TwinMOS on winbond chips in few latest weeks - they are on the cheapest modules. But they are not very good - from maybe 50 pieces the best works at 225-230MHz 2-2-2 @3V in memtest86. Of course they like voltage like typical Winbonds.
3. Who will check what chips are under the VX's heatsinks? Winbonds are very easy to recognize...

P.S. maybe a title of this topic should be: 'OCZ VX, winbonds again'?

[Tony]

PCS is a make...a manufacturer.

Speed binning of VX ensure every module does at least 240fsb 2-2-2.

[enzoR]

Oh, k that makes sense. Unless Winbond err infineon makes tested chips rated at PC3200 with specific timeings it is much more worth it to just buy VX, since those are actually tested and will perform good.

[HiJon89]

PCS is a make...a manufacturer.

Speed binning of VX ensure every module does at least 240fsb 2-2-2.
Only AMD though right?

[enzoR]

well that shouldnt matter, if the chips are capable of 240 2-2-2 on amd the chips should be capable of that in any platform aslong as they dont have compatibility probs. like the 875p only likes them with cas 2.5 or something. thats a chipset thing.

[Rabbi_NZ]

bigtoe, ryangroovey...

how about making a PC3500 VX or a PC3700 VX that is still rated at 2-2-2 @ 3.2v?

I know both my modules do over 250 2-2-2 @ 3.2v.... sure the price could be higher to account for the extra labour, but I'm certain they would sell just as well as the current brew

[ojdr2001]

I have tested almost all bios on MSI K8N Neo2 + FX55 and my VX does max 230/235 2-2-2-5 1:1 (3,5 to 3,7V) 1T DC

What´s wrong????? or... where are the good ones??? :)

[Rabbi_NZ]

tried another CPU?

[cuddles]

Guaranteed 240 @ 2-2-2. I love you OCZ ;)

[ojdr2001]

Why another cpu? My mushkin Level2 (BH5) 2x256 do 280 1,5-2-2-5 with same cpu....

problem cannot be cpu or system.....

[Rabbi_NZ]

2x256 & 2x512 are not the same... maybe your memory controller has an issue with 2x512 setups.
maybe it is the RAM too, but it would pay to try another CPU to see if the memory all of a sudden goes higher....

[enzoR]

2x256 is a big difference on A64 than 2x512 no?? i dunno if its that significant but still.

Edit: damn rabbi u beat me to it by some seconds :D

[Rabbi_NZ]

Edit: damn rabbi u beat me to it by some seconds :D
hehehehe 2 voices are better than one

[ojdr2001]

Ya quite a difference of course. I know I cannot expect 270 or above from this ram but at least 250/260 yes..

Don't believe it´s cpu or mem controller. I have tested 2x512 OCZ 4200 and got stable 295 1:1 2,5-3-3-6 1T DC (there are screens in this forum)

So my setup sure likes 2x512 :)

may be I was unlucky or the mems are not that good.... :(

[enzoR]

hmm contact someone from OCZ for some help

[Grov]

Is it ok to run the VX at 1.5-2-2-10. CPU-Z doesn't report the 1.5, but i know it's not accurate for stuff.

I ran 3dmark fine, is it woth keeping it at that or 2?

[cuddles]

Run benches at 2 and 1.5 and find out yourself. :p:

[conrad.maranan]

Is it ok to run the VX at 1.5-2-2-10. CPU-Z doesn't report the 1.5, but i know it's not accurate for stuff.

I ran 3dmark fine, is it woth keeping it at that or 2?
Check the official test thread. Rabbi_NZ runs his at 1.5-2-2-10. CPU-Z doesn't report CAS1.5, but it doesn't negate the fact that it's running at that latency value.

I tested with CAS1.5 on my VX modules and it made no difference. Everyone has a different set-up, even when the components are exactly the same. Test with 1.5 and 2.0 to discover what setting is best for your system.

[Tony]

Guys we test at 240fsb on the Neo2. The modules seem to not do as well on Intel rigs although WXchaser hit 260 with 3-2-2 and I hit 250 on the 875P-T with 2-2-2 so they do perform as long as the board does.

We have no plans to release them at 3700 speed as this may drive price up, (things do change though)

We are doing our best to keep prices as keen as we can with the extra testing involved...im sure you understand.

[cantankerous]

I know I do. I'm just dying to get a new board to really push these things. I need to sell off one my kits for cash. Whichever does worse. Buying 2 gigs of ram broke me.

[Playful_Buffalo]

oj, people usually benefited from burn in


LOL @ letter

[SteveOCZ]

LOL @ letter

Agreeeeeeeed :banana:

[RocKer]

If you suspect your harddisk, try running memtest86+ from a floppy.
All run test5 of oczmemtest at 2-2-2-10.

10x230 at 3.50vdimm 59195 errors 10passes
10x228 at 3.50vdimm 209 errors 10passes
10x227 at 3.50vdimm 48 errors 10passes
10x226 at 3.50vdimm 4 errors 10passes
10x226 at 3.40vdimm 920 errors 10passes
10x225 at 3.40vdimm 146 errors 10passes
10x224 at 3.40vdimm 0 errors 10passes


9x234 at 3.60vdimm 12.352 errors 10passes
9x232 at 3.60vdimm 1052 errors 10passes
9x230 at 3.60vdimm 0 errors 10passes

8x234 at 3.50vdimm 9603 errors 10passes
8x233 at 3.50vdimm 237 errors 10passes
8x232 at 3.50vdimm 8 errors 10passes
8x231 at 3.50vdimm 0 errors 10passes

Q: what is the best performance setting,is it 10x224,9x230 or 8x231,its now running at 10x224 at 3.40vdimm and prime stable.

[Rabbi_NZ]

Rocker, what is your VCore?
10x224 will smoke your other scores!

[RocKer]

Vcore is 1.550/104% so something like +-1.62v.

And what woot be the best setting to burn'in the mem,becaus 10x224 isnt that great hè.

When it just got them the hit 10x250 at 3.40vdimm and 265 at 3.50vdimm,what is the matter with this mem :confused: .

[Rabbi_NZ]

Should be enough....
Could you list:
*ALL your memory timings
*ALL voltage settings
*ALL multis & dividers
*Is you CPU a CO or CG stepping
*What slot is RAM in

[RocKer]

The a64 rig is not connected to inet,but i will make sreenshot of a64Tweaker and post it here.

a64 3200+ Newcastle
ADA 3200 AEP4AX
CBAEC 0430 RPMW

1 stick 512mb VX

its running at 1T 10x224 at 2-2-2-10-15-16,mem is in slot 2 at the moment,before it was in slot 1 also in slot 3 and i had the same results.

vcore 1.550v 104%
vdimm 3.40v
chipset 1.6v

No AGP becaus a use a pci TNT2 32mb.

Multies or dividers:its running 1/1 that is the only one am intrested in.

But i am gone try my bh5 to see if the can hit higher,and to see if the board is still ok or not i hit 9x275 (http://users.pandora.be/RocKer/A64%200430%20RPMW/Everest%20Bench/Everest%20bench.JPG) with my corsairs on the dfi.

[Rabbi_NZ]

RocKer,
I just remembered!
wait for Conrad to come back online... from his testing he found his NC would not run in SC with VX, but in DC it would run fine.
My CH was fine with either configuration...

[RocKer]

OK

But what i dont understand is that this VX stick did run at 9x265 3.5vdimm the first week i got it in the house,and now i can only hit 9x230 :( at 3.6vdimm,i dont understand it anymore.


from his testing he found his NC would not run in SC with VX, but in DC it would run fine
I can only do SC Rabbi_NZ ;).

[Zeus]

How's your PSU?
Maybe your voltages go down too much?

[conrad.maranan]

But what i dont understand is that this VX stick did run at 9x265 3.5vdimm the first week i got it in the house,and now i can only hit 9x230 :( at 3.6vdimm,i dont understand it anymore.

Yo RocKer. Check this (http://xtremesystems.org/forums/showthread.php?p=635515#post635515) out. What motherboard are you using? If you're using an MSI K8N Neo2 Platinum, see what happens if you run dual channel.

EDIT:
I just noticed your sig states that you're on the DFI 754 board. I'm not sure why you can't get up to 265MHz anymore. When you were at 265MHz, how many loops of Memtest Test #5 did you run? 10 loops usually isn't enough. I run 20 to 30 just to get an idea of what direction I'm heading. When I hit a wall, I'll loop Memtest overnight at my last error-free frequency.

[Bluebeard]

well - this letter is more then funny... :wierd:

M.tec is TwinMOS low-cost brand - there could be every IC underneath... but of course cheepest IC out there (which atm is also Winbond UTT) used and as you can see it can perform as good as the VX

Winbond is not dead... they have a deal with Infineon to use Infineons Technology for the new production - on the other side with this agreement they are not allowed to put Winbond Label on the ICs - afaik...

[Sgt_Strider]

I'm curious here, is it the reason why these RAM are so good is because you can pump some serious voltages into them? How does it overclock if you use <3.0 v? Like maybe 2.8v?

bump, can someone answer my question in bold?

[cuddles]

Sucks at 2.8v. It's made for 3.2+, it's rated for 2-3-3 @ 2.8v. It's called Voltage Xtreme dude, 2.8v isn't eXtreme :p:

[RocKer]

Yo RocKer. Check this (http://xtremesystems.org/forums/showthread.php?p=635515#post635515) out. What motherboard are you using? If you're using an MSI K8N Neo2 Platinum, see what happens if you run dual channel.

EDIT:
I just noticed your sig states that you're on the DFI 754 board. I'm not sure why you can't get up to 265MHz anymore. When you were at 265MHz, how many loops of Memtest Test #5 did you run? 10 loops usually isn't enough. I run 20 to 30 just to get an idea of what direction I'm heading. When I hit a wall, I'll loop Memtest overnight at my last error-free frequency.
When it was at 9x265 3.5vdimm it run test5 for +- 2hours with 0 errors,then it folded D2OL for 3day's and 2 nights 24/7,thats 100% cpu-load,if thats not stressed enough i dont know what will:),and to be sure i run +- 2 hours prime ontop of D2OL and that was also rock solid stable.

Then afther those 3days i reboot to go up with htt and it crasht,i try'd everything i know but it did'nt help,then i reinstalled XP(i thought maybe that help)but it did'nt,i thought maybe its the bios(J&J 4.0v bios)i reflasht the bios with the one with that memtest inside in hope that woot help,but again it was a nono,i had allsorts off troubel with volts on vdimm, then i reflasht it again to the J&J bios and the lines where back ok(checkt with multiemeter),but i was not abel to go up with HTT on whatever vdimm,so now i am running at 10x224 :( at 3.40vdimm,memtest 0 errors and prime stable in XP.

O yes the test where run all at 2-2-2-10 or 2-2-2-5.

My 520w OCZPowerstream works ok.

I still have to try my 2x256mb bh5 to see if the can run as before(the did 9x275 error free :) )did'nt have the time to do it yet,but if i find time today i will try them.

Can you give some info on what to do to burn'in the mem and easy go up with htt,or something becaus i dont know what to do anymore :( ,it was looking so good the first week but sadly it did'nt last.

So now am wondering if the dfi(again) is to blaam or cpu or the memory,i can't try it on an other board becaus i dont have an other s754 board.

I'am now serious begin to consider to quit s754(sell it or make it a D2OL rig) and go over to s939 in DC.

I have got no luck yet with a64 on s754,it was my first one,it was allways something that whent wrong,almost everything is gone in rma,first it was the DFI then cpu then the OCZ 520w,and what is it now ppffff.

[Bluebeard]

@ ryan

so can you say that u use new winbond (of course selected) IC for the VX?

u r absolutely right M.tec can be everything - Winbond, PSC, ...

but new (not the ones out on market atm which are selected PSC) TwinMOS Speed Premium which will be out soon will all come with these new Winbond IC.



Alonso

FAE TwinMOS Germany ;)

[georgesod]

@ ryan

so can you say that u use new winbond (of course selected) IC for the VX?

u r absolutely right M.tec can be everything - Winbond, PSC, ...

but new (not the ones out on market atm which are selected PSC) TwinMOS Speed Premium which will be out soon will all come with these new Winbond IC.



Alonso

FAE TwinMOS Germany ;)
Have u done any testing on those? What can we expect?
:toast:

[xgman]

Alonso ,

As far as I can tell MTEC is normaly PCS die , so if people want to take there chances buying mtec , we really are not bothered.


The funny thing is early today we got a blackmail letter , saying the same thing your posting , that we have found some majic new winbond chips.

The one comment I have is even if people figure out whos making these chips , there still not going to be speed screened , and only 20-30% of the current raw parts pass our internal test spec. If you really want to take your chances on something I can get the raw un- binned parts made into a special part number , which you can then order at a much lower cost ... no guarantees it will do as well as vx in every case though


Let me know :-)


Oh , here is the funny blackmail letter

" Hello,

i know what the chips are on your VX ram. They are Winbond UTT blank chips :> Very cheap chips.

And also know that you, ocz, don’t want that this is going to be puplic, cauz more than 100% win on one 512mb DDR-Ram is very much ;)

I’ve got 50 Mtecs with Winbond UTT…

So I think you’re pleased that I keep my Mouth shut … cauz u would like to make much money as possible, that’s normal I know.…

(mtec oem about 39Eur/512mb)

I think there could maybe a deal… for keeping my mouth closed or do you want that everybody know that this ram have one of the cheapest chips on the market…



Hope for a fast reply.

In Rar archive some screens of my Winbond UTT…


Grats Martin"

That is the silliest thing I have heard in a long time. You should arrange to meet him to drop off the loot and we will all jump him and take his clothes. :grr:

[Bluebeard]

Have u done any testing on those? What can we expect?
:toast:

Speed Premium with WB to come out soon (the ones already out are PSC) - should perform well if the chips are the same as on the M.tec modules we tested and talking about.

[Zebo]

micron g? I've heard ballistix close to this... maybe OCZ just selects them better than crucial?

[perkam]

You should arrange to meet him to drop off the loot

Off topic...


and we will all jump him and take his clothes.

VERY OFFTOPIC !!! :eek:

Seriusly though, crucial just selects them better for the reviews not the consumers :D

Perkam

[DoGMaN]

Funny, I have some SuperTalent memory that acts a lot like VX. Hrmmmz I wonder what is under the spreaders.

Couple things I'd like to note though for all you looking for a cheap way to get the same ICs as VX. First as we have seen from the countless TCCD reviews the OCZ/PDP/PQI modules always seem to work better because of the better PCB design. Much the same will come from buying cheap modules with the same ICs. The OCZ VX will just work better due to the better PCB design.

[hollywood]

Thank you Conrad. We don't have Fry's on the East Coast and thats the only place I saw people talking about.

oop's hit edit..my bad ment to quote.

hey...whats going on here?!?!? I NEVER wrote the part in bold...Why is someone changing my posts???

[Shroomalistic]

heres pc ice's vx running on my dfi at 3.6volts

[Playful_Buffalo]

Stable?

[Shroomalistic]

not yet, still playing around. I was able to get it to 280 2,2,2,5 but frozen on screen shot

[Dumo]

is it possible that vx is tccd 437(week 37 and up)?

[Playful_Buffalo]

^sometimes the week 37+ TCCD's didnt like voltage


in was inconsistent

[MarilynMX]

I have a question. Should I run memtest86 with default #7 test? or I need to run #5 only since lota ppl preferring it?
Thanks in advance

[jinu117]

I would clear the #5 first than go #7. If I have time I would run all test but these 2 usually tells me if I am going to hit wall or not.

[MarilynMX]

I left it with #7 for 2 hours(about 8 passes) and no errors, should I move on to #5 now?

[jinu117]

Sure sounds good :) After all that is done, I suggest testing it with windows memtest as well. Even the ones that pass memtest 86 have tendency to fail on windows memtest especially on A64,

[MarilynMX]

Yes that's true. At the very first time there were more than 4 errors within 10% but now seems fine with about 2000%
And one thing that is confusing me, why my OCCT CPU toture failed at 250x10 before I went for memtest86 and now it could last for 1 hr + now? is there something to do with the on die memo controller or just about the memory?(I quit the program since I had to order pizza online)
Thanks for getting back to me Jinu117

[jinu117]

There seems to be certain break in happening with VX from all reports I've seen. I don't know as i've only had about 1 hours tops on each VX chips I tried yet and more worried about divider atm. You might have to keep testing over time to see where your break in ends :)
Unless your mobo is brand new, I doubt it is mem controller or mobo that makes you able to fly now. Gotta be something with memory itself.
I need to ask you a favor :) Have you tried divider on memory and see if it causes you problem? Thanks in advance. Most of us seems to have problem over 260FSB once divider is set to 166.

[G H Z]

Wow, is this stuff discontinued now????

MWave say's so on their site. I hope that's not the case :(

Only place I can find it is NCIX out of Canada for $281 US

[jinu117]

Interesting... either Fry's sucked up all inventory dry or the memory batch isn't as good as it used to be to mass produce this. Or RMA rate is too high where it isn't worth selling the product. Think first one is valid if the deal wit Fry's went through.

[jikdoc]

it would suck if the vx was discontinued, i run my system at 260 and was looking forward to trying out a pair

[Playful_Buffalo]

i dont think it can be discontinued 1 month after its launch....

[kromosto]

discontinued i hate this :( last time i decided to buy ocz 3700eb ordered it to a friend coming from usa he cant find 3700eb and get me a 3500eb now vx gone :( hey ocz people can you do rezervation for me at least for ddr booster i missed the vx train lol :D

[Noktar]

How you cool the Memor at these voltages ?
Only the HS or a Fan (more) ?

[Playful_Buffalo]

you can just place a fan over the memory modules^^^^

EDIT: here is a thread on the topic (http://xtremesystems.org/forums/showthread.php?t=47413)

[DevilsRejection]

and i grabed a pack from mwave of vx and then 5 hr's later it was off there site.

like yesterday.i think i got there last pack..

but it would be great new's for me if fry's get's those in..but realy fry's never gets the good stuff..so i feel that fry's geting the good vx is 100 % incorect.
finally someone other than me who bought vx from mwave.

i havent tested it yet due to extreme shortages of 3500+ processors but i think it will be better post xmas.

[JasonDTM]

I've been eyeballing this memory since the original VX thread was posted, I'd like to see a 2x256 mb kit.... to compare it with my BH-5 :/

Now I just need to get an A64 and I'd be set :D

[jinu117]

Extreme shortage of 3500+? Interesting. FYI, there will be 4 coming up for sale (3 winchester and 1 newcastle in few days). used less than 1 hour each :P For winchesters 1 boots no problem with -55c or below, 1 boots fine -55c above, 1 doesn't like -30c below. New castle hasn't arrived yet (monday) but I have feeling I am going all out FX-55 in few days anyway.

Check the FS section on CPU tuesday or wed :)

[stevehat1]

and i grabed a pack from mwave of vx and then 5 hr's later it was off there site.

like yesterday.i think i got there last pack..

but it would be great new's for me if fry's get's those in..but realy fry's never gets the good stuff..so i feel that fry's geting the good vx is 100 % incorect.
Gone already......glad I got some ordered from MWave on Wed. It's sitting in the Fedex dist. center, should have it Mon.
Same thing happened with 3700EB....got some from MWave, and a week later it's gone :lol:

[Dumo]

Prolly samsung will come out with another 'good' ones. As for vx, imo ocz already exhausted their chips that passed speed binning. Is this gonna be a trend? Released a batch of high perf. module, discontinued and move on.....

[MarilynMX]

To Jinu117, no I haven't tried memo divider yet but I will do that for you. Last night I ran 288 passes #5 ot 250MHz and there's been 1 error per pass since loop 5x. And later I tried 255MHz and there were 20 error per pass average. So should I go back to 250MHz or stick on 255MHz and see what's gonna happen?

[Tony]

Prolly samsung will come out with another 'good' ones. As for vx, imo ocz already exhausted their chips that passed speed binning. Is this gonna be a trend? Released a batch of high perf. module, discontinued and move on.....

No this is not a new trend...the issue is we pre test ram and not all modules pass the test....we could just realease without pre testing but im sure you would all hate us within a few days.

[JDizzle]

So what's going on? Is OCZ pre-testing slower than the demand?

[JDizzle]

Sorry, I didn't mean to make it sound so harsh, I was just trying to make sure that they weren't discontinued like everybody was saying.

[Playful_Buffalo]

OCZ is teh pone :banana:

[FallenAngel]

OCZ is teh pone :banana:
It's true I'm not a good english writer, I don't want to be annoying but can you tell me what did you wrote, please... :(

[MarilynMX]

me dont understoodz too XD

[Playful_Buffalo]

LOL, it means OCZ is amazing

[Revv23]

its brings back a good feeling when i hear about the burn ins and voltage that i remeber from bh-5, its almost like winbond makes living chips that give you a big kiss when you push them ;)

though it is wierd how differently these react with low and high voltages then previous winbond designs...

i can only hope the increased sales of these IC's would inspire winbond/infineon to make some higher end IC's

[LikwidKool]

VX is some special ram. I have mine running at 260 2-2-2-10-1t with 3.5v Tried up to 3.6v and saw nothing more. Still burning in. Trying for 270x10 as that is the sweet spot on my 3500+.

BTW OPP, nice to see your seal of approval on these. The OCZ guy's deserve it!:toast:

[kromosto]

i will try to get vx and booster in a short time if i find a friend returning from usa

[Tony]

I hope you guys can now see why we chargle a little more for our modules. All are pretested, and backed by a lifetime warranty. We also have a growing team of forum whores who are here to look after you guys and answer all your questions...;)

So as you see we pretty much are the most hands on enthusiast based company there is...and we are pushing to bring you even more awesome products and customer service.

More VX is coming...you just need to keep your eyes out for stock as it sells out in mins on every site it is posted on.

[bachus_anonym]

yeah, this IS great piece of RAM! and actually, i don't think that $259 is expensive for such great performing sticks. i remember, when i first bought 1GB of Mushkin LvL (almost 2 years ago) i spend $310 on it and it performs far worse than OCZ VX i have here!
ATM, i'm flying @ DDR520 6-2-2-2.0 1T - something i could dream of with 2x512MB Mushkin LvL II.

excellent job on selectioning and pre-testing whatever ICs you guys found ;)

thanx guys,

[georgesod]

I hope you guys can now see why we chargle a little more for our modules. All are pretested, and backed by a lifetime warranty. We also have a growing team of forum whores who are here to look after you guys and answer all your questions...;)

So as you see we pretty much are the most hands on enthusiast based company there is...and we are pushing to bring you even more awesome products and customer service.

More VX is coming...you just need to keep your eyes out for stock as it sells out in mins on every site it is posted on.

However the prices in Europe are outrageous comparing with the ones in usa. Not only talking about ocz but for all.

[kromosto]

georgesod it is still a luck for you in here we dont even find good rams

[Nube]

Can sombody PM me when someone in UK sells them? ?
I wanna get some soon and I missed the gladiator selling them, they not sell 1gig anymore, only 512VX ram, but if i buy two 512sticks, I can run them in dual channel and there just like buying 1 gig (2x 512) ?

[georgesod]

georgesod it is still a luck for you in here we dont even find good rams
Neither do we. We just order them from online shops, pcmemoryupgrade for example

[georgesod]

Can sombody PM me when someone in UK sells them? ?
I wanna get some soon and I missed the gladiator selling them, they not sell 1gig anymore, only 512VX ram, but if i buy two 512sticks, I can run them in dual channel and there just like buying 1 gig (2x 512) ?


link (http://www.pc-memory-upgrade.co.uk/memory/ddr-ram-pc3200.asp)

[jinu117]

I hope you guys can now see why we chargle a little more for our modules. All are pretested, and backed by a lifetime warranty. We also have a growing team of forum whores who are here to look after you guys and answer all your questions...;)


errhhh... sorry to say, but YOU WERE one to begin with (the P W) before you joined OCZ :) Not that it was bad thing. Lot of people benefitted from your experiments and some conflicting investigation with eva2000 or someone :)

Let's hope we can find something about the divider issue of VX soon before we can claim it to be best memory made since BH5 though :). A little surprised that there was no divider test done on MSI K8N neo2 or I bet half the people with Phase cooling wouldn't have bothered with it to begin with...
(Bios 1.3,1.36,1.37,1.4,1.41,1.51 doesn't fix this issue at all)
Let's see how it unfolds.

[bachus_anonym]

yeah, that divider issue really suxx.... i still believe it must be BIOS related / a way board handles VX...
OCZ really needs to look into it....

[OPPAINTER]

errhhh... sorry to say, but YOU WERE one to begin with (the P W) before you joined OCZ :) Not that it was bad thing. Lot of people benefitted from your experiments and some conflicting investigation with eva2000 or someone :)

Let's hope we can find something about the divider issue of VX soon before we can claim it to be best memory made since BH5 though :). A little surprised that there was no divider test done on MSI K8N neo2 or I bet half the people with Phase cooling wouldn't have bothered with it to begin with...
(Bios 1.3,1.36,1.37,1.4,1.41,1.51 doesn't fix this issue at all)
Let's see how it unfolds.
I never tryed a divider with this memory.
What seems to be the issue?
I'll have to plug in the VX today and see what your reffering to.

OPP

[bachus_anonym]

I never tryed a divider with this memory.
What seems to be the issue?
I'll have to plug in the VX today and see what your reffering to.

OPP
check that, OPP please....
well, what happens is that above 270ish (no matter what multi, it seems) we get NO POST with memory divider of 166...
i'm playing with it now and it seems that once mems hit DDR446 with 166 divider system either doesn't POST or crashes (if upping HTT with clockgen in Windows)...

[jinu117]

And yes, OPP. Would appreciate if you can find out if other boards such as Asus or s754 have same issues if you happen to have them handy :P Want to know if it is mobo issue, chipset issue, or bios issue, or something.

[Nube]

link (http://www.pc-memory-upgrade.co.uk/memory/ddr-ram-pc3200.asp)
Stock:
Out (no ETA)

[Tony]

I have been able to run 275+ 166 mode on the epox 9NDA3+ and had no issues 133 mode either...i know some boards may have some issues but it may not be the ram here but I will look into it for you.

TBH you guys are going to want to run 1:1 as it is the fastest mode, don't go off buffered sandra use the unbuffered and use everest to compare..you will see what i mean.

[bachus_anonym]

I have been able to run 275+ 166 mode on the epox 9NDA3+ and had no issues 133 mode either...i know some boards may have some issues but it may not be the ram here but I will look into it for you.

TBH you guys are going to want to run 1:1 as it is the fastest mode, don't go off buffered sandra use the unbuffered and use everest to compare..you will see what i mean.
but u see, the problem is that with divider issue this memory is useless for those that wish to run it in phase change systems... 256x10=2560MHz is OK with div 200 but e.g 285x9=2565MHz or 285x10=2850MHz with div 166 will be impossible.... :(

EDIT: have u tried e.g. 280HTT with div 166 ??? i can POST up to 279HTT but only 270HTT makes it's way all the way to Windows. but unfortunately it's totally unstable there :(

[Rabbi_NZ]

Mine will POST and go into windows at 275 on 200MHz (1:1) divider... will try divider and >280HTT tonight

[OPPAINTER]

Well sure enough, I just tryed the 166 divider on MSI Neo2. The system was all FUBAR after 270HTT:D.
I don't have other boards to mess with at the moment. Waiting on some NF-4 and ATI 480 boards next week.
I can see how some people with locked multis won't appreaciate the memory.
I myself am not worried about a divider issue for my rigs. 1:1 is my goal.

OPP

[bachus_anonym]

Well sure enough, I just tryed the 166 divider on MSI Neo2. The system was all FUBAR after 270HTT:D.
I don't have other boards to mess with at the moment. Waiting on some NF-4 and ATI 480 boards next week.
I can see how some people with locked multis won't appreaciate the memory.
I myself am not worried about a divider issue for my rigs. 1:1 is my goal.

OPP
so in other words, if one's going for OCZ VX on phase change beyond 2900Mhz - better one gets himself FX.... :D hmmm.... that FX just got few places up on my "to-upgrade" list :)

[OPPAINTER]

FX is the only way to Fly:D

OPP

[STEvil]

VX for FX .. kinda has a rhyme to it.. :D

[jinu117]

If I get really nice batch of VX for this RMA... than maybe. 270 2-2-2-8 dual channel will be nice enough to consider it's fault (or mobo's fault) with divider and move on to FX. We are entering the first batch of 875p era kind of problem. (It was chaotic era with everyone scrambling to figure out what works and what not)

[formyfaith]

above 275mhz with the 166 divider i can't boot either.

also is my friend, too. guess everyone with the neo2 board may have the

same problem.

[Tony]

above 275mhz with the 166 divider i can't boot either.

also is my friend, too. guess everyone with the neo2 board may have the

same problem.
Im going to point MSI at this thread and see if they can help.

[xgman]

why are you guys messing with dividers? doesn't that sort of defeat the purpose?

[eva2000]

why are you guys messing with dividers? doesn't that sort of defeat the purpose?
yeah high 1:1 is best for AMD64

[LikwidKool]

We also have a growing team of forum whores who are here to look after you guys and answer all your questions...;)

Hey I'm no whore!! :smileysex :devil:

For real though, OCZ is the most devoted company we have for us xtreme enthusiasts. Just be patient and they will get us more VX soon. I know none of us like to wait, but would we rather have un tested IC's on the VX that'll only do pc3200?

[LilGator]

eva, there is no such thing as 1:1 on A64, memory runs based on a CPU divider, not HTT (FSB).

Dividers such as 166, 183, etc... just change the divider that is used based on the CPU speed.

Many times you can optimize your CPU speed, by using a divider, effectively increasing your multiplier (if you imagined the multiplier to be RAM speed x Multi = CPU speed)

Of course if you have an FX chip, then there is no reason to play with dividers :) But for those stuck with a 9x multi, dividers can be a huge improvement...

BTW, for those looking to buy VX, these guys look like they are in stock (ship worldwide as well) if you don't mind paying a little extra:

http://www.microstorm.com/OCZ_1GB_512MB_x_2_DDR_OCZ4001024ELDCGEVX_K/memory/partinfo-id-529218.html

[charlie]

it's out of stock, pretty much EVERYWHERE......

[DevilsRejection]

glad i jumped in on it when i did!

[jinu117]

Don't worry, unless I get 260+ dual channel 2-2-2-x, I will be selling my 4 pieces or something and look for BH-5 instead. Per price, it's 1:1 seems good but that doesn't take into account of divider unfortunately which is big turn off for me. (even if I get FX).

[gandalf027]

I've got my VX on a trial base for now, so i'm just waiting a bit to see if there is a solution to this divider problem. My 3000+ is running at 3000mhz 1.6v prime stable for hours .. i was hoping on the vx to use the 166 divider. Let's wait and see.

[Tony]

166 issue update.

We have programmed VX with a TCCD and EB spd and it made no difference...the modules still didn't like 166 mode.

We have a theory as to why but its best we work with bios engineers to see if we can resolve the issue.

So for now we are advising VX is best in 1:1 mode, some will do ok in 166 mode but your not going to see 300+HTT.

I have started asking if bios files can help, as soon as i know more i will post.

[bachus_anonym]

166 issue update.

We have programmed VX with a TCCD and EB spd's and it made no difference...the modules still didn't like 166 mode.

We have a theory as to why but its best we work with bios engineers to see if we can resolve the issue.

So for now we are advising VX is best in 1:1 mode, some will do ok in 166 mode but your not going to see 300+HTT.

I have started asking if bios files can help, as soon as i know more i will post.
so it definetely is some sort of VX compatibility issue, right? hope you guys can fix this with co-op with BIOS engineers.

on a side note - ATM, after being able to do DDR512 6-2-2-2.0 1T i'm not even able to memtest @ DDR400 :( checked in NF2 mobo and it runs fine with no memtest error. put back some BH-5 in my Neo2 and it works fine too..... i thiunk i exhausted all ideas i had :(

update: and for this situation to look even "better" - i already pulled off heatspreaders :(

[Tony]

Guys pulling off heatspreaders is always a bad move...with our ddr2 it can pull off IC's also :eek:

bachus_anonym you saying your ram is ok in NF2 but not your MSI?

[gandalf027]

166 issue update.

We have programmed VX with a TCCD and EB spd and it made no difference...the modules still didn't like 166 mode.

We have a theory as to why but its best we work with bios engineers to see if we can resolve the issue.

So for now we are advising VX is best in 1:1 mode, some will do ok in 166 mode but your not going to see 300+HTT.

I have started asking if bios files can help, as soon as i know more i will post.

Thanks m8 on the update. We will be praying for some kind of solution. :)

[eva2000]

Guys pulling off heatspreaders is always a bad move...with our ddr2 it can pull off IC's also :eek:

bachus_anonym you saying your ram is ok in NF2 but not your MSI?

tiny BGA have weaker joints to pcb so yup pulling off heatspreader will take a few things off the pcb hehe

[conrad.maranan]

I gained 2MHz (http://xtremesystems.org/forums/showthread.php?p=628155#post628155) by pulling off my heatspreaders. I'm pretty sure I was the first one here to do this. :D If voiding your lifetime warranty for 2MHz sounds good to you, go for it - just be careful not to accidentally yank anything else off.

[charlie]

I gained 2MHz (http://xtremesystems.org/forums/showthread.php?p=628155#post628155) by pulling off my heatspreaders. I'm pretty sure I was the first one here to do this. :D If voiding your lifetime warranty for 2MHz sounds good to you, go for it - just be careful not to accidentally yank anything else off.

Which brings up a good point....
"OCZ, if you guys spend so much time and expense binning and testing chips, then stick heatsinks on them and reduce their performance?????

Why not just include heatsinks in the package with frag tape?? User option, you know?

[conrad.maranan]

Why not just include heatsinks in the package with frag tape?? User option, you know?
I think this would be a good idea. For sure, case modders love the bling effect of gold and silver heatspreaders. But enthusiasts who care more about the bottom-line performance of a component can definitely appreciate a naked stick of RAM. Off the top of my head, the G.Skill modules serve as a fine example.

[Playful_Buffalo]

heh, reminds me of the thread in which the guy watercooled his ram

EDIT: here you go (http://www.xtremesystems.org/forums/showthread.php?t=38945&page=1&pp=25)

[jinu117]

PCS is a make...a manufacturer.

Speed binning of VX ensure every module does at least 240fsb 2-2-2.

Well, you sure about that? 6 sticks tested so far (2 from RMA) and 1 didn't even hit stock speed 2-2-2-x on 3.5v, now I am just toying with one of RMA'd stick that is keep going down to what seems to be stock speed from 240 below. (@ 215 testing right as I write this....)
No, it isn't my machine as the other 4 sticks did just fine above 240. If not higher (to 267).
I plan to call and RMA again. However, the cost of shipping this memory back and forth are really adding up. So far, amount of money I had to spend on OCZ RMA is adding up close to $30 with memories, etc.

[bachus_anonym]

Guys pulling off heatspreaders is always a bad move...with our ddr2 it can pull off IC's also :eek:

bachus_anonym you saying your ram is ok in NF2 but not your MSI?

yes sir.... VX is in my day-to-day rig, NF7-S v2, running at DDR433. i've tried memtest on them and it runs just fine. when i had it in MSI, i get errors right after test #1 :confused: and 99% of them at the same address 308.5MB...
so for now i put my 2x256MB BH-5 in Neo2 and it runs absolutely fine @ DDR520 (no errors in memtest).
i now recall that this happened BEFORE i removed heatspreaders but AFTER i started playing with memory dividers.... i think i need to look into little bit closer coz it looks like RAM is fine....

EDIT: i just want to repeat that just few days ago i was error-free @ DDR512 and benchmark-stable up to DDR524....

[conrad.maranan]

Looks like the same thing that happened to trans am is also happening to you. In the end, I believe he was only able to run at stock speeds.

[arctic-k20]

*interested onlooker - now concerned with his recent order of vx*

[Funny_S]

well it seems that I'm one of the lucky owners of this VX memory, got it today together with the DDR Booster, I'll be picking up my ASUS A8N-SLI later this day since they held one back for me :toast:

Now my question is has anyone yet tried the VX together with the Booster on the A8N-SLI? I've saved quit a while for this beast and don't want to fry anything (well at least not in the first month :p: )

[DevilsRejection]

hmm i wouldn't try it on a board that will be hard to RMA, might want to wait and see what andy says, he trys the booster on a variety of boards.

[trans am]

but u see, the problem is that with divider issue this memory is useless for those that wish to run it in phase change systems... 256x10=2560MHz is OK with div 200 but e.g 285x9=2565MHz or 285x10=2850MHz with div 166 will be impossible.... :(

EDIT: have u tried e.g. 280HTT with div 166 ??? i can POST up to 279HTT but only 270HTT makes it's way all the way to Windows. but unfortunately it's totally unstable there :(

I think VX is great for FX where you can use high multi and run VX at it's highest stable ceiling, or a single channel setup. But I think if you ran a divider on the VX, you are better off getting some good tccd and running that at high speeds. I've had 2 sets of VX with 3 different cpus and 3 boards. they both max around 240-245. I think opp got really lucky and this is not normal for VX> don't get your hopes up and think you are going to hit 260mhz out of the box with these. IMO you will be extremely lucky to hit 250mhz stable. I'm not trying to stomp anyones buzz, I'm just being realistic and don't want people to commit suicide because they didn't get their VX past 260mhz. As for the divider issue, I coun't get any divider to work on the neo2 no matter what speed, on any bios except the 100 divider out of the bios., I could change dividers 166/133 etc with A64 tweaker and it worked, but would fail to load os when dividers set in the bios. My heatspreaders came off the day I got mine. I think it's a good idea if you have a decent set . They run much much cooler with the spreaders off because air can actually pass between them even using dimm slots 1 and 2. when heatspreaders are on, they end up touching each other generating tons of useless heat. Take it all off.

[eva2000]

I think VX is great for FX where you can use high multi and run VX at it's highest stable ceiling, or a single channel setup. But I think if you ran a divider on the VX, you are better off getting some good tccd and running that at high speeds. I've had 2 sets of VX with 3 different cpus and 3 boards. they both max around 240-245. I think opp got really lucky and this is not normal for VX> don't get your hopes up and think you are going to hit 260mhz out of the box with these. IMO you will be extremely lucky to hit 250mhz stable. I'm not trying to stomp anyones buzz, I'm just being realistic and don't want people to commit suicide because they didn't get their VX past 260mhz. As for the divider issue, I coun't get any divider to work on the neo2 no matter what speed, on any bios except the 100 divider out of the bios., I could change dividers 166/133 etc with A64 tweaker and it worked, but would fail to load os when dividers set in the bios.
have you burnt in the VX yet ?

[trans am]

have you burnt in the VX yet ?

Burn in?
as in countless hours of test 5 and hot cpu tester? Yes, been there done that. lol

[jinu117]

Burn in seems real for me... by about 3-6mhz on each stick except the one that is doing 267-2-2-2-8. Wonder if it has to do with heatspreader settling down or something. BTW, whoever keeps saying this memory runs cool even at high volt is smoking something or running it @ stock speed with high volt. It seems like temperature escalates rather quickly when you have over 3.3+v and actually stressing memory. I've seen so far up to 50c on pcb section of memory. with 2 32cfm 80mm fan blowing directly above that area.

PS) I am getting another set sent out from OCZ today. By thursday hopefully I will have some magic clocker. One thing I found out, they pretest memory without heatspreader. Don't want to invalidate warranty yet but when my stick becomes old it might be worth the fun to take out.

[trans am]

Burn in seems real for me... by about 3-6mhz on each stick except the one that is doing 267-2-2-2-8. Wonder if it has to do with heatspreader settling down or something. BTW, whoever keeps saying this memory runs cool even at high volt is smoking something or running it @ stock speed with high volt. It seems like temperature escalates rather quickly when you have over 3.3+v and actually stressing memory. I've seen so far up to 50c on pcb section of memory. with 2 32cfm 80mm fan blowing directly above that area.


I'm not smoking anything nor am I running it at stock. my spreaders are off and I have 90mm panaflo blowing on ddr booster and dimms. Of course memory is not cool, it's not scalding my fingers when I touch it either. THe heatspreaders were really hot before I removed them, even with the panaflo.

[formyfaith]

Im going to point MSI at this thread and see if they can help.


If the divider issue gets solved with the neo 2 board,

I think VX would start to rock~! :)

I hope msi gets this problem fixed really fast~ :D

thanks a lot bigtoe :toast:

[jinu117]

I'm not smoking anything nor am I running it at stock. my spreaders are off and I have 90mm panaflo blowing on ddr booster and dimms. Of course memory is not cool, it's not scalding my fingers when I touch it either. THe heatspreaders were really hot before I removed them, even with the panaflo.

I wouldn't dare accusing you of smoking something... maybe sniffing cat stuff lol :)
Got any temperature measure probe to check the temp real time while it is running? I really think this heatspreaders are nothing more than heat trappers.

[conrad.maranan]

I really think this heatspreaders are nothing more than heat trappers.
Truth.

[andyOCZ]

Burn in seems real for me... by about 3-6mhz on each stick except the one that is doing 267-2-2-2-8. Wonder if it has to do with heatspreader settling down or something. BTW, whoever keeps saying this memory runs cool even at high volt is smoking something or running it @ stock speed with high volt. It seems like temperature escalates rather quickly when you have over 3.3+v and actually stressing memory. I've seen so far up to 50c on pcb section of memory. with 2 32cfm 80mm fan blowing directly above that area.

PS) I am getting another set sent out from OCZ today. By thursday hopefully I will have some magic clocker. One thing I found out, they pretest memory without heatspreader. Don't want to invalidate warranty yet but when my stick becomes old it might be worth the fun to take out.

I tested at 3.5v 260mhz with VX with some good ventilation in the form of a strong 80mm fan and using my digital thermometer if found nothing higher that 32C. I also probed the actual IC's under the heatspreader (from the side of course) and found nothing higher than 32c.

I suggest that either your voltages are off (high) or you need better ventilation. :)

[mcnbns]

Would it not be cheaper to produce VX without heat spreaders? I know some people think RAM cannot be high-performing without heat spreaders, but those kind of n00bs would not be into buying RAM that needs 3.3V to perform. You could call this new heat spreaderless RAM OCZ VX Lite or something. I'd buy some.

[OPPAINTER]

All my testing was done inside the case with no fan blowing on them. I can't wait to get these on the Test Bench try some cooling, might get me some extra hertz:D

OPP

[lalPOOO]

Would it not be cheaper to produce VX without heat spreaders? I know some people think RAM cannot be high-performing without heat spreaders, but those kind of n00bs would not be into buying RAM that needs 3.3V to perform. You could call this new heat spreaderless RAM OCZ VX Lite or something. I'd buy some.

I think one of the reasons they use heatspreaders is to help prevent ESD, otherwise they probably would be making them without heatspreaders. Just a guess though, I may be wrong.

[trans am]

i think it's more of a "BLING BLING" factor. It's a status thing lol. you gotta have 'Bling' in order to have street cred. these days.

[conrad.maranan]

It's alright if my system is lacking in the rave lighting and hip hop shine department. I can always throw in a couple of pairs of the illest Air Max 87s and 90s to make my rig really stand out in a picture. Better yet, I can commission some skanks on MySpace to pose with my rig for added effect. :D

[trans am]

It's alright if my system is lacking in the rave lighting and hip hop shine department. I can always throw in a couple of pairs of the illest Air Max 87s and 90s to make my rig really stand out in a picture. Better yet, I can commission some skanks on MySpace to pose with my rig for added effect. :D

LOL conrad, I was thinking about getting some chrome 20" rims for my case. It's so heavy with all this cooling, I need to be able to roll it around. Maybe adding a few 12" subs and a few LCDs in the near future.....Please MTV...Pimp my ride!

[Rabbi_NZ]

Im going for the ghetto thug angle... lay down a GAT & AK47 beside my system and paint 2Pac on the case window... now that's gangsta! hahaha

Conrad, you realise those modules you have are 10x better than mine... either that or your sis has a CPU with one beaut of a mem controller... granted I can't give my modules more than ~3.35vDimm but they max at bout 261 memtest stable. So either your 3.5v burnin really did wonder, you jus got brilliant sticks, or that CPU you got there needs to be sent this way and fast! :)

[trans am]

I think COnrad had so much bad luck in the recent past with hardware that the overclock gods were finally merciful. He simply got some amazing VX.

[ryanpgroovy]

it could be the burn in , I mean some of the results people get after burning in are amazing

[Rabbi_NZ]

Those that have tried, do you really see much of a gain between 3.3v and 3.5v?

I ask because I can't feed my modules more than ~3.35v and that drops to near 3.3v under load, 3.3vRail is too low.
So, I'm wondering if upping my 3.3vRail is going to net me any benefits? I would jus go do it but voiding warrantee on a PC P&C is not something I'm dying to do.

[cantankerous]

I hope it does. With 3.3v I can't do anything over 243 let alone 261. More volts better bring me more mhz or these things are out the door. my BH-6 is doing better right now.

[Tony]

Well, you sure about that? 6 sticks tested so far (2 from RMA) and 1 didn't even hit stock speed 2-2-2-x on 3.5v, now I am just toying with one of RMA'd stick that is keep going down to what seems to be stock speed from 240 below. (@ 215 testing right as I write this....)
No, it isn't my machine as the other 4 sticks did just fine above 240. If not higher (to 267).
I plan to call and RMA again. However, the cost of shipping this memory back and forth are really adding up. So far, amount of money I had to spend on OCZ RMA is adding up close to $30 with memories, etc.

We have tested every module that has come back to us and many that were unstable on customers rigs are fine on the test rigs we use.We still swap out though to brand new modules.

I have a feeling there is more to the reasons for the lack of stability than meets the eye here...maybe its our boards or CPU's...its just hard to tell.

On a side note guys we test to 240 or so 2-2-2 but going by results seen on this thread overall results can be VERY different. I said this before and i feel i must repeat that this is NOT DDR500 ram, we don't test it that high as there are just to many factors involved in getting a solid guarantee at that speed.

[OPPAINTER]

Those that have tried, do you really see much of a gain between 3.3v and 3.5v?

There is deffinitly a difference. Better yet, between 3.3V and 3.7V, bigger difference. :D

OPP

[bachus_anonym]

there are some oddities with this mems.... ATM i'm memtesting at DDR460 again but i had to lower vdimm quite significantly to be able to do so... few days ago i posted a screenie of 10hrs of memtest @ DDR512 with 3.51v.... today at the same volts even DDR400 is not stable... i lowered vdimm to 3.20v and i'm memtesting @ DDR430 again :confused:
i don't think i will be able to hit that DDR512 again, though :(

[Rabbi_NZ]

There is deffinitly a difference. Better yet, between 3.3V and 3.7V, bigger difference. :D

OPP
lol
What sort of diff between 3.3v & 3.5v in terms of MHz?

[OPPAINTER]

lol
What sort of diff between 3.3v & 3.5v in terms of MHz?
Can't tell off the right off the bat. I'd have to test it, but I know there's a gain.

OPP

[andyOCZ]

I hope it does. With 3.3v I can't do anything over 243 let alone 261. More volts better bring me more mhz or these things are out the door. my BH-6 is doing better right now.

On your NF2 board. ;)