The Official OCZ VX Thread - Strictly Test Data Only!!!

[DevilsRejection]

Same here, everything stable but I have had zero success running P95 on more than one A64.

did you try stock speeds and it still fails?

[Joe Camel]

as per above, ive tested the 2 sticks 1 @ a time...

got 1 stick ("good")to:
255x10 (cpu:1.6V) 1:1 222,10,12,14 @3.4V
the other stick ("bad")to:
250x10 (cpu:1.6V) 1:1 222,10,12,14 @3.4V

Benches with ONLY "bad" stick @250:

sandra MEM.......Sandra CPU........01...........03.........Aqua......S.Pie (1M)
3854/3854.........11,531............26,698.....14,179.. ...76,520.....35 sec

BOTH sticks (DUAL) @250:
sandra MEM.......Sandra CPU........01...........03.........Aqua......S.Pie (1M)
7002/6917.........11,532............27,361.....14,205.. ...78,081.....33 sec


BUT!!! single sick tests were P95 stable (30min)...DUAL ISNT...yet

2xSticks (bad=slot3, good=slot4, booster=slot1) <2 sec P95
2xSticks (bad=slot4, good=slot3, booster=slot1) >7 MIN P95...

now upping V for CPU and Mem...

(tests above: CPU: 1.6...MEM: 3.4...GPU @ 456/1216 )
CPU: 1.65V MEM: 3.4... >10min P95 (will update)
CPU: 1.675..MEM: 3.4....>13min
CPU: 1.7.....MEM: 3.4....>2min
CPU: 1.675..MEM: 3.45...>11min
CPU: 1.675..MEM: 3.5.....>6min...(huh... )
CPU: 1.675..MEM: 3.6.....<1SEC...mobo problems (readings go 0.0 >3.5V BIOS/MBM)
.........................3.3?...>6min closed by windows

NEW (lower) OC: CPU 245x10, MEM: (see above)

CPU: 1.675..MEM: 3.4....>9min...:grr: P95 & Winnie mem controllers

240x10
.......1.60............3.4....>23min
.......1.625..........3.4.....>15min...see, this is where i start loosen it!

going back to my GSkill TCCD (@3.3v 233,10,12,14) to compare...
250x10 (1.6vCPU) for 33min...error...
250x10 (1.65V) 1hour 20min and still going...(sorry OCZ, this mobo likes TCCD better than VX) went for 1H 28min...errored...
250x10 (1.675v) 1H 35min...error...

whelp, thats 11 hours of testing...

[conrad.maranan]

I said I wasn't going to overclock this Winchester that I "borrowed," but I was tempted just to push it a little bit. I won't have a DDR Booster to play with until I receive my new unit, so I decided to see how much I can push these VX modules using 2.85VDIMM and 2-3-3-10 timings. I'll be burning-in at this setting until I have some new parts to play with.

Here's a few shots with default A64 Tweaker settings:





Here's a few more good at the same frequency and timings + A64 Tweaker enhancements:





235MHz @ 2-3-3-10 using 2.85VDIMM in dual channel mode...not too shabby.

[aCidbAbY]

so far i have found my 250mhz 1:1 clock to be totally stable.
going to try higher tonight.

i think that 250mhz is good for me i want to try to tweak the settings for a higher pi score though. my cpu loves tight timeings


edited:::: 253.3 update final pic

[trans am]

Guys, I have new news to report.... My set is bad!! one at any speed won't boot wixxp. THe other boots fine at 2-2-2-10 1t @256*10 memtest stable! The strange thing is that I can run both sticks together at 240mhz mem stable, but when I swapped them to run alone to see which one goes faster, one of the sticks wouldn't even let me post no matter what volts and timings. Strange huh? I tried different dimm slots too. Also I am unable to use any ram dividers in the bios 1.36-1.51 (neo2) with this ram. Is this possible? I can change divider with clockgen fine. THe ram is very cool and I am using 1 92mm fan. I plan to use my el4200 until ocz can send me a replacement.

And for those who want to blame it on my poor winnie memcontroller, think again....
These are with el4200 platinum 2x512 in dual channel......

http://xtremesystems.org/forums/atta...id=20567&stc=1
http://xtremesystems.org/forums/show...408#post623408

[DevilsRejection]

when/where u order from?

[conrad.maranan]

OCZ Representatives
The gentlemen in your group have had the pleasure of playing around with the VX since early November. Out of curiosity, what is the highest frequency you are able to run stable in dual channel at 2-2-2 timings and 2.85VDIMM (no DDR Booster)?

Basically, what I'm trying to find out is whether the VX can run with lesser VDIMM over time (following a lengthy burn-in period). Please post your results at your earliest convenience.

Thanks!

EDIT:
Do these things really get better with time? I suppose only time will tell. Here's my current maximum with 2.85VDIMM. Prime95 Torture Test ran for an hour, but I'll run it again overnight just to be sure. The interesting part of all of this is the fact that I couldn't even pass 20 loops of Memtest last night at 220MHz with these timings and voltage.


Just keep on truckin', ladies.

[SteveOCZ]

I'll give it a try later today

[Joe Camel]

will be getting/testing a MSI Neo2 (nForce3) mobo later this week...maybe its the mobo?

[trans am]

did anyone have issues running dividers with this ram on the neo 2 in the bios? I've used 1.36-1.51 it won't boot into os.

[andyOCZ]

Hi guys,

I see you are all hard at work.

I have some findings of my own. As a control setup I put some TCCD at 2.5-3-2-10-11-13 into a Gigabyte K8NS Ultra-939 running a Winnie 3500+ (week 37). FSB was at 201mhz. Prime95 Blend failed in 5hrs 34 minutes.

I have an FX53 and VX coming from OCZ Sunnyvale to my office. I will throw the FX53 into the Gigabyte and use the same settings to see if P95 will run. I suspect that there is a problem with the memory controller on the Winnie that is causing "instabilty". Of course everything else runs just fine and appears perfectly stable.

I am not trying to discount the value of P95, but we need to find out if the 90nm A64 has a flawed memory controller.

I will update you ASAP.

[andyOCZ]

Ah yes, I picked up an MSI Neo2 939 yesterday that I will be winding up to test also. I will post results on that platform as well.

[cuddles]

Andy, you could always just send your VX sample to me you know, my Neo2 and FX-53 would love to do your job

[trans am]

Ah yes, I picked up an MSI Neo2 939 yesterday that I will be winding up to test also. I will post results on that platform as well.

Andy try to run dividers in the bios for me with the vx and see if you have problems. I have a week 40 taiwan nf3 ultra. I definitely think the memcontroller is flawed on the winnie. Either that or there is a compatiblilty bug with Prime95 and the instructions from memcontroller. I can run Prime95 for a hours, but like you too, it will eventually get a fatal error. And I could loop 3dmark2001se for days and run everything like play heavy 3d games like hl2 or ut2004 without having problems. I like hotcpu tester, toast, and CPU Stability Test 6.0
http://www.majorgeeks.com/download215.html


Toast? http://www.majorgeeks.com/download867.html

prime95 might be failing after a long period of time because cpu is just getting too hot.

[conrad.maranan]

230MHz with a wimpy 2.85VDIMM was gold with Prime95 Torture Test running overnight. I just passed 20 loops of Memtest at 235MHz - same timings, same wimpy voltage.



Will it do 240MHz @ 2.85VDIMM? Stay tuned.

trans am
No problem with dividers here, even when my original Winchester and DDR Booster were still working. I have two boards: 0442A2 KOREA and 0443A2 KOREA.

andyOCZ

I am not trying to discount the value of P95, but we need to find out if the 90nm A64 has a flawed memory controller.

I have suspected that the Winchester has a weak memory controller since day one of ownership. Please prove me right.

[trans am]

This whole divider thing has me

[cuddles]

That wouldn't explain why I fail Prime, but fail nothing else at 2.64, would it? Or is this only a Winnie issue?

[xgman]

Man Conrad, I said they saved the best for last for a reason. I was the 1st to order and mine aren't doing jack. 241 2-2-2-10 is the best mine will do so far. I've been running 3.3v and burning in since friday. I'm really disappointed because everyones getting at least 250 from these and mine are stuck at 240 basically. I tried cas 2.5/3 and didn't help. I found a weird bug with my neo2. If I try a divider like 166 or 133 the system won't boot into windows. I should be able to run my cpu at 280*9 no problem, but when trying 280*9 166 divider there are immediate problems. WTF? I am out of ideas and I'm ready to send this back to ocz.

Why on earth would you want to run a divider at such low htt anyway?

Put the 4200plats back in and smile.

[trans am]

Because I was trying to max out my cpu with the sweet spot on this ram. I figured just try some low htt with a divider just to see if it would even work and no matter what, it doesn't. Only way was in clockgen. And I wanted to see how much the tighter timings could do over the el4200 platinum. My memcontroller won't let me run the 4200 past 280 mhz without getting errors. I tried ras/cas 4 and that didn't seem to help. so far it's 270*10 1:1 2.5-3-3-10 1T. Just dropped the VX off at fedex. ocz is going to replace mine. So honestly what do you think would be faster VX @ 250-260 2-2-2-10? or el4200 @270 2.5-3-3-10?
....I think we both know the answer.

[xgman]

After reading through most of the posts on the VX in this and other threads, it is pretty clear to me that for the most part, the ram is doing exactly what Ryna and Toney etc have said it would do, up to 260 at tight timings, given a few exceptions which may have many different reasons behind them as is with all ram we see. I think some folks expected to pop these in and be able to run 280 2-2-2.

This is great if you run at 260 or less 24/7 and want tightest timings and have adequate voltage access or a booster. For those others that compensate by running 260-330 htt sync, there is always the plat tccd based series that has proven that at 2.5-3-3-7-1T you can get close and in many cases over 300htt and at moderate voltage, which more than, makes up for 2-2-2 up to 260.

[xgman]

Because I was trying to max out my cpu with the sweet spot on this ram. I figured just try some low htt with a divider just to see if it would even work and no matter what, it doesn't. Only way was in clockgen. And I wanted to see how much the tighter timings could do over the el4200 platinum. My memcontroller won't let me run the 4200 past 280 mhz without getting errors. I tried ras/cas 4 and that didn't seem to help. so far it's 270*10 1:1 2.5-3-3-10 1T. Just dropped the VX off at fedex. ocz is going to replace mine.

Seems to me that 279 2.5-3-3-7-1T is more than enough to make up for the faster timings at much lower htt.

[xgman]

So honestly what do you think would be faster VX @ 250-260 2-2-2-10? or el4200 @270 2.5-3-3-10?
....I think we both know the answer.

Actually someone with both that can run the VX at at 250-260 2-2-2 should do a test to see just what the differences would be if they are at the same cpu mhz.

[Tony]

Guys I knew this would happen, you are all expecting way to much here and xgman is bang on the money with his statement.

We rate the ram at 2-2-2 PC3200 with 3V or so but we don't say every module is going to run 260fsb as they won't. I said a good portion are good to 250fsb and some higher but a lot depends on CPU memory controllers, PSU's, Boards, etc etc and ultimately the dimms themselves.

Now we will do our best to look after you guys but a wholesale cherry pick is not on the cards here. Ryan has quoted the modules are tested to 240fsb 2-2-2, this in its self is an awesome speed and we are trying to keep priced low here, any additional testing above this speed will add to the final cost of the modules, and we want everyone to love the price/performance on these dimms.

[enzoR]

yea you cant RMA a pc3200 stick cuz it wont run 240mhz... sell it and get a new seperate one. their is nothing defective about it.

[trans am]

yea you cant RMA a pc3200 stick cuz it wont run 240mhz... sell it and get a new seperate one. their is nothing defective about it.

If you are reffering to me, my reason for the rma is because the system won't even post with one of the sticks (at any speed). There is no foul play here, I just want some ram that works.