MEMORY Voltage Scaling, please contribute... [Archive]

View Full Version : MEMORY Voltage Scaling, please contribute...

Pages : [1] 2

Rabbi_NZ

01-18-2005, 01:29 PM

This thread is dedicated to collecting UTT/VX/NEW-BH5 related data... especially how they scale with voltage.

To contribute please do the following:
* Set timings to 1T-2-2-2 (I run 1T-2-2-2-8-7-15, but that is just for reference)
* Find the max MEMTEST STABLE MHz at 3.1v, 3.2v, 3.3v, 3.4v, 3.5v, 3.6v.
* MEMTEST STABLE (for this thread) = 25+ loops of test #5
* Where possible... also give chip dates & PCB revision.

*********************VX/UTT Memory-Timing Tips*********************
* i865/i875 TWEAKER can be found here (http://www.cpuid.org/download/Tw865.zip)
* A64 TWEAKER v0.31 can be found here (http://home.exetel.com.au/codered/A64Tweaker/A64Tweaker_V031.zip)
* nForce2 Tweaker 0.25b can be found here (http://forum.abit-usa.com/attachment.php?s=&postid=379291)

DFI nF2 Infinity , Mobile 2600+ , 2x512 TwinMOS UTT 50D (uwackme)
http://www.xtremesystems.org/forums/attachment.php?attachmentid=29455&stc=1

DFI nF3 754 LanPartyUT , Mobile 3400+ , 2x512 OCZ VX (rabbi)
http://www.xtremesystems.org/forums/attachment.php?attachmentid=29452&stc=1

DFI nF3 754 LanPartyUT , 3000+ , 2x512 OCZ VX (aCidbAbY)
http://www.xtremesystems.org/forums/attachment.php?attachmentid=29453&stc=1

DFI nF4 939, 3500+ Winny , 2x512 TwinMOS UTT 04514/05044 (esdee)
http://www.xtremesystems.org/forums/attachment.php?attachmentid=29454&stc=1

Abit i865 AS8, 560J? , 2x512 TwinMOS UTT (Hipro5)

Best is - as for ABIT based mobos - and IF someones mobo/Rams can "handle" that, to put : 2 - 5 - 2 - 2 , 32min , StreetRacer , Enchanced , 4 , Enable , Enable as for 1:1 and 2 - 5 - 2 - 2 , 32min , Auto , 5 , Enable , Enable as for 5:4......and as for 3:2 , 2 - 5 - 2 - 2 , Auto , Encanced or Normal , 6 or 7 , Enable , Enabe or Disable....... ;)

************************************************** *********

********************AMD Socket 462 Results********************

enok (2x512, TwinMOS UTT 50D 0451 KO-6633, Abit nF2 AN7):
3.1v = 240 MHz
3.2v = N/A
3.3v = N/A
3.4v = N/A
3.5v = N/A
3.6v = N/A

TEDY (2x512, TwinMOS UTT 0503 44D KO6633, DFI nF2 Infinity)
3.1v = 235 MHz
3.2v = 240 MHz
3.3v = 245 MHz
3.4v = N/A
3.5v = N/A
3.6v = N/A

mrlobber(1x512, OCZ VX, DFI nF2 Infinity)
3.1v = N/A
3.2v = 243 MHz
3.3v = 247 MHz
3.4v = N/A
3.5v = N/A
3.6v = N/A

godsa(1x512, TwinMOS UTT 50D 0503 KO-6633, DFI nF2 LanPartyB):
3.1v = 254 MHz
3.2v = 257 MHz
3.3v = 261 MHz
3.4v = N/A
3.5v = N/A
3.6v = N/A

EMU(2x512, TwinMOS UTT 50D, DFI nF2 Infinity):
3.0v = 198 MHz
3.1v = 217 MHz
3.2v = 228 MHz
3.3v = 243 MHz
3.4v = N/A
3.5v = N/A
3.6v = N/A

Fr3ak(2x512, TwinMOS SP UTT, Shuttle nF2 SN45G):
3.1v = N/A
3.2v = 245 MHz
3.3v = N/A
3.4v = N/A
3.5v = N/A
3.6v = N/A

uwackme (1x512, TwinMOS UTT 50D, DFI nF2 Infinity):
3.1v = 240 MHz
3.2v = 245 MHz
3.3v = 245 MHz
3.4v = 245 MHz
3.5v = N/A
3.6v = N/A

uwackme (2x512, TwinMOS SP UTT 50D, DFI nF2 Infinity):
3.1v = 215 MHz
3.2v = 230 MHz
3.3v = N/A
3.4v = N/A
3.5v = 250
3.6v = N/A

3.1v = N/A
3.2v = N/A
3.3v = N/A
3.4v = N/A
3.5v = 245
3.6v = N/A

************************************************** *********

********************AMD Socket 754 Results********************

Rabbi (2x512, OCZ VX, DFI nF3 754):
3.1v = 242 MHz
3.2v = 259 MHz
3.3v = 263 MHz
3.4v = N/A
3.5v = N/A
3.6v = N/A

HermS (1x256, TwinMOS UTT, DFI nF3 754):
3.1v = N/A
3.2v = N/A
3.3v = 254 MHz
3.4v = N/A
3.5v = 264 MHz
3.6v = N/A

Jupiler (1x512 & 2x512, TwinMOS UTT 44D, DFI nF3 754):
3.1v = 242 MHz @ (1x512) , 230+ MHz @ (2x512)
3.2v = 250 MHz @ (1x512)
3.3v = 253 MHz @ (1x512) , 246 MHz @ (2x512)
3.4v = 255 MHz @ (1x512)
3.5v = N/A
3.6v = N/A

aCidbAbY (2x512, OCZ VX, DFI nF3 754)
3.1v = N/A
3.2v = 250 MHz
3.3v = 256 MHz
3.4v = N/A
3.5v = N/A
3.6v = 260 MHz

3.1v = N/A
3.2v = 255 MHz
3.3v = 257 MHz
3.4v = 260 MHz
3.5v = N/A
3.6v = N/A

v3n (2x512, OCZ VX, DFI nF3 754)
3.1v = N/A
3.2v = N/A
3.3v = 250 MHz
3.4v = N/A
3.5v = N/A
3.6v = N/A

************************************************** *********

********************AMD Socket 939 Results********************

conrad.maranan (2x512, OCZ VX, MSI Neo2 nF3 939):
2.8v = 220MHz
2.9v = 220 MHz
3.0v = 225 MHz
3.1v = 235 MHz
3.2v = 250 MHz
3.3v = 255 MHz
3.4v = 263 MHz
3.5v = N/A
3.6v = N/A

jikdoc (2x512, OCZ VX, MSI Neo2 nF3 939)
3.1v = N/A
3.2v = 250 MHz
3.3v = 253 MHz
3.4v = 257 MHz
3.5v = 258 MHz
3.6v = 260 MHz

Garrett(2x512, TwinMOS SP UTT 44D, MSI Neo2 nF3 939)
3.1v = N/A
3.2v = N/A
3.3v = 257 MHz
3.4v = N/A
3.5v = N/A
3.6v = N/A

MarilynMX(2x512, OCZ VX, MSI Neo2 nF3 939)
3.1v = N/A
3.2v = N/A
3.3v = ~240 MHz
3.4v = 249 MHz
3.5v = N/A
3.6v = 258 MHz
3.7v = 261 MHz
3.8v = 264 MHz
3.9v = 266 MHz

mrlobber(2x512, OCZ VX, DFI nF4 SLI-D)
3.1v = N/A
3.2v = N/A
3.3v = 258 MHz
3.4v = 262 MHz
3.5v = N/A
3.6v = N/A

bobmanfoo(2x512, OCZ VX, MSI Neo2 nF3 939)
3.1v = N/A
3.2v = 242 MHz
3.3v = 257 MHz
3.4v = 263 MHz
3.5v = 266 MHz
3.6v = 268 MHz

odb (2x512, OCZ VX, MSI Neo2 nF3 939):
3.1v = N/A
3.2v = 245 MHz
3.3v = 252 MHz
3.4v = 255 MHz
3.5v = N/A
3.6v = N/A

arj(2x512, OCZ VX, MSI Neo2 nF3 939):
3.1v = 209 MHz
3.2v = 222 MHz
3.3v = 242 MHz
3.4v = 259 MHz
3.5v = 261 MHz
3.6v = 263 MHz

Wingz(2x512, OCZ VX, MSI Neo2 nF3 939):
3.1v = 238 MHz
3.2v = 241 MHz
3.3v = 255 MHz
3.4v = 265 MHz
3.5v = N/A
3.6v = 275 MHz (1x512mb "good" stick only)

neuvas(2x512, TwinMOS SP UTT 44D ,MSI Neo2 nF3 939):
3.1v = N/A
3.2v = N/A
3.3v = N/A
3.4v = N/A
3.5v = 263 MHz
3.6v = N/A

esdee(2x512, TwinMOS UTT 04514/05044,DFI nF4 939):
3.1V = 255
3.2V = 259
3.3V = 263
3.4V = 267
3.5V = 270
3.6v = N/A

TEDY (1x512, TwinMOS NEW-BH5, DFI nF4 Ultra-D)
3.0v = 240 MHz
3.1v = 245 MHz
3.2v = 250 MHz
3.3v = 253 MHz
3.4v = 257 MHz
3.5v = 260 MHz

3.0v = 240 MHz
3.1v = 246 MHz
3.2v = 251 MHz
3.3v = 254 MHz
3.4v = 258 MHz
3.5v = 262 MHz
3.6v = 265 MHz

************************************************** **********

*************************Intel Results*************************

hipro5(1x512, TwinMOS UTT, Abit i865PE AS8):
3.1v = N/A
3.2V = 239 MHz
3.3V = 259 MHz
3.35 = 261 MHz
3.5V = 264 MHz
3.6V = 268 MHz

3.1v = N/A
3.2V = 217 MHz
3.3V = 239 MHz
3.35V = 252 MHz
3.5V = 255 MHz
3.6V = 257 MHz

bwan(2x512, TwinMOS SP UTT 44D 0503,Abit i865PE AS8):
3.1v = N/A
3.2v = 200 MHz
3.3v = 225 MHz
3.4v = 250 MHz
3.5v = 253 MHz
3.6v = 255 MHz

kindnuguz(2x512, OCZ VX,Abit i875 IC7):
3.1v = N/A
3.2v = 238 MHz
3.3v = N/A
3.4v = 250 MHz
3.5v = N/A
3.6v = N/A

************************************************** *********

AMD Socket 462 AVERAGES
3.1v = 237 MHz (5 Entries)
3.2v = 243 MHz (6 Entries)
3.3v = 248 MHz (5 Entries)
3.4v = 245 MHz (1 Entry)
3.5v = N/A
3.6v = N/A

AMD Socket 754 AVERAGES
3.1v = 242 MHz (2 Entries)
3.2v = 254 MHz (4 Entries)
3.3v = 256 MHz (6 Entries)
3.4v = 258 MHz (2 Entries)
3.5v = 264 MHz (1 Entries)
3.6v = 260 MHz (1 Entries)

AMD Socket 939 AVERAGES
3.1v = 234 MHz (4 Entries)
3.2v = 244 MHz (7 Entries)
3.3v = 255 MHz (9 Entries)
3.4v = 260 MHz (9 Entries)
3.5v = 265 MHz (5 Entries)
3.6v = 265 MHz (5 Entries)

Intel AVERAGES
3.1v = N/A
3.2v = 224 MHz (4 Entries)
3.3v = 241 MHz (3 Entries)
3.4v = 253 MHz (4 Entries)
3.5v = 257 MHz (3 Entries)
3.6v = 260 MHz (3 Entries)

***** uwackme's bh-5 burn-in guide thread at DFI-Street.COM ***** (http://www.dfi-street.com/forum/showthread.php?threadid=1860&highlight=uwackme+BH5+burn)

***** uwackme's bh-5 burn-in guide as Word document - hosted by NiCKE^***** (http://www.xtremesystems.org/forums/attachment.php?attachmentid=27628)

Cheers :toast:

TEDY

01-18-2005, 01:54 PM

oh you getting nice 260 with only 3.3v ? :)

that is great... CPC on ? 1T ?

Did anyone get any decent ocing with them on DFI NF2 ?

EnJoY

01-18-2005, 01:56 PM

Rabbi, with your 3.3v rail at it's current setting, with the 4.0vdimm bios you should be able to get 3.4-3.5volts. Why aren't you doing that?

Rabbi_NZ

01-18-2005, 02:03 PM

Rabbi, with your 3.3v rail at it's current setting, with the 4.0vdimm bios you should be able to get 3.4-3.5volts. Why aren't you doing that?
My 3.3vRail = 3.6v with DMM when not loaded... in Smart Gaudian software it is actually read at 3.47v-3.5v... and Im pretty sure the 3.3vRail needs to be at least 1.5v higher than the vDimm.
I tried running it at 3.4v and it was only at that level unloaded... in Sisoft sandra it actually dropped to 3.3v so my 3.3vRail needs to be cranked up to get anymore vDimm, and I dont want to mod this PSU.

Rabbi, with your 3.3v rail at it's current setting, with the 4.0vdimm bios you should be able to get 3.4-3.5volts. Why aren't you doing oh you getting nice 260 with only 3.3v ? :)

that is great... CPC on ? 1T ?

Did anyone get any decent ocing with them on DFI NF2 ?that?
1T 2-2-2-10-7-15 3.3v :)

EnJoY

01-18-2005, 02:23 PM

I had mine on my OCZ upt o 3.8, I suggest you try that and set it to 3.4volts. If that doesn't work, somethings wrong with your board.

Rabbi_NZ

01-18-2005, 02:31 PM

I had mine on my OCZ upt o 3.8, I suggest you try that and set it to 3.4volts. If that doesn't work, somethings wrong with your board.

My 3.3vRail = 3.6v with DMM when not loaded... ...my 3.3vRail needs to be cranked up to get anymore vDimm, and I dont want to mod this PSU.
...

STEvil

01-18-2005, 02:40 PM

try measuring the mosfet, mine was overvolting by .1v

You should try running CAS 1.5, too... and getting TRCD and TRP down to 1 (note: VX cant get TRCD-r and TRP down to 1 since it is not winbond chips, but UTT probably can because it is Winbond chips).

Rabbi_NZ

01-18-2005, 02:51 PM

STEvil,
1) I ran 10x270 ~3.3v 1.5-2-2-0-7-15 on SuperPI no worries... but the performance diff wasn't even noticeable in benchmarks compared with 2-2-2-10.

2) I will be shocked if TwinMOS UTT and OCZ VX dont use the same chips... what gives you the idea they dont dude? What else could VX possibly be?
People have taken spreaders off and actually seen the two dimples on their VX chips..

3) Measured at mosfet and under load in Sisoft Memory Bench it was bout ~3.33v which is very close to the loaded voltage when vDimm is set to 3.3v... i think my 3.3vRail simply needs more voltage.
Also, remember that reading comes before the big resistor bank and more traces so it's possible the value drops even further after that... in fact, it probably does. Just because it's overvolting there doesn't mean the memory is getting that much voltage.

MarilynMX

01-18-2005, 04:22 PM

I dont know how to define memtest stable since I'm running 258/259 at 3.6V first 50 loops no errors but avg 2~6 errors per 100 loops after #50. Also my memory controler limits the HTT up to 245 I think? I can't use office under 250x10 or 255x10 but rock solid on 240x11 wtf?

Rabbi_NZ

01-18-2005, 04:28 PM

Thankyou MarilynMX :)

Sounds like it may be a heat issue for you if the first 50 tests of #5 never give errors.
With 260 I can pass #5 all day long no problem with 3.3v but #7 gives a single error every 10 passes or so... I think it is a memory controller issue too to be honest. I can Prime all day at 260 aswell, and I can run SuperPI 32MB up to 268 2x512 so it's def wierd.

Would you mind testing you RAM from 3.1v to 3.6v please?
I would be very interested in your results.

STEvil

01-18-2005, 04:44 PM

STEvil,
1) I ran 10x270 ~3.3v 1.5-2-2-0-7-15 on SuperPI no worries... but the performance diff wasn't even noticeable in benchmarks compared with 2-2-2-10.
Thats why I suggested TRP 1 and TRCD-r 1 also.

2) I will be shocked if TwinMOS UTT and OCZ VX dont use the same chips... what gives you the idea they dont dude? What else could VX possibly be?
People have taken spreaders off and actually seen the two dimples on their VX chips..
Things that give me the idea UTT is not found on VX:
- VX cannot do TRP 1. Winbond can
- VX cannot do TRCD-r or TRCD-w of 1. Windbond can.
- VX can do CAS3, Winbond cannot (have not seen reliable example as of yet)
- Dimples come with the package type, not a feature of Winbond chips only.

3) Measured at mosfet and under load in Sisoft Memory Bench it was bout ~3.33v which is very close to the loaded voltage when vDimm is set to 3.3v... i think my 3.3vRail simply needs more voltage.
Also, remember that reading comes before the big resistor bank and more traces so it's possible the value drops even further after that... in fact, it probably does. Just because it's overvolting there doesn't mean the memory is getting that much voltage.

Measure at the DIMM slot then, mine was the same there as it was at the mosfet and there was no droop detectable by any of the three multimeters I own..

Rabbi_NZ

01-18-2005, 05:02 PM

Thats why I suggested TRP 1 and TRCD-r 1 also.

Things that give me the idea UTT is not found on VX:
- VX cannot do TRP 1. Winbond can RABBI: "will try tonight, are you sure UTT can?"
- VX cannot do TRCD-r or TRCD-w of 1. Windbond can. RABBI: "will try tonight, are you sure UTT can?"
- VX can do CAS3, Winbond cannot (have not seen reliable example as of yet) RABBI: "UTT can with no problems"
- Dimples come with the package type, not a feature of Winbond chips only. RABBI: "name one other chip type that uses that package"

I'm not trying to start an arguement with you STEvil, I've seen you round here long enough to know you're no idiot... but I DO think VX is UTT.

sjohnson

01-18-2005, 05:08 PM

Agree, TwinMOS SP (UTT, right?) does cas 3 NP. Don't know about the other values, might just have to try them ;)

Rabbi_NZ

01-18-2005, 05:16 PM

Agree, TwinMOS SP (UTT, right?) does cas 3 NP. Don't know about the other values, might just have to try them ;)
If you could try it on TwinMOS UTT it would be very much appreciated sjohnson... also, if you could see how they scale between 3.1v-3.6v as mentioned in 1st post it would be great.

EnJoY

01-18-2005, 05:48 PM

You say you don't want to mod your psu bet yet you alread have the 3.3 at 3.6? And PC Power & Cooling allows up to any voltage on all the rails, so they are already adjustable from the factory. Am I wrong? Sorry, not trying to pick a fight, I'm just confused.

Rabbi_NZ

01-18-2005, 05:51 PM

It has pots from factory and turning them does not void warrantee, but they only go +/- 10% which limits my 3.3vRail to ~3.6v.

Once hipro5 finishes his DDR Maximizer I will buy that and give my RAM more juice

EnJoY

01-18-2005, 06:07 PM

Ooooh, I see. Well best of luck with that then, the DDR Maximizer does look awesome.

sjohnson

01-18-2005, 06:10 PM

If I get the time to try volts lower than 3.3, I'll post them as well.

Two 256M sticks TMSP here, unmodded Infinity NF2.

#1: 253 FSB SC, 3.3v, 2-2-2-11 all aggressive error free 24 hours memtest #5
#2: 257 FSB SC, 3.3v, 2-2-2-11 all aggressive error free 24 hours memtest #5

Both, 240 FSB DC, 3.3v, 2-2-2-11 all aggressive error free 24 hours memtest #5

BIG difference in SC & DC, I'm still playing around with BIOS' to find one with less difference. BTW, it makes no difference how loose I set memory timings, those are the max error-free speeds.

Rabbi_NZ

01-18-2005, 06:31 PM

BIG difference in SC & DC, I'm still playing around with BIOS' to find one with less difference. BTW, it makes no difference how loose I set memory timings, those are the max error-free speeds.
Same for me sjohnson... there are only a few timings that change stability and it's none of the "BIG 4" timings.
Trc, Tref, and a few others can kill performance... usually once a stable speed is found I can set all timings to absolute max and it will remain stable.

STEvil

01-18-2005, 08:28 PM

sjohnson, can you provide 2-2-2-11 vs. 2.5-2-2-11 vs 3-2-2-11 for the UTT to show the bios or SPD are not interferring?

sjohnson

01-18-2005, 09:30 PM

What are you asking? Tests at those speeds? Sorry, I don't have a windows build ATM, my testing is all in memtest86.

At 2-2-2-11 set in BIOS and showing same in memtest86, memtest86 shows memory bandwidth of 1807 MB/s@240 DC

At 2.5-2-2-11 set in BIOS and showing same in memtest86, memtest86 shows memory bandwidth of 1708 MB/s@240 DC

At 3-2-2-11 set in BIOS and showing same in memtest86, memtest86 shows memory bandwidth of 1708 MB/s@240 DC

That's the best I can give at this time. I'm not up on the memory math, does the reported bandwidth at cl 2.5 and 3 of 1708 have any significance?

I'm pretty sure that my BH-5 won't post at CL 3, so I don't *think* anything is amiss but I'll admit I don't know for sure!

(Edit) whoa!, at CL3 I'm getting tons of errors. Going to try CL 2.5 to see what I get there. Back soon with another edit :)

(2nd Edit) CL3 gave 2006 errors in the first run of test 5, at the 99% level. CL 2.5 is error free after two loops. I'll try dropping the FSB at CL3 and see what happens.

(3rd Edit) 3-2-2-11 is error free at 210 FSB, 215 gave 68 errors at 99% on test 5. There may be a magic set of settings that give better results with CL3, but I'm not trying to find out, at least not tonight.

Skip

01-18-2005, 10:03 PM

i had thought test 5 always gave the errors at the end, at least that is the only time i will ever get them.

also trfc is a very important timing, going from 16, the usual, to 12-11 or 10 really helps my bandwidth scores.

sjohnson

01-18-2005, 10:05 PM

The end of test 5 involves relocating the running memtest program so as to test all the memory. So, it's a double-relocation plus the program has to continue to execute.

At least, that's the way it was described to me to explain the 99% error tendency at high FSB. If I push the memory high enough, I'll generate errors before 99%, though.

reject

01-18-2005, 10:15 PM

i thought that the program just didnt list erros till the end so as not to interfere with the test

sjohnson

01-18-2005, 10:29 PM

Nah, it shows the errors as they occur. The program is running in memory not being tested so isn't affected by the errors in memory under test.

memtest86 can err out itself, that's when you get stack errors, weird behavior and crashes.

STEvil

01-18-2005, 10:50 PM

showing no bandwidth difference between cas 2.5 and 3 may show the board is not actually changing from 2.5 to 3.. its really up in the air though.

In windows going from cas 2 to 2.5 shows a small performance hit, barely noticeable. Going from 2.5 to 3 is nearly double but still barely noticeable.. so we'll have to wait on some results from someone else or some numbers from you if you get around to it later.

Thanks!

enzoR

01-19-2005, 01:21 AM

Thats why I suggested TRP 1 and TRCD-r 1 also.

Things that give me the idea UTT is not found on VX:
- VX cannot do TRP 1. Winbond can
- VX cannot do TRCD-r or TRCD-w of 1. Windbond can.
- VX can do CAS3, Winbond cannot (have not seen reliable example as of yet)
- Dimples come with the package type, not a feature of Winbond chips only.

Measure at the DIMM slot then, mine was the same there as it was at the mosfet and there was no droop detectable by any of the three multimeters I own..

VX is Winbond UTT, winbond doenst mean it can do TRP 1, TRCD-r or TRCD-w 1 and cannot do CAS3. Heck my CH-5 does CAS3. this stuff is manufactured on 0.11 together with infineon. Finally Bigtoe already said "The secreat is out" :)

STEvil

01-19-2005, 01:36 AM

I have yet to meet any BH/CH 5/6 that cant get TRP to 1. Havent seen a single stick of VX do it yet.

Got a link to where he said that? I know its out too, but he was referring to its on the web, not that it is winbond based if it is the same one I remember.

What is your system, can you provide cas 2 vs 2.5 vs 3 benchmarks and a picture of your ram?

enok

01-19-2005, 01:41 AM

what memtest are u using? have u guys tried the last memtest86+ 1.40?

now my UTT results:

3.1v - 244mhz - PASSING all tests of v1.40, only giving one error always on same bit at #7/#8
3.2v - ...
3.3v - ...
3.4v - ...
3.5v - ...

1T 1:1 2-2-2 , default vdd

my AN7 can only handle till 244 fsb... not too bad. :rolleyes:

HermS

01-19-2005, 02:06 AM

Heres the results from my DFi NF3

3.3v - 254mhz
3.4v - ...
3.5v - 264mhz

1T 1:1 2-2-2

Will do a more thorough test when my NF4 arrives.

enzoR

01-19-2005, 02:09 AM

STEvil

01-19-2005, 02:21 AM

I dont know, nobody has even tried as far as I can tell and I dont have any to test.

googles

01-19-2005, 03:36 AM

i thought that the program just didnt list erros till the end so as not to interfere with the test

same here...ive never had errors occur until the 98/99% mark on test 5

enzoR

01-19-2005, 03:41 AM

I dont know, nobody has even tried as far as I can tell and I dont have any to test.

if it didnt then i guess that would be 100% proof VX is UTT right?

Rabbi_NZ

01-19-2005, 11:35 AM

Herms, how much memory are you using... 2x256 or 2x512?

I have added both of the results from Herms and enok to the first post. Thanks for the feedback :toast:

STEvil,
how many people have tried running those settings on VX?
how many people have tried running those settings on CH/BH?
maybe it's jus a case of you having good BH/CH and poor VX?
do you have any idea what chips MIGHT be on VX if not winbond-based?

From what I've gathered UTT is based on the same process as CH, all be it improved... BUT, the chips are not speed binned by Winbond themselves, they are just sold off as UnTesTed chips.
That's why TwinMOS speed bins the chips as 60D/50D/44D themselves... 44D is, in effect, CH44.

EnJoY

01-19-2005, 11:39 AM

Both my BH5 and BH6 could not do TRP1 without getting a lot of errors. And both of those sets did 265 in my A64 and ran every other timings as tight as possible. So,,,I think trp1 just varies per stick, not per type (ie BH5, CH5, BH6, UTT, etc)

HermS

01-19-2005, 11:50 AM

Rabbi, its just one stick of 256Mb of AA4T 50D, unfortunately one of the pair I ordered was D.O.A, should have a replacement next week.

*edit* doh just saw you wanted 2 x 512 results, sorry to confuse things Rab.

Rabbi_NZ

01-19-2005, 11:55 AM

Sorry if I mentioned I only wanted 2x512... I didnt mean to.
Your results are appreciated AND useful :)

aCidbAbY

01-19-2005, 12:48 PM

currently getting 260mhz 2-2-2-10 1T memtest86 stable @3.56-3.6vdimm.

Rabbi_NZ

01-19-2005, 12:55 PM

currently getting 260mhz 2-2-2-10 1T memtest86 stable @3.56-3.6vdimm.
Would you mind running tests from 3.1v-3.6v please aCidbAbY?
I'm trying to see how these UTT chips scale with voltage...

STEvil

01-19-2005, 01:32 PM

if it didnt then i guess that would be 100% proof VX is UTT right?

No, it would show that UTT may not be capable of what the "old" winbond chips are.

STEvil,
how many people have tried running those settings on VX?
Myself so far, two a7n8x-dlx with about 4 bioses, an nF7 with about 8 bioses, and a DFI Lanparty-UT with 3 different ones. They would not POST at those settings.

how many people have tried running those settings on CH/BH?
Here is a year old thread... there are more examples in the extreme overclocking section and I think I may have seen one or two here before too although nobody pays them much attention it seems. http://www.xtremesystems.org/forums/showthread.php?t=27206

maybe it's jus a case of you having good BH/CH and poor VX?
My VX seem to keep up with every other stick of VX out there, give or take 5mhz depending on the memory controller my CPU had when they were tested on the 754 setup.

do you have any idea what chips MIGHT be on VX if not winbond-based?
If I did I wouldnt be allowed to say. It was stated once or twice that they were not winbond based in the VX thread iirc, though..

Rabbi_NZ

01-19-2005, 01:46 PM

If I did I wouldnt be allowed to say. It was stated once or twice that they were not winbond based in the VX thread iirc, though..
Minor technicality, or straight out lie in my opinion.
They may have been refering to the fact they purchased them from Infineon seeing as they buy a LOT of the UTT chips or maybe they just lied to us.
Plain and simple, I see NO arguement that proves these are NOT Winbond chips.
NO OTHER chips I have seen have the 2 dimples on top, and NO OTHER chips I have ever used, or heard of even, react to voltage or low latencies like Winbond chips.
The odd TCCD module and very select few Micron based ones have, but that is very rare... and they DO HAVE DIMPLES on top.

I realise you may not be allowed to say, even if you do know, but unless something concrete is put forward they will remain UTT in my mind.

Anyway, I didnt want to start a debate in this thread, just wanted to get voltage scaling data.

STEvil,
if you still have your VX or some UTT would you mind posting your experiences between 3.1v-3.6v?

Cheers for a good debate. No ill feelings are meant if I come across that way.

Rabb

aCidbAbY

01-19-2005, 01:50 PM

ok so what you want me to do is see what clocks i can get with 3.1v - 3.5v

STEvil

01-19-2005, 01:57 PM

I already did, but I only tested 3.2v+.

255 @ 3.2v, 260 @ 3.3v but unstable

Have the VX at 2.9v in my a7n8x right now, 2-3-2-11.

Samsung TCB3 acted like BH-6 back in the early days.. ;)

Rabbi_NZ

01-19-2005, 02:07 PM

STEvil & aCidbAbY,
if you get a chance could you do this please:

Find where the memory is ERROR FREE in MemTest (+ or 86) Test #5 for 25-50 loops (50 preferable but I realise it takes a while to test.
Find the max MHz your RAM will do at 2-2-2 1T with 3.1v-3.6v under those conditions... ERROR FREE.

Thanks :toast:

aCidbAbY

01-19-2005, 02:09 PM

what do you need for tras? 10,8,6?

Rabbi_NZ

01-19-2005, 02:10 PM

Tras & Trc make no difference in terms of max clock for me, but Trfc should be 15.

For reference only, I run 2-2-2-10-7-15

enzoR

01-19-2005, 02:14 PM

this UTT is actually infineon... made on winbond machines. stateing that VX isnt winbond could still mean its UTT.

Rabbi_NZ

01-19-2005, 02:22 PM

this UTT is actually infineon...
Winbond still exist, they sell "most" of the chips to Infineon, but not all.

made on winbond machines. stateing that VX isnt winbond could still mean its UTT.
agree 100%

trans am

01-19-2005, 02:57 PM

aCidbAbY

01-19-2005, 03:17 PM

wow you really got them up there now trans_am... nice job for a while i was thinking that you got some bad sticks:)

trans am

01-19-2005, 03:25 PM

wow you really got them up there now trans_am... nice job for a while i was thinking that you got some bad sticks:)

Yeah, they clocked way better on the dfi over the neo2. the VX still holds the crown. 264 for 2x512 is simply amazing! you can't put a price on that. the twinmos was more to test my curiousity, but you can't do much with 2x256 these days. You can browse xs pretty fast though....

aCidbAbY

01-19-2005, 03:41 PM

lol yeah the dfi LPUT is all i have to test on:(
yeah these sticks are great and 3.675vdimm warranteed for life you cant beat that 5h!T*

STEvil

01-19-2005, 03:46 PM

tried lower values for TRCD or TRP on that UTT trans am?

I dont have a 754 or 939 setup to test with at the moment, but 250 was memtest error free for an extended period.. 255 was error free but windows complained at 3.2v. 3.3v made it benchable up to 260 but memtest would throw an error once every 10-20 passes or so... seemed to be a bit random so it may have been right on the edge of the processors memory controller.

Posting up to 280 was no problem but errors galore ;)

Rabbi_NZ

01-19-2005, 04:00 PM

Highest possible memtest86 stable:
On single channel DFI board, my VX does 264 2-2-2-10 2x512

Twinmos SP UTT does 261 with 2x256
one of the 256 sticks can max out at 267mhz.
best performance occurs @ 3.4-3.5v
Thanks trans am but if you could post 3.1v-3.6v in 0.1v increments it would be great... Im just trying to see how they scale on average.
Cheers

trans am

01-19-2005, 04:19 PM

Rabbi_NZ

01-19-2005, 04:35 PM

OK. I will try to when I have some more time. Most buying this ram will not care about scaling since they will be feeding them over 3.3 volts in the 1st place. The big concern is what do they max out at? We are not here for general performance. We want maximum performance from overclocks and MAX being the key word. I saw no gains after 3.5v and the highest point(MAX) was around 3.4-3.5v.
one thing that should be noted is that the VX did much better on single channel DFI board than in dual on neo2 with a booster. Conrad had similar results when he put his VX in his sister's DFI board from is neo2 he saw a huge gain in oc.
Whether or not people buy this RAM with 3.3v in mind or 3.5v in mind doesn't really matter... it jus gives us a better picture of what this new Winbond (I'm pretty sure that's what we're dealing with here) Process behaves like on average.

I have also heard about how well these sticks behave on 754 vs 939, I have been in contact with Conrad for a while now via PMs and emails.

But, a prime example of why I want this data is some guys are saying they still continue to gain even past 3.6v... whereas others are saying 3.2v-3.3v gets them to the max of the sticks. Obviously platform limitations will effect this, but all data is useful.

Thanks for taking the time trans am :), look forward to seeing your results.

sjohnson

01-19-2005, 05:21 PM

Sigh - tried TRP of 1 on my 2x256 sticks, DFI Infinity NF2. No post at 240, 230, 220, 200 FSB.

Toasted the BIOS on the 200 FSB try. :(

Good thing I have another BIOS chip for a hotflash :)

STEvil

01-19-2005, 06:07 PM

hmm... :/

craig588

01-19-2005, 06:19 PM

conrad.maranan

01-19-2005, 06:24 PM

2.8VDIMM - 220MHz
2.9VDIMM - 220MHz (no increase in speed)
3.0VDIMM - 225MHz
3.1VDIMM - 235MHz
3.2VDIMM - 250MHz
3.3VDIMM - 255MHz
3.4VDIMM - 263MHz (3.41VDIMM = actual reading on digital multimeter)

This is accomplished using the setup in my signature. I've realized well over 263MHz error-free and stable as of several weeks ago on the same system. I'll announce my status when I have my new NF4 rig in place. :D

Rabbi_NZ

01-19-2005, 06:39 PM

Cheers conrad, thats EXACTLY what I've been wanting :toast:

I have a feeling your next rig is going to be a BEAST! hehehehe, can't wait :D

[EDIT] so much for waiting till your time off... once addicted, always addicted!
I realise you're keeping the new max as a suprise, but would you mind telling us what voltage the max is at now... just curious if you see an increase past 3.5v on the Neo2?

STEvil

01-19-2005, 08:08 PM

You must have the values wrong, the only way people were able to run TRCD at 1, even BH5/6 is by slowing down the memory to well below 150MHz. TRP was only obtainable with the speed below 200MHz. (The closest to 200MHz that I remember was 198MHz, but that was only one guy)

Even at 150MHz I can't run my BH5 (Each stick individually will do 270MHz at 1.5-2-2 with all of the other "BH5" timings) with 1.5-1-1 with 3.8V, it gave an instant crash. I was acctually able to put the CAS at 1 though.

TRCD-w is what you need to slow down for (if you dont have seperate TRCD settings in the bios then both will be set to 1 when you select 1), I have both TRCD-r and TRP at 1 on my ch-5 and am doing 236mhz on an nF7 with 2x512 of it at 3.45v at the moment..

11.5x230 was good for a 34 minute 27 second superpi 32m run and did 39 seconds for 1m.

sjohnson

01-19-2005, 08:19 PM

...
- VX cannot do TRP 1. Winbond can
- VX cannot do TRCD-r or TRCD-w of 1. Windbond can.
- VX can do CAS3, Winbond cannot (have not seen reliable example as of yet)
...
I know it's just one guy testing, but my TwinMOS SP appears to run at CAS3 and it would not post at TRP 1 (no, I'm not trying any FSB lower than 200, one toasted BIOS per night is my limit ;) )

So, doesn't this limited evidence point to TMSP as being "VX" - at least at the module level?

STEvil

01-19-2005, 08:39 PM

It points out that UTT are not as good as old winbond chips anyways.

trans am

01-19-2005, 08:45 PM

true for 2x256BH5, but how many bh-5 2x512 kits do you see doing over 260mhz?

sjohnson

01-19-2005, 10:04 PM

lol, it isn't "not as good" so much as different. I'm joining the UTT="VX" crowd ;)

STEvil

01-19-2005, 10:42 PM

you're all going against me.. I feel left out lol :P

alexio

01-19-2005, 11:23 PM

I will be having 2 stick of 512mb within one week.

They'll only get 2.75V untill the Venice-core comes out and I'll be buying my new AMD rig.

That really sounds funny, burning-in @ 2.75V :D

cinders

01-20-2005, 01:55 AM

I found that my VX responded best at 3.3v, anything over that improved no gains.

Both my sticks max'd out at 255 HTT @ 3.3v running 1T 2-2-2-5

craig588

01-20-2005, 07:26 AM

TRCD-w is what you need to slow down for (if you dont have seperate TRCD settings in the bios then both will be set to 1 when you select 1), I have both TRCD-r and TRP at 1 on my ch-5 and am doing 236mhz on an nF7 with 2x512 of it at 3.45v at the moment..

11.5x230 was good for a 34 minute 27 second superpi 32m run and did 39 seconds for 1m.

I have the same mobo as you with the beta 4 bios and I don't have seperate RCD read and write settings.

Can you post a A64 tweaker shot just to make sure I am thinking about the correct settings?
http://img127.exs.cx/img127/5702/bh5233c.jpg is mine

enzoR

01-20-2005, 07:59 AM

Winbond still exist, they sell "most" of the chips to Infineon, but not all.

agree 100%

i thought they sold their whole dram department to infineon?

STEvil

01-20-2005, 02:41 PM

I have the same mobo as you with the beta 4 bios and I don't have seperate RCD read and write settings.

Can you post a A64 tweaker shot just to make sure I am thinking about the correct settings?
http://img127.exs.cx/img127/5702/bh5233c.jpg is mine

You have an Abit nF7 rev2? :stick:

nF2tweaker, abit doesnt let me set them seperately in the bios.

craig588

01-20-2005, 06:00 PM

I was going off of your 3dmark link:
Mobo Manufacturer DFI Corp,LTD

Mobo Model LP NF3 250GB

STEvil

01-20-2005, 09:46 PM

sold that setup a while back.

enzoR

01-22-2005, 02:53 PM

noone posting resutls!

Jupiler

01-22-2005, 03:27 PM

Haven't had much time to test my 2*512MB Twinmos Speed Premium sticks (AA4T / E44D) but the early test don't seem that bad.

Tested it in my DFI NF3 board, 2-2-2-5, 1T timings :

1*512MB, Memtest error free with "only" 3.35V was 252Mhz. Don't have a booster yet, still waiting for it to arrive.

1*512MB, at 3.1V, max memtest error free was 242Mhz.

2*512MB, same settings and 3.35V. Only tested it at 245Mhz, haven't tried it higher yet, but will soon. Maybe monday.

http://users.skynet.be/MetalHead/twinmos/memtest252.jpg

http://users.skynet.be/MetalHead/twinmos/memtest245x2.jpg

http://users.skynet.be/MetalHead/twinmos/memtest242.jpg

enzoR

01-22-2005, 03:29 PM

pretty good for only 3.35

Jupiler

01-22-2005, 03:33 PM

Thx.
Hope to squeeze some more out of it. :)

jikdoc

01-22-2005, 03:46 PM

3.2 = 250 25 passes
3.3 = 253 25 passes
3.4 = 257 25 passes
3.5 = 258 25 passes
3.5 = 259, errors

memtesting was done with tweaks in the ocz version, @ 258x11, mem speed was 2696. best combination i found was tras8, trc 10, 2-2-2. i started getting errors at 259, so before burning in overnite, i decided to boot into windows for other benchies. tried out some everest benchies, @ 3.4v 257x12 on fx55/neo2 i got mem write: 7381, mem read: 2859, latency: 38.3. i fired up pifast and blue screened. had to clear cmos to get it to boot and computer hasn't been stable at anything but stock since. this is the 2nd time this has happened to me on a neo2 and 3rd time i've seen it happen in person, same thing happened to trans am. i can't wait to get rid of this board, dfi comes in a week.

enzoR

01-23-2005, 01:29 AM

i notice these mems hit their voltage ceiling pretty fast. that jump from 3.3-3.4 is pretty big.

Supertim0r

01-23-2005, 10:30 AM

my VX died @ 3.6v :(

aCidbAbY

01-23-2005, 11:39 AM

are you going to get a new set? it is still under warranty.

Rabbi_NZ

01-23-2005, 12:55 PM

Thank you for the contribution Jupiler & jikdoc :toast:

Jupiler,
if you could do some more testing from 3.1v-3.6v it would be great too.

jikdoc,
what brand of memory are you using, what size sticks, and how many sticks?
if they are UTT you will most likely get better results with Trfc=15. Everyone else cannot run Trfc tighter than 15 from what I have heard... well, they can but not nearly as high as when Trfc is 15 or looser.
would you mind doing the testing again with 2-2-2 1T Trfc=15 please?
2-1-2 Trfc 10 will effect the overclock significantly.

I have added both of your results to the first post :)

Supertim0r

01-23-2005, 09:17 PM

are you going to get a new set? it is still under warranty.
yes i will :D

its was stable @ 258fsb 3.6v
pushed it @ 260fsb 3.67-3.7v and burned it like 2-3 days
it was pratically stable until it froze after 2h of memtest (147pass 6err)
rebooted = beeping / no post
finally restarted after a few hours and a clear cmos but same problem happened every reboot ("beeping")...had to wait (cool down ?) until it finally died :stick:

it could be great if they (OCZ RMA) could send me back some 4200plat for the 3200 VX + $$

errr...sorry my poor english

aCidbAbY

01-25-2005, 01:02 AM

mine died to night it has been acting a little fickle lately:( then all the sudden it just stopped working. i did know what it was so i checked beep codes and it a ram error then it worked again. now it gives me the same beep so i try a new set and boom im back up. this ram has died for me:(

Rabbi_NZ

01-25-2005, 05:11 PM

Sorry to hear man, what VDimm did you run 24/7?

Seems a few have died round these forums... between the VX & Booster there must be a few RMAs going OCZs way.

What cooling did you use?

Look on the bright side though... if I was doing an RMA I would be glad it's with OCZ and not another company

Supertim0r

01-25-2005, 08:07 PM

jikdoc

01-25-2005, 08:27 PM

G13

01-25-2005, 09:05 PM

same exact problem as me !!!
look post #85
My first post is an unfortunate one. Same happened to me on 1/17 but I was @250, 3.5V. Was running memtest fine but when I turned it down and started back up I got the long repeating beeps. Three dimms (on Neo2) and memory dead. It was happy @240 3.2V.

Supertim0r

01-25-2005, 11:59 PM

My first post is an unfortunate one. Same happened to me on 1/17 but I was @250, 3.5V. Was running memtest fine but when I turned it down and started back up I got the long repeating beeps. Three dimms (on Neo2) and memory dead. It was happy @240 3.2V.
maybe a problem related to this mobo ?

Rabbi_NZ

01-26-2005, 02:02 AM

@ rabbi

my mistake, trfc is 15, trc is 10, tras is 8, 2-2-2. when i wrote 2-1-2 i was referring to trw, twr, etc timings. i'm referring to the timings in ocz memtest 1.4 under the a64 options menu.

i'd like to add 2 things. voltage scaling as we are recording has some flaws as we're not accounting for the effects of burning in. i'm sure that i could run my mem at lower volts, now that i've burned in a bit than i could before. the results i'm reporting are from virgin sticks.

my sticks are VX, 2x512.

i noticed my timings did affect my stability, i was at 260 2-2-2-7 and was getting a few errors every 2nd or 3rd pass. i loosened timings to 2-2-2-8 and got no errors for 50 passes. perhaps i was exactly at the cusp of stability and that is why loosening the tras helped, perhaps i was just seeing the effects of burning in.

i'm currently at 260 2-2-2-8, 50 passes stable, 3.6v
Thanks for the update bro, keep me posted bout how high you get with 3.6v... I'm real interested in seeing your results :toast:

I realise this isnt the perfect data, but it would be impossible to only get results after X-Hours of burnin, or from virgin sticks... no one would be able to contribute.
Whether or not they are burned in I dont mined, Im not looking for the effect in burnin, Im looking for the effect of voltage scaling... but I agree, things will change after a burnin for sure.

TEDY

01-26-2005, 02:32 AM

240 2-2-2-11 25 passes and more stable 3.2v

cantankerous

01-26-2005, 05:26 AM

Keep it up Tedy. You are running on NF2 right? Let me know how high you can get error free with 3.2 and 3.3v volts. So far I am up to 243 error free. 244 is fine for awhile but errrors start to creep in after a bit and don't let up. I'm hoping a heat issue.

BTW... anyone know how hot their ram gets when burning in? Though perhaps not the most accurate, my P160 case has a built in thermometer with 2 probes that I can stick wherever I like. I stuck one of them right into the heatspreader through the slots on the back of the ram and while memtesting it reads about 45c. I have had it as high as 51c if I turn down my fan and have the room get a little humid. While gaming they get to about 41-42c and idle 36.

jikdoc

01-26-2005, 07:11 AM

Thanks for the update bro, keep me posted bout how high you get with 3.6v... I'm real interested in seeing your results :toast:

atm i'm not really keen on running it at 3.6 given the last few posts of the memory dying. i'm going to wait until dfi comes and i can directly control vdimm through the motherboard and not have to rely on the booster.

saaya

01-26-2005, 08:05 AM

mmhhh yummy :D

looking good

enzoR

01-26-2005, 08:20 AM

yep definately!

TEDY

01-26-2005, 09:38 AM

245 2-2-2-11 25 passes and + stable as rock 3.2v

W000T :D

Rabbi_NZ

01-26-2005, 11:08 AM

245 2-2-2-11 25 passes and + stable as rock 3.2v

W000T :D
Very nice results TEDY! :D Congrats on the great memory :toast:

Once you find the max at 3.2v could you do the same for 3.1v & 3.3v please?
And, could you tell me if you chips are 44D, 50D, or 60D please?

Cheers

aCidbAbY

01-26-2005, 12:56 PM

sorry about the limited results..
i forgot to write down all of them..
but im back up so far i got,

240 3.2v <<<original burnin 15hrs to start out and test,
250 ^^
254 3.3v
256 ^^
258 need more burnin
260 ^^

next burnin will be @ 250 3.2v no sweat it just takes time...

all 2-2-2-10 1T

240 and 250 are good for the same voltage... nice!

Rabbi_NZ

01-26-2005, 01:00 PM

just to clearify...

are you saying the modules are memtest stable at 250@3.2v, 260@3.3v?

cantankerous

01-26-2005, 01:12 PM

Yea very nice! Tedy your ram seems to be doing good. I can't do 245 error free. Is that on your infinity? Is it actively cooled?

aCidbAbY

01-26-2005, 01:24 PM

no 250 is fine @ 3.2 but 260 needs more time it is just the results ive gotten so far!

if it has a ^^ that means same voltage as b4. 258 and 260 are not good yet.

cantankerous

01-26-2005, 01:36 PM

Bigtoe mentioned in another thread that the newer VX are doing better. He is getting some soon to test.

Rabbi_NZ

01-26-2005, 01:51 PM

Thanks aCidbAbY, I have added your results and put you down as using a DFI nF3 754 board... let me know if I'm wrong.

cantankerous,
where did you read that... if these get much better we'll be seeing 270 2-2-2 on a regular basis :D

Garrett

01-26-2005, 01:52 PM

My Twinmos Speed Premium will do 257 in Dual Channel (socket 939) @ 3.3v... for now

MSI K8N Neo2 Platinum 0412 with bios 1.4 (official)
AMD Athlon64 3500+ Newcastle @ 10x257@ 1.6v for now... mem won't go higher @ 3.3v yet...
Twinmos Speed Premium week 48 2x512 @ DDR514 2-2-2-5
OCZ Powerstream 520W adjustable PSU

Rabbi_NZ

01-26-2005, 02:00 PM

Thanks Garrett, are your chips 44D,50D,60D?

Garrett

01-26-2005, 02:22 PM

Thanks Garrett, are your chips 44D,50D,60D?Oh I forgot I'm sorry... 44D :)

Rabbi_NZ

01-26-2005, 02:25 PM

Thanks dude... feel free to test 3.1v-3.6v too hehehe

Jupiler

01-26-2005, 03:12 PM

I'll try 3.2 and 3.4V tomorrow.
Time to cranck the 3.3V line of my Powerstream up a little more. :D

aCidbAbY

01-26-2005, 05:07 PM

let me know if I'm wrong.
:D

well my first set OCed to 260 at 3.56-3.6vdimm

Rabbi_NZ

01-26-2005, 05:11 PM

well my first set OCed to 260 at 3.56-3.6vdimm
i have a feeling this one will do better... so I will leave your previous results out, otherwise it wont show an accurate description of how these ones scale with voltage. Cool?

Are you using a Neo2 or LanPartyUT?

aCidbAbY

01-26-2005, 05:37 PM

LPUT 250 i will try and get some better numbers this time.
this is a newer batch and they seem to be clocking better at lower voltage...

cinders

01-27-2005, 05:05 AM

I removed the heatspreaders from my OCZ VX last night, it would do 260 at 3.4v but after about 5 passes would start creating errors....

After removing the spreaders it's been looping over night without any errors!.... This stuff is great!

I think the next time I'll upgrade my memory is if any DDR comes out that can do 300HTT @ 2-2-2-5 (1T) :banana:

Maybe one day.... :D

Ulti

01-27-2005, 05:14 AM

Haven't had much time to test my 2*512MB Twinmos Speed Premium sticks (AA4T / E44D) but the early test don't seem that bad.

Tested it in my DFI NF3 board, 2-2-2-5, 1T timings :

1*512MB, Memtest error free with "only" 3.35V was 252Mhz. Don't have a booster yet, still waiting for it to arrive.

1*512MB, at 3.1V, max memtest error free was 242Mhz.

2*512MB, same settings and 3.35V. Only tested it at 245Mhz, haven't tried it higher yet, but will soon. Maybe monday.

images

So the DFI is compatible with the DDR Booster? I've been looking for this info for like 2 weeks :toast:

Jupiler

01-27-2005, 05:31 AM

So the DFI is compatible with the DDR Booster? I've been looking for this info for like 2 weeks :toast:

Haven't tried it with booster (which should arrive tomorrow or monday).

Upped the 3.3V line on Powerstream 420W to 3.5V. So 3.5V is bios would give 3.35-3.36V.

cantankerous

01-27-2005, 07:00 AM

So removing the heatspreaders really does help eh? Must be a heat issue. Too bad you lose warranty from that. Heatspreaders should be an option.

Garrett

01-27-2005, 08:16 AM

Thanks dude... feel free to test 3.1v-3.6v too heheheWell the $h1t has hit the fan if you ask me :(

these 2 double sided dimms 512mb each: massive errors in memtest
errors all @ 1022mb... if I switch the 2 around it's @ 1.8 and 1.9mb...

Could it be that my A64 mem controller doesn't like the 2 double sided dimms @ 1T anymore? I mean I even get errors @ 200 2-2-2-5 1T :(

Putting my 2x256 BH5 back in results in no errors whatsoever...

If I try only 1 doublesided DIMM (doesn't matter which 1 of the 2) then I get no errors at all...

TEDY

01-27-2005, 08:28 AM

rabbi please correct

it's not SP but 2 sticks just happens to work DUAL too...hehe...

AA4T / 05034

i can do 25 loops with 1 stick at time at 3.3vdimm

will do some burn in at weekend.

Jupiler

01-27-2005, 10:27 AM

Well,
After my earlier results, which looked promising, I'm kinda disappointed with the end result.

1*512MB

3.22V : 250Mhz (not to bad, IMO)
3.35V : 253Mhz (gained 1Mhz compared to last time)
3.46V : 255Mhz (not that impressive)

As you can see, increasing VDimm by 0.25V only gave me 5Mhz more, not that much after all.

Didn't have much time in testing the 2 sticks together, but managed to do 1, at 3.35V.

3.35V : 246Mhz (+1Mhz compared to last time.) Any higher would result in errors.
That's it for now. Results on the Neo2 might follow later (can't promise anything)

aCidbAbY

01-27-2005, 11:31 AM

ok i tried to burn in at stock speeds and the 2-3-3-8 stock speed (200) errored out @ 3.2 vdimm.. WTF is up with these things?
then i went back to 2-2-2-10 and it was fine wierd it isnt running at stock speeds..
maybe because i burned in at 2-2-2-10?

new one BTW

200mhz @ 3.2vdimm 2-2-2-10

EDIT:: im thinking of getting another set and finding the best between the 2 :confused: do you guys think it is worth it? i mean this one seems to be acting wierd right off the bat. :confused:

Rabbi_NZ

01-27-2005, 11:55 AM

Garrett,
try clearing CMOS and then loading optimised defaults... sounds like a timing has changed maybe

TEDY,
done :)

Jupiler,
updated yours too :)... sounds like it's your board/cpu holding you back right now

aCidbAbY,
standard timings are 2-3-3 @ 2.6v or 2-2-2 @ 3.2v.... no where does it say 3-3-3 timings as stock.
I dont think you need a new set though bro, yours are doing very good for not even being burned in yet... personally, I think it's your board/cpu causing the wierdness with CAS 3. If they work fine at 2-2-2 dont worry bout CAS 3 results... 256 @ 3.3v is above average I think and if sheeet does happen later on jus return them for an RMA

aCidbAbY

01-27-2005, 12:00 PM

lol yeah i meant 2-3-3-8 sorry..
did you get the new one? 200mhz?

how long should i burnin?

also these new ones have OCZ in black on the side. my first set didnt have anything on the side just pressed into the heatsink..

craig588

01-27-2005, 12:01 PM

Winbond chips never really liked to run at cas 3

Rabbi_NZ

01-27-2005, 12:05 PM

lol yeah i meant 2-3-3-8 sorry..
did you get the new one? 200mhz?

how long should i burnin?
I dont know much bout burnin to be honest... I have never seen a gain but I know of many who have, and I trust a lot of them. I've heard a few people say that 24-48hours is a good timeframe, then see if they have improved and burnin at the new speed.

Bout the new results, are you saying your sticks MAX at 200 MHz with 3.2v?
Previously you said they did 250 MHz :confused:

Jupiler

01-27-2005, 12:12 PM

Jupiler,
updated yours too :)... sounds like it's your board/cpu holding you back right now

Don't know.
I was able to run 1 stick of PC4800 Gsill at 290, 2.5-3-3-7, 1T on the DFI.
But it might be, indeed.

Rabbi_NZ

01-27-2005, 12:28 PM

also these new ones have OCZ in black on the side. my first set didnt have anything on the side just pressed into the heatsink..
Mine are only pressed too, and it seems most of them had the black badge right from the start... even the picture on their site has always had the black badge.
I'm wondering if the pressed ones were the beta sticks or something....

aCidbAbY

01-27-2005, 12:29 PM

I dont know much bout burnin to be honest... I have never seen a gain but I know of many who have, and I trust a lot of them. I've heard a few people say that 24-48hours is a good timeframe, then see if they have improved and burnin at the new speed.

Bout the new results, are you saying your sticks MAX at 200 MHz with 3.2v?
Previously you said they did 250 MHz :confused:

no i am saying that the best i could do @ 200 2-2-2-10 is 3.2

yeah thats what i was thinking about these too..
i got mine from the first batch thx to trans_am...

Rabbi_NZ

01-27-2005, 12:32 PM

no i am saying that the best i could do @ 200 2-2-2-10 is 3.2
SOrry jus to clearify... your modules WONT do 200 2-2-2 without at least 3.2v right?

aCidbAbY

01-27-2005, 12:34 PM

yes sir

Ulti

01-27-2005, 12:44 PM

yes sir

Same problem here, 2-2-2-5 3.1V isnt memtest stable. I can 3D bench and pi tho.

Rabbi_NZ

01-27-2005, 12:54 PM

yes sir
That's real strange... your modules are doing real well with 3.2v but anything below they jus wont play. Looks like the "2-3-3 @ 2.6v 2-2-2 @ 3.2v" rating was very well thought out after all... some of these modules NEED 3.2v to get going.

Funny thing is mine did 220 2-2-2 @ 3.0v straight out the box... might test again soon to see how they go with 3.0v after the burnin.

Either way though, I dont see you putting less than 3.2v into those badboys so it's really not important :toast:

cantankerous

01-27-2005, 01:06 PM

Mine need at least 3.2 just to get going as well.

arj

01-27-2005, 01:33 PM

I've just received my VX replacements today and so far the results looks very promising, better than the first pair. The new pair just did 2 hours of memtest Test #5 and one hour of 3D Mark Lobby test at 255Mhz 2-2-2-5 3.5v (multimeter). I think I can squeeze out a few more Mhz out of this babies. My first pair can only do about 245Mhz 2-2-2-10 3.5v.....

cantankerous

01-27-2005, 01:48 PM

wow, very good news. Did your first ones die?

I can't wait to get my new setup so I can see what mine can do. This is killing me being stuck on NF2 with crap results.

Jupiler

01-27-2005, 01:53 PM

I experienced about the same.
After I finished benching, went back to normal settings (222*11). Knowing that they need volts, set VDimm at 3.0 and started the system.
I then noticed that it took a long time for the system to boot, until it suddenly froze.
Rebooted and increased vdimm to 3.1. Same thing happened, system felt sluggish and slow.
Took the sticks out, and put my Corsairs BH5 back in and rebooted.
System booted as normal, and felt quick and responsive again. Weird.

As you said, perhaps they need 3.2V to start with.

arj

01-27-2005, 02:00 PM

wow, very good news. Did your first ones die?

I can't wait to get my new setup so I can see what mine can do. This is killing me being stuck on NF2 with crap results.

After running the first pair with 3.5v for nearly two weeks 24/7 (Folding), I began to experience frequent reboots and the frequency of reboots increases over time. When I tested the rams individually, one of the sticks have thousands of memtest errors, even at stock settings. The other stick has no problems at all.

Rabbi_NZ

01-27-2005, 02:06 PM

Jupiler

01-27-2005, 02:10 PM

Would you mind finding the max MemTest stable MHz @ 3.1v please Jupiler?

With 2 sticks, I suppose?
Will do, but can't post results until tomorrow evening.

felinusz

01-27-2005, 02:13 PM

Guys, as far as burning in your VX, the best way for you to do it, while charting to see if you are getting any gains, would be uwackme's BH-5 burnin.

http://www.dfi-street.com/forum/showthread.php?threadid=1860&highlight=uwackme+BH5+burn

Basically, you overvolt your RAM as high as you can, running it with 2-2-2 timings. You run the memory at a speed that is unstable, that causes errors in memtest86 test #5.

You run test #5 looped for ~6-12 hours, recording the number of errors that appear by the time the test period is over.

Then, you repeat the cycle for the exact same amount of time (~6-12 hours), at the same voltage, and unstable speed. You record the number of errors again.

After several cycles, you should see that the number of errors your memory causes starts to decrease (you record the number of errors, so you can chart, and see if your RAM is having any gains from the burnin process). Eventually, the errors may go away entirely.

Once you are getting <100 errors, boost the FSB up 1-2 MHz, and repeat the process.

This worked quite well for my BH-5, and gave me a good visualization of the gains that my memory was seeing, throughout the burn-in process. Worth a shot for you guys :)

Great thread by the way, an interesting read.

Rabbi_NZ

01-27-2005, 02:20 PM

Cheers felinusz :toast:
I'll add uwackme's guide to the first post... I've actually read it a long time ago but forgot about it

enok

01-27-2005, 03:28 PM

ok, i think this is the best thread were to implement this Rabbi_NZ. :)

let's list our stick pcb aswell as oc results, to try to understand wich is the best pcb for UTT chips.

mine pcb is KO-6633 and btw i'm using TwinMOS DC kit (non SP) 50D with 0451 chips. (if i'm not mistaken)

PS: as OCZ VX has heatspreaders, it will be more dificult to do this for them. only some will be bold enought to take them off in order to see pcb rev... :p:

Jack

01-27-2005, 03:45 PM

a copy-paste from my reply in the xtreme twinmos thread

tested my mem's today :)
# 1 : 261MHz
# 2 : 254MHz
# 3 : 240MHz
# 4 : 257MHz
# 5 : DOA
# 6 : 253MHz

3,65V, according to the multimeter
not too bad for out of the box, without burn in

pity one module is dead,
gonna rma that one with # 3...

next week burn-in :D

cantankerous

01-27-2005, 03:47 PM

can I have module 1?

hehe

Garrett

01-27-2005, 03:54 PM

Rabbi, I already flashed to a newer bios, 1.51 but that didn't help either...

When I pop in my old trusted BH5 sticks the problems are gone.... looks like RMA time :(

I wanna test them elsewhere though, maybe it's my cpu messing about (memcontroller)

Rabbi_NZ

01-27-2005, 04:25 PM

enok,
thanks enok I have added your info... if everyone else could do the same I will add all info we have here.
eventually I'll post requirements on the first post but I have little time now.

Jack,
thanks for that. how many loops of MEMTEST #5 did you do to get those figures?

Garrett,
try this...
clear CMOS, put UTT sticks in, go into BIOS and load optimised defaults, set timings to something like 2-2-2-10-12-15 and all the rest to how you usually have it, run memtest again.
you said sticks work individually but not together... are your BH5 2x256 or 2x512?

Jack

01-27-2005, 05:56 PM

Jack,
thanks for that. how many loops of MEMTEST #5 did you do to get those figures?

5 loops
i'll report back after i did a burn-in :)

aCidbAbY

01-27-2005, 06:59 PM

MarilynMX

01-28-2005, 02:40 AM

check out my VX
3.4V 240~249MHz
3.6V 250~258MHz
3.7V 259~261MHz
3.8V 262~264MHz
3.9V 265~?
same timings 2-2-2-5 1T
I found if I add volt during memtet86 bench will be more stable, e.g 265MHz 4 errors in 20 loops #5, no errors and superPi&3Dmark stable after 0.1V added
I stopped at 265MHz since booster is burning at 3.9V, maybe I can gain few more MHz but I won't hit 4V at all I mean isn't it scary?

MarilynMX

01-28-2005, 02:48 AM

I've been running 255MHz for few weeks, just upped my FSB up to 265MHz yesterday and system runs fine expect the massive heat

Pictures are horrible which were taken by my cam phone

Add: For every 2~3Mhz increasement I ran 20 loops #5 and #5~10 loops #7 plus 5 loops std test. If error free then jump on superPi and 3Dmark and Sandra etc

enok

01-28-2005, 02:59 AM

have anyone tried memtest86+ 1.50 (http://www.memtest.org/#downiso) ? :)

cantankerous

01-28-2005, 04:24 AM

Marilyn, very impressive. That is what I hope my sticks can do! 265 is great. I would be happy with 260 anything over would be pure bliss.

What cooling are you using on those sticks as they must be scorching!

MarilynMX

01-28-2005, 04:47 AM

I haven't got any fan to blow on ram and booster. only 6 aluminium hs stuck on booster. partially cooled by xp90 and great f***king cold air in Canada you know it.
I will try 266+ later today if no go with 3.9V max then I'll give up and stick with 265

cantankerous

01-28-2005, 05:17 AM

I live in Canada too but my house is warm!

No fan is crazy. I can't believe your ram isnt erroring due to that. I notice when my temps get up there I start to error.

Stuck with 265? I would kill to get that. You care to share an everest read, write and latency with us at those speeds?

Ulti

01-28-2005, 07:49 AM

Would you mind finding the max MemTest stable MHz @ 3.1v please Jupiler?

My TwinMos 44D can't do 200 2-2-2-5 memtest at that setting :stick:

alexio

01-28-2005, 08:00 AM

Just for reference, this is with 2.75V. It's error-free for atleast 4 hours at this timings/speed.

Enjoy :p:

http://img.photobucket.com/albums/v342/AlexKok33/ScreenHunter_003.jpg

*It's Twinmos Speed Premium

conrad.maranan

01-28-2005, 11:40 AM

MarilynMX - You've got an excellent set of memory. Let's get some fans blowing on them just to make sure they live a long and happy life. :D

mrlobber

01-28-2005, 11:43 AM

Ok, couldn't finally resist while waiting on the DFI nF4 and put the VX in my DFI Infinity rig.

First of all, they didn't want to run 2x512Mb dual channel 1T 2-2-2 even 220fsb .

But that's probably because the memory controller on my Infinity just plain sucks (it did the same with 2x512 Mb Kingston BH-5).

So I just started playing with them separately, that is, in single channel configuration (1x512Mb).

The first results (I'm doing the burn-in simultaneously as well, so only results from one stick, the other will follow).

3.2v: 243 fsb 2-2-2-9-12-14 memtest stable for 100+ passes.
3.3v: 247 fsb 2-2-2-9-12-14 memtest stable for 100+ passes.

2-2-2-11-13-15 in memtest gave ~20 Mb/s less bandwidth so I just remained with the more aggressive timing values.

Heatspreaders are on, and a fan is blowing on the ram.

The board is unmodded, so the voltages are supplied by the mobo, my PCP&C (vanilla 510W) 3.3 rail is at 3.42 (with the built in pot maxed out).

We will see how it goes further.

arj

01-28-2005, 11:48 AM

Here's my max so far, with 3.6v. Will do the full 3.1v - 3.6v test tomorrow.

http://img.photobucket.com/albums/v625/aimanrifhan/260VX.jpg

cantankerous

01-28-2005, 11:58 AM

wow no errors till 195 passes then none again up to 260. Weird.

conrad.maranan

01-28-2005, 12:04 PM

It's probably just some small insect getting zapped by one of the PSU rails. I'd still call arj's overclock stable.

mrlobber

01-28-2005, 12:12 PM

It's probably just some small insect getting zapped by one of the PSU rails

Ouch, that poor insect :p:

I'd call it stable as well.

Just to add from personal experience, I wouldn't call memory memtest stable in the 5th test without at least 50 passes... Often you see first errors popping up around the 20-30th loop and then the number of errors building up slowly (the situation where usually the burn-in should come into play :toast: )

Garrett

01-28-2005, 12:17 PM

Jupiler

01-28-2005, 12:30 PM

Stickied. ;)

Garrett

01-28-2005, 12:35 PM

Well I cleared the CMOS... no good :(

TEDY

01-28-2005, 12:39 PM

3.1Vdimm

235 2-2-2-11 50 loops error free
238 2-2-2-11 10 errors after 2hours of memtest#5

gotta burn this mofo :D

cantankerous

01-28-2005, 12:55 PM

keep it up matey!

aCidbAbY

01-28-2005, 01:10 PM

man the method that i am trying isnt working yet i dont know why but, the errors stay the same...

cantankerous

01-28-2005, 01:55 PM

unfortunately not all memory benefits from burning in. Your's just may be stubborn. How many errors are you trying to rid of and at what speed and voltage?

Rabbi_NZ

01-28-2005, 02:18 PM

MarilynMX,
could you test 3.1v-3.3v too please?
and is it fair to say your memory does this? I wasnt 100% sure from your post
3.4V 249 MHz
3.6V 258 MHz
3.7V 261 MHz
3.8V 264 MHz
3.9V 265 MHz

Ulti,
Could you post your results from 3.1v-3.6v please.... or as much of that range as you can? even if 3.1v only lets you do 190 MHz 2-2-2, results is all we want :toast:

mrlobber,
thanks dude, I'll put you down as 1x512mb for the following speeds...
3.2v: 243 fsb 2-2-2-9-12-14 memtest stable for 100+ passes.
3.3v: 247 fsb 2-2-2-9-12-14 memtest stable for 100+ passes.
Interestingly, I've never seen VX doing Trfc=14 on AMD64, but you seem fine with it :D

Garrett,
2x512 does put more strain on the memcontroller so maybe your CPU is on it's way out and it jus wont take it anymore :(
Either that or it's time to get that RMA in motion hehehehe

Tedy,
are these results with 1x512 or 2x512?
* 235 2-2-2-11 50 loops error free
* 238 2-2-2-11 10 errors after 2hours of memtest#5

Jupiler,
Biggy Cheers for the sticky bro! My first one at XS :D

Adding new results now, thanks everyone!

Rabbi_NZ

01-28-2005, 02:27 PM

I wont be here for a few days so no updates till I get back... I expect results from everyone tho! :)

Jupiler

01-28-2005, 02:31 PM

NP, Rabbi_NZ. ;)

Still testing 2*512MB on my DFI at 3.1V, as you requested.

220 and 225Mhz passed test #5 (25 runs).
So far, 230Mhz, 2-2-2-5,1T seems ok, no errors.
Will try some more tomorrow, getting late here.

MarilynMX

01-28-2005, 02:36 PM

Rabbi_NZ you are about right. every level represents the max expect 265MHz since I just stopped at 265MHz, I was kinda too chicken to jump on 266+ at that moment. Tonight I will try higher FSB with VDIMM NO HIGHER than 3.9V
some screenies taken from Sandra 2005 and everest home edition, please see for yourselves

I don't have any soft to show my VDIMM and I know some people don't believe my ram could survive and kick around. I recalled andyOCZ said any more than 3.4V would not help with FSB increase but only brings mass heat, seems it doesn't work on mine.

I also have some shots for ddr500 if you guys want to make a comparison I can post them too.

Add: Rabbi_NZ I used to test on 3.1~3.2V well I can guarantee 200MHz 2-2-2-5 1T stable or maybe 210ish; 3.3V let's say 240MHz max. Before I was running 240x11 since I got best Pi result(250x11 bootable but Pi unstable).

bobmanfoo

01-28-2005, 02:38 PM

have anyone tried memtest86+ 1.50 (http://www.memtest.org/#downiso) ? :)
i use 1.5, works fine..... one of my UTTs died on me so can't contribute any info on that, but now have some VX :banana: loving it, hit 245mhz reading 3.26v under load, need to burn it in and will report back

cantankerous

01-28-2005, 02:38 PM

Wow. 34ns latency. Now that is something to be reckoned with. When exactly did you get your chips?

bobmanfoo

01-28-2005, 02:51 PM

k, got a question about the voltages, should list them according to what the voltage reads or what i set the booster at in bios? voltages drop about .1v under load, tia

MarilynMX

01-28-2005, 02:54 PM

cantankerous The date was Dec 18th or a day before. Actually I'm worried about my mobo way more than the RAM~~

bobmanfoo I put down the voltage shown on booster during test. please bear with my limited knowledge, why would you need the booster if you had your board vdimm moded? according to my situation my booster always keeps same VDIMM no matter idle or load. So both my RAM and booster are behaving really weird...

bobmanfoo

01-28-2005, 03:50 PM

my vdimm mod only goes as high as my 3.3v, with my +3.3 rail at 3.34v i get 3.27vdimm, i don't really feel like running my +3.3 rail at 3.5v 24/7, i'll be running my VXs at 3.5v 24/7 when i fully break it in

your vdimm doesn't drop at all? is this according to the ddr booster led or from a multimeter? my ddr booster doesn't show the drop,only when i use my multimeter. my vdimm doesn't drop when priming or gaming only when memtesting

aCidbAbY

01-28-2005, 09:13 PM

ok well mine doesnt seem to be responding to burnin:( i gues im gonna get another kit...

:::update:::i cant even get 2-2-2-6 @ 200mhz 3.0vdimm WTF :stick:

:::update#2:::ok it is final i cant even get 200 mhz @ 3.2 :stick:

::update#3:::i just sent an email to andyOCZ about this...

i get memtest errors all the time. the thing i thought yuou might find interesting is it errors out @ 3.2 vdimm 200mhz 2-2-2-6 in an hr and 13min.. i got 8 errors in 46 passes...VVVVVVVV

failing adresses:::
0002cc5d418 716.8mb
0002dc5d3f8 732.8mb
0002cc5d418 716.8mb
0002dc5d3f8 732.8mb
0002cc5d418 716.8mb
0002dc5d3f8 732.8mb
0002cc5d418 716.8mb
0002dc5d3f8 732.8mb
is there a may to desipher this?

in a previous post i have also said that it was stable at even stock speeds like 2-3-3-8 200mhz

MarilynMX

01-29-2005, 01:50 AM

I've been out for whole day...now back on business.
265x10 superPi 32s, 266x10 superPi 31s ??
I had little time for tests so now I can only provide a hard fact that 3.9V gets me as far as 266MHz, same timings.
Guys I just realized my DC is even worse than my cam phone, look at these pictures!

I ran memtest at 265 again thought I should obey Rabbi_NZ's 25 passes rule as everybody else

cantankerous

01-29-2005, 05:13 AM

Holy crap im jealous. 266 and 33ns latency?

]=(

where in ontario are you and where did you get your sticks? Have you ever thouched them at 3.9v?

Garrett

01-29-2005, 08:02 AM

...Garrett,
2x512 does put more strain on the memcontroller so maybe your CPU is on it's way out and it jus wont take it anymore :(
Either that or it's time to get that RMA in motion hehehehe...
I've tested another cpu, a 3000+ and it doesn't help $h** you know...
Exactly the same error, also @ 1022.0mb... bah... I really think this memory is phu**ed :(

I'll definitely get some G.Skill... that way I'm 100% sure about getting 275 DC @ decent timings... it's time for some MHz power ;)

Jupiler

01-29-2005, 10:11 AM

TEDY

01-29-2005, 10:31 AM

seems this twinmos is really weird,eh ?

bobmanfoo

01-29-2005, 12:25 PM

With 3.1V, 231Mhz is the max I can run 2*512 sticks Memtest errorfree.

Strange thing though is that, at 221Mhz and same settings, I get a registry error when XP boots, and a memory error after some time.
After giving it 3.2V, all seems ok now. Weird.
thats not so strange, my VX doesn't do too well until 3.2v, then it takes off, i've only gotten to 255mhz so far with 3.3v, with 3.1v i can't get 230mhz stable

alexio

01-29-2005, 12:38 PM

Well, it is strange that he has no problem at a higher speed wile at 10mhz lower he gets a memory-error an d a registry-error.

MarilynMX

01-29-2005, 04:36 PM

cantankerous I bought my ram from Poklik (http://www.poklik.com)
as an authorized OCZ retailer in Montreal. I was told by a XS member that they got only 1 set left so I took it home. Yes I'v been feeding 3.9V to my ram since two days ago. I'm going to get some TCCD to replace these VX during summer.

aCidbAbY

01-30-2005, 12:53 AM

3.9 is a rediculous voltage to be running 24/7 those will die fast.
does anyone else notice how many of these kits are bad?
i mean i love OCz and all but, WTH ive only had my hands on 2 and both of mine are bad..
ive also heard about alot of dead sticks and rma's..

Garrett

01-30-2005, 02:45 AM

TEDY

01-30-2005, 03:26 AM

update

3.2

240 2-2-2-11 25 loops memtest#5 OK

3.3

245 2-2-2-11 25 loops memtest#5 OK

:D

bobmanfoo

01-30-2005, 12:33 PM

ram isn't burned in yet but heres my results so far

3.2v - 242mhz

3.3v - 257mhz

3.4v - 263mhz

Yeah, IF those sticks live until summer... 3.9v is foolish
:eek:
i agree, i thought these sticks don't give much if any above 3.6v

cantankerous

01-30-2005, 01:04 PM

Bob, seems like you have some very good chips as well. Lucky guy.

mrlobber

01-30-2005, 01:26 PM

ram isn't burned in yet but heres my results so far

3.2v - 242mhz

3.3v - 257mhz

3.4v - 263mhz

0.1v additional Vdimm and max stable speed goes up by 15 Mhz? :slobber:

One of my VX is now doing 250 fsb 3.3v single channel memtest stable. The other still remains untested.

bobmanfoo

01-30-2005, 01:38 PM

0.1v additional Vdimm and max stable speed goes up by 15 Mhz? :slobber:

One of my VX is now doing 250 fsb 3.3v single channel memtest stable. The other still remains untested.
yeah, kinda strange but i think it has to do with the voltage drop under load and the 3.2v take-off point

Rabbi_NZ

01-30-2005, 08:50 PM

Jupiler,
added your 3.1v results....
Sounds like you may have one stick weaker than the other... might pay to check them individually to make sure

MarilynMX,
Nice sticks dude.... Take good care of them! :D
If you insist on using 3.9v it will be interesting to see how long they last... with your COLD temps it will def help the cause but I would still be very worried at 3.9v.
Updated your results.

bobmanfoo,
You look to have some of the better modules around... congrats :D
Did you read those voltages from the booster itself or from a software monitoring program?
Apparently if your 5v line on PSU is not at 5v it will cause the Booster to read incorrectly... is yours at 5v?

aCidbAbY,
sorry to hear that bro... those sticks looked like were going to be real nice too.
Hopefully OCZ will get you some nice handpicked modules for all the trouble you've been through! Good luck

TEDY,
Updated your results... cheers :toast:

TEDY

01-31-2005, 05:53 AM

DUAL TWINMOS

3.3vdimm
245 2-2-2-11 #5 32 errors in 7 hours....wtf is liming me ? I need more vdimm ?

odb

01-31-2005, 06:19 AM

3.2v 245mhz 2x512
3.3v 252mhz 2x512
3.4v 255mhz 2x512

might do higher i dunno, crashes after 255.

memtest stable, 30 passes on #5, 2-2-2-10 upto 255

TEDY

01-31-2005, 06:25 AM

odb so raising vdimm doesnt help a lot afterall ....just few mhz ...and on AMD64 system :(

can you try 3.6 ?

odb

01-31-2005, 06:35 AM

i never tried 3.5v+ yet, but will tonight. im still happy with 255 tho, when i first burned in when i got them the end of november, highest i could get was 242 at 3.4v

bobmanfoo

01-31-2005, 08:44 AM

bobmanfoo,
You look to have some of the better modules around... congrats :D
Did you read those voltages from the booster itself or from a software monitoring program?
Apparently if your 5v line on PSU is not at 5v it will cause the Booster to read incorrectly... is yours at 5v?

my +5v rail is at 5.08v, but i don't use the ddr booster's leds anyway, not really accurate, i measured the voltages with a multimeter, ill be testing higher voltages later...making the fun last longer ;) i've hit 265mhz stable with a minute turn on the ddr booster from 3.4v....

Wingz

01-31-2005, 08:55 AM

anyone have a pic showing the measuring points for the vdimm on the Msi Neo 2 Plat, i currently have my Booster @ 3.4v @ 255 for my VX but want to use my multimeter to make confirm the voltage.

arj

01-31-2005, 09:20 AM

anyone have a pic showing the measuring points for the vdimm on the Msi Neo 2 Plat, i currently have my Booster @ 3.4v @ 255 for my VX but want to use my multimeter to make confirm the voltage.

There's a circle above the forth memory slot with two dots, one in the white area and the other in the black area. The black one is where you measure the VDIMM.

Wingz

01-31-2005, 09:28 AM

There's a circle above the forth memory slot with two dots, one in the white area and the other in the black area. The black one is where you measure the VDIMM.

thanks arj http://www.pclincs.co.uk/Forums/images/smilies/thumb.gif

bobmanfoo

01-31-2005, 09:52 AM

3.5v - 266mhz

3.6v - 268mhz ---MAX tried 3.7v @ 270mhz not stable

TEDY

01-31-2005, 09:58 AM

bob you reached max then.

cantankerous

01-31-2005, 09:58 AM

DAMN 268! You must get like 30ns latency with that.

bobmanfoo

01-31-2005, 10:18 AM

i dunno about 30ns, i only get 34ns at 260, you are talking about everest right? i haven't tried getting into windows at 268mhz because i'd have to use 9multi, my cpu maxes out at 2.67ghz, 268mhz in dual channel, wonder what i could get in single channel.... might try it when i get bored

mrlobber

01-31-2005, 10:29 AM

It seems at least one of my VX sticks doesn't respond to burn-in at all. 4 days (& nights ;) ) of testing, I've tried also various chipset voltage adjustments just in case, have had 3 sessions of continuous 8hour memtest runs, and many not so long ones, the errors just keep coming where they did previously.

Rabbi_NZ

01-31-2005, 11:30 AM

odb, thankyou I have added your results :)

TEDY, it's almost certain your chipset is limiting you... wait till you go to a nice AMD64 setup and the MHz should climb ;)

bobmanfoo, updated your results too dude... very nice :toast:

arj

01-31-2005, 12:24 PM

Here's my results. They all pass at least 25 loops of Test #5 at 2-2-2-5 and the voltages are taken with a multimeter and are within 0.03v

3.1v - 209Mhz
3.2v - 222Mhz
3.3v - 242Mhz
3.4v - 259Mhz
3.5v - 261Mhz
3.6v - 263Mhz

TEDY

01-31-2005, 12:38 PM

Rabbi_NZ

01-31-2005, 12:51 PM

enok, changed your results as per PM request

arj, thanks a lot for the thorough testing. it's appreciated :)
your results have been added.

TEDY,
if CPC OFF lets you get higher it usually points to chipset limit... unfortunately, though, with 2x512mb extra vdd does very little in my experience with nF2... once you hit the limit, you've hit the limit.

TEDY

01-31-2005, 01:04 PM

nah also with cpc off 255 i get tens of errors 2*512...

so must be chipset eh ? i should go to someone with DFI NF3 and check there ? :P

cantankerous

01-31-2005, 01:25 PM

with CPC off my chips do no better either. ARJ your chips are seeming to be just like mine. About 242-243 max at 3.3v. I hope I can get as high as you with 3.5v though.

cantankerous

02-01-2005, 06:26 AM

Oh that is weird. Exact same error and only 1 per pass. Maybe that one part of the IC is very sensitive to voltage? Do you think you can burn that out over time?

enok

02-01-2005, 07:09 AM

PLEASE, POST YOUR PCB REVISION ASWELL AS CHIP'S DATE AND LATENCY, IN ORDER TO MAKE A BETTER "CONTRIBUTION"! :p:

sorry but i had to put it on caps! :D

TEDY

02-01-2005, 07:18 AM

/me :slap: enok

Rabbi_NZ

02-01-2005, 11:16 AM

PLEASE, POST YOUR PCB REVISION ASWELL AS CHIP'S DATE AND LATENCY, IN ORDER TO MAKE A BETTER "CONTRIBUTION"! :p:

sorry but i had to put it on caps! :D
Modified 1st thread... thanks for the suggestion :toast:

mrlobber

02-01-2005, 11:41 AM

Modified 1st thread... thanks for the suggestion :toast:

Good point. But I'm personally not going to remove the heatspreaders until I test my VX in A64 operation to see their true (in)capability. :p:

BeSaiD

02-02-2005, 11:09 AM

and what about P4 results?

pik-ard v1.1

02-02-2005, 11:42 AM

hmmm, looks like you need to reconfigure your averages... you've got four 3.6v examples, yet the average is of only two. dunno what else is outdated.

just figured i give you a nudge to get that sorted. ;)

Rabbi_NZ

02-02-2005, 11:43 AM

No one has posted them... it would be great to get some though

pik-ard v1.1

02-02-2005, 11:47 AM

uh... looking through the 1st post, i see four people with mhz's beside the 3.6v part....

Rabbi_NZ

02-02-2005, 12:20 PM

pik-ard,
jus a misunderstanding bro... i was answering the Intel question, we must have posted at the same time.

i havent updated averages in a few days.. i will tonight wen i get home from work. jus havent had time last few days.

pik-ard v1.1

02-02-2005, 12:21 PM

oh, i see now. sounds good. :)

El Snorro

02-02-2005, 01:26 PM

Msi k8n neo2 plat
2x512mb OCZ VX
booster

3,2:
240mhz
memtest stable.

3,5v:

http://img135.exs.cx/img135/5972/273sandra1fp.jpg

:banana: :banana: :banana:

Thats not memtest stable though

Garrett

02-02-2005, 01:41 PM

How high can you go memtest stable?

El Snorro

02-03-2005, 05:44 AM

How high can you go memtest stable?

Dont know yet, but i'm running 230mhz 2-2-2-5@ 3,2 24/7

Wingz

02-03-2005, 05:54 AM

i got one stick that can do 270 @ 3.5v and one that could do 255 and now it does not like voltages above 3.3v or i get errors .

godsa

02-03-2005, 07:10 AM

I will contribute to this thread soon as my DFI is modded.
Seems me and Tedy have the same sticks...Hopefully I can get 260 :toast:

Great job everyone..Begining to love this forum..Team work

Esdee..Good luck..shoot for 270 :toast:

esdee

02-03-2005, 07:13 AM

ok good news

my MSI is crap!
my twinmos (the ones i had since the beginning - not speed premioums)
did 240 at 3.2V (Vio modded the MSI)
i you can recall 240 was my limit with 3.55 too!

odd wall. Monday ill have the DFI ultra, so we will have some 260+ action!

forgot to mention

MSI Neo2 1.3b6 BIOS Vio modded
Winchester 3500+
Twinmos UTT (not speed premioums) 04514 AA4T ending P/N

240@2-2-2-5@3.2V full load,3.3V idle

Rabbi_NZ

02-03-2005, 11:18 AM

El Snorro,
could you test your modules to find their max MEMTEST STABLE point for voltages 3.1v,3.2v,3.3v,3.4v,3.5v,3.6v please?
for more info bout MemTest/timings requirements check out first post in this thread.

Wingz, you too please :)

godsa, thanks dude... look forward to your results.

esdee, once you get the DFI would you mind posting results too please?

cheers all :toast:

justwOo

02-03-2005, 11:29 AM

Hello

Sorry I dont have any contributions, but I have a question
My prime95 keeps giving me calculation errors? Even at pc3200 speeds with the correct timings
Sometimes it will work, and sometimes not

But I've been able to do 2-2-2-4 at 220 with 3.35 volts so far

mobo wont boot above 220 :(
ordered a dfi ultra-d so it should be here wednesday then the fun begins ;P
Hopefully theres nothing wrong with the ram and its just my mobo acting funny

Wingz

02-03-2005, 11:36 AM

275 @ 3.6v

http://www.xtremesystems.org/forums/attachment.php?attachmentid=24002

Rabbi_NZ

02-03-2005, 11:46 AM

Holy sheeeeeet!
Wingz, what are your complete specs dude?
Timings, Components, Cooling...

That is amazing!

aCidbAbY

02-03-2005, 12:23 PM

275 @ 3.6v

http://www.xtremesystems.org/forums/attachment.php?attachmentid=24002

amazing bro holy sheeeeet indeed :toast: :banana4:

maybe it is because you have version 1.0?

Wingz

02-03-2005, 01:53 PM

Holy sheeeeeet!
Wingz, what are your complete specs dude?
Timings, Components, Cooling...

That is amazing!

argg rabbi your the 5th person to ask me that question and the specs are in my sig.

i have a 92mm fan attached to the zalman bracket cooling the ram.

Rabbi_NZ

02-03-2005, 01:59 PM

argg rabbi your the 5th person to ask me that question and the specs are in my sig.

i have a 92mm fan attached to the zalman bracket cooling the ram.
Could you give complete timings please?
Also, how good is your other stick?

godsa

02-03-2005, 02:55 PM

Ay guys..im trying to test out those TMII400 sticks I have and need some assistance on my timings as well

Tras 6
Trcd(read)3
trcd(write)2
trp2
cas latency 2
trc 12
trfc 14

all the rest are on auto

Scratch the 250 i just finished 255 for 15 passes..looks ok so far..260 did give errors tho

I did 250 with no errors..but I am clueless if these timing are correct or not..
3.3 v

Wingz

02-03-2005, 03:20 PM

Could you give complete timings please?
Also, how good is your other stick?

for some reason the other stick now does 260 but still does not like voltages 3.4v or above so i am now able to run both sticks @ 260 - 2-2-2-10 with 3.3v on the booster (multimeter shows 3.35v)

http://www.hvweb.co.uk/wingz/260%2Ejpg

Rabbi_NZ

02-03-2005, 03:56 PM

Would you mind testing them from 3.1v-3.6v as described in first post please? :)

pik-ard v1.1

02-03-2005, 04:07 PM

Rabbi_NZ

02-03-2005, 04:11 PM

erm... :confused:
beer distracted me

pik-ard v1.1

02-03-2005, 04:13 PM

lol, you've gotta catch one of those "sober" moments while you've got the chance. ;)

godsa

02-03-2005, 07:09 PM

Remenber this stuff was http://www.newegg.com/app/viewProductDesc.asp?description=20-218-062&depa=0

Pics of the memory and stuff all the identifying marks
http://www.xtremesystems.org/forums/showthread.php?p=696753#post696753

Ok Im done with my contribution..here ya go..and guys thanks everyone here is awesome.
All this was memtest stable for 25 passes
Tras 10
Trcd(read)2
trcd(write)2
trp2
cas latency 2
trc 10
trfc 15

I did manage to go further with the trfc 15

3.1v 254
3.2v 257
3.3v 261

Enjoy when i get the other sticks of Speed I will post as well

Specs let me c...
92mm Tornado on the chips...
DFI "LANPARTY NFII ULTRA B" stock bios..no mods
2600 Mobile (Stock Cooling)

The Memory chip numbers..

TWINMOS
512mb chips
05034
TMD7608F8Ε50D
Pic of them...
http://www.xtremesystems.org/forums/attachment.php?attachmentid=23897
TwinMOS Dual Channel Kit 184-Pin 1GB(512MB x 2) DDR PC-3200, Model TMII400/1GB - OEM

Want to add..one stick absolutely sucked...only did like 240 and thats just horrible...

MarilynMX

02-03-2005, 10:41 PM

does lower cpu multi really help in memtest?

Rabbi_NZ

02-03-2005, 10:55 PM

godsa,
thanks :)
Could you list your specs please...:
Brand & Amount of memory
Motherboard/chipset

does lower cpu multi really help in memtest?
Not for me

conrad.maranan

02-03-2005, 11:01 PM

MarilynMX - I just noticed that you're running at 3.9VDIMM. That is beyond wicked, man. :slobber: Did you measure that voltage on a multimeter to confirm it?

Rabbi_NZ

02-03-2005, 11:46 PM

Upgraded averages on first page... they will not be accurate for all platforms due to the fact they include all dimm sizes, chipsets, brands, cooling, etc... but some will find it useful :)

conrad, il pm soon as i get some food in my belly! good to see you round again bro.

MarilynMX

02-04-2005, 03:16 AM

Conrad: I dont have multimeter so I cannot find out if the volt shown on booster is accurate. My ram refuse to work at 266MHz w/ VDIMM less than 3.9V(4V idle). I know it's more than weird since I'm the only one with such experience. Another thing is booster is not affected by tunning in-BIOS VDIMM(which is totally different from my previous RMA'd booster).

PS I noticed FSB(HTT) range ocassionally varies in Corecenter. For example, I set 240x11 1:1, FSB range varies from 23xMHz to 25xMHz sometimes, that was abit freaky but same thing doesnt happen on cpu-z.

My computer reboots ocassionally without bluescreen or any other warnings. vcore 1.50~1.54, 3.3/5/12 = 3.34/5.08/12.56~12.74, VDIMM idle up to 4V depending on FSB. anyone can help me figure out the probable reason?