cherry picking

[Zeus]

Only 7095MHz? What a pos!

edit: From what i've seen on a german forum, the low VID chips like 1.275V for example do very well on air/water. Shame you have to unpack the cpu and drop it in to find out the default VID though.

[Mad222]

It's a new batch & has diff. behavior

BTW...guys, as Chew said "N" can be easily killed with high volt.
This one is "T"

I wish I could test a dozen same batch to find out the fastest way to bin

1.275v VID should be the lowest for Deneb I've seen so far
1.25v VID for Thuban

Only 7095MHz? What a pos!

[Sn0wm@n]

well and VID, vid is the ref voltage the board recognizes chips at.

so a chip that shows up at auto at 1.20v is super duper high leakage, and a chip showing up at 1.40 ref is super low leakage.

huh ... so more voltage = less leakage????

[eXa]

And here i thought lower VID=better (from a air/water point of view), now i know better

[chew*]

huh ... so more voltage = less leakage????

yes......its all based on tdp.

TDP is the power draw or power threshold by witch amd bins chips.

A 140w cpu must not exceed 140w draw at ref voltage, so if they are running a lower ref voltage on the chip it's to keep it under 140w power draw. If they ran higher it would exceeed the 140w rating, etc it has higher leakage.....

An example of super high leakage. The ES chips all the "PROS" were using had vid's in the 1.20-1.25v range.......Of course some of them would argue that this gave them no advantage whatsoever You show me a chip in retail with that low a vid

TWKR had high leakage to but TWKR was the rejects from the "955 ahem retails" err I mean HL ES 955 retail lookalikes that people were competing with.

[stangracin3]

so what are the weeks we want to stay away from?

bump

[[XC] Oj101]

Too much confusion...

CACAC AC 1009GPMW:
CACAC = revision code
1009 = year and week, 10th year, 9th week
G = 7th day:
A/R = first day
B/S = second day
C/T = third day
D/U = fourth day
E/V = fifth day
F/W = sixth day
G/X = seventh day
PMW = location, line, batch - I don't know further details at the moment
String below all of the above = serial number. The lower the last few digits are the closer to the center of the wafer.

Your chip was manufactured on the seventh day of the ninth week of 2010

[chew*]

bump

we already answered that question backwards.......

we told you the weeks to look for week 10 deneb, week 15 thuban........if you can deduce those are the weeks we suggest then you can deduce to avoid everything else?

[stangracin3]

so the chip i have is 1.272 vid. doesn't that mean i hae a high leakage chip? and don't high leakage chips do beter at cold than low leakage?

[chew*]

so the chip i have is 1.272 vid. doesn't that mean i hae a high leakage chip? and don't high leakage chips do beter at cold than low leakage?

ummm isn't that what we have been saying in this thread the entire time?

[skier]

what about when the VID bounces back and forth? lol mine goes from 1.3v-1.45v !

[BeepBeep2]

So, my 965BE C3 has a 1.40v VID.
However, when I first got it (it has degraded slightly, ever so slightly) I was not able to hit higher than 3.75 Ghz on 1.4v.

I needed 1.49v for 3.8 in fact, and 1.53-1.54v for 4 Ghz. (Now I need ~1.55)

Wouldn't that contradict low-leakage chips being better on air/water? I run ~4100 @ 1.6v 24/7, because I don't do anything too intensive.

*makes sadface*

[chew*]

what about when the VID bounces back and forth? lol mine goes from 1.3v-1.45v !

What does bios claim your VID as with all C@Q and turbo core disabled.......

Beep your forgetting batch/week makes the most diff.......you could have the highest or lowest leakage chips........but if the batch week sux....its still going to suck.

[BeepBeep2]

What does bios claim your VID as with all C@Q and turbo core disabled.......

Beep your forgetting batch/week makes the most diff.......you could have the highest or lowest leakage chips........but if the batch week sux....its still going to suck.

whoops.

I was nice and content with the CPUNB though, does 3 Ghz stable enough for day to day on 1.4v.

[chew*]

whoops.

I was nice and content with the CPUNB though, does 3 Ghz stable enough for day to day on 1.4v.

HL ES do that with 1.25nb

[stangracin3]

HL ES do that with 1.25nb

ES's and retails need to swap places lol

[skier]

What does bios claim your VID as with all C@Q and turbo core disabled.......

with the gigabyte board it would say 1.3v one boot, and 1.45 the next, didn't notice any pattern of what it would say per overclock, perhaps it just always said 1.3v as default when i CLR CMOS, then it goes to 1.45 on first reboot (which really sucks because Gigabyte changes voltage in increments BASED on the VID, this caused many a black screen)

this asrock board sets it to whatever voltage I want (ie, if i set vcore in BIOS to 1.375v, coretemp reads the VID as 1.375v, where the gigabyte board would only ever be 1.3 or 1.45)

the ASRock board defaults to 1.3v, and don't know if that also changes to 1.45, as I manually set all my settings to their initial defaults now because it goes by raw value, not increment based on defaults/fake VID like GA

[stangracin3]

with the gigabyte board it would say 1.3v one boot, and 1.45 the next, didn't notice any pattern of what it would say per overclock, perhaps it just always said 1.3v as default when i CLR CMOS, then it goes to 1.45 on first reboot (which really sucks because Gigabyte changes voltage in increments BASED on the VID, this caused many a black screen)

this asrock board sets it to whatever voltage I want (ie, if i set vcore in BIOS to 1.375v, coretemp reads the VID as 1.375v, where the gigabyte board would only ever be 1.3 or 1.45)

the ASRock board defaults to 1.3v, and don't know if that also changes to 1.45, as I manually set all my settings to their initial defaults now because it goes by raw value, not increment based on defaults/fake VID like GA

Turn off Turbo if you don't want to see it fluxuate that much

[skier]

Turn off Turbo if you don't want to see it fluxuate that much

i've used turbo once i think, to get a 4.496GHz validation (used cputweak with XP to get a x16 multi on core#0) other than that it stays off

[chew*]

i've used turbo once i think, to get a 4.496GHz validation (used cputweak with XP to get a x16 multi on core#0) other than that it stays off

what he is saying turbo and all of amd's little features need to be off to get a correct default vid value.

[skier]

what he is saying turbo and all of amd's little features need to be off to get a correct default vid value.

I know, on the gigabyte board it seemingly defaults to 1.3v VID once they are all off...except the one screenshot i have at 3.9GHz with 1.45v VID lol

with the ASRock it appears to stay 1.3v, here it is with the Vcore, NB and HT settings on auto and all others manually set to their defaults, turbo and technologies disabled including IO C State whatever the heck that is in the main OC settings tab

(I'm clocking from scratch now because I just got this board on saturday- and yes it idles at 19C on air without the window open, and fans are at ~medium)

[chew*]

I know, on the gigabyte board it seemingly defaults to 1.3v VID once they are all off...except the one screenshot i have at 3.9GHz with 1.45v VID lol

with the ASRock it appears to stay 1.3v, here it is with the Vcore, NB and HT settings on auto and all others manually set to their defaults, turbo and technologies disabled including IO C State whatever the heck that is in the main OC settings tab

(I'm clocking from scratch now because I just got this board on saturday- and yes it idles at 19C on air without the window open, and fans are at ~medium)

so your vid is kinda high, thuban goes down to 1.250.

You have a middle range chip.

[I.nfraR.ed]

Too much confusion...

CACAC AC 1009GPMW:
CACAC = revision code
1009 = year and week, 10th year, 9th week
G = 7th day:
A/R = first day
B/S = second day
C/T = third day
D/U = fourth day
E/V = fifth day
F/W = sixth day
G/X = seventh day
PMW = location, line, batch - I don't know further details at the moment
String below all of the above = serial number. The lower the last few digits are the closer to the center of the wafer.

Your chip was manufactured on the seventh day of the ninth week of 2010

And then what MPMW stands for? My chip is 1010MPMW and it has quite good CPU-NB. I was thinking that the first letter is the weekday, but couldn't figure out what is M for...


Cores are not so weak, either. The number of the cpu is 546. I'm not sure about the default VID, but it's no more than 1.28V. I'm currently running my lapped 955, so I can't check.



One of the 965 I had was good for benchmarks, but weak for 24/7. It loved voltage, but was unstable over 4.1GHz on water. It's serial number was 44.

[cadaveca]

MPMW is multi-wafer batch.

[NoM$_YesLinux]

MPMW is multi-wafer batch.

MW = Multi-wafer, ok, but what is AW?



This 1090T from pic are very similar to I.nfraR.ed one for nb oc on air and is capable to run wPrime 32m @ 6200 on LN2. I am not sure about VID, but i think it's 1.3v, will confirm this later.