cherry picking

[demonkevy666]

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

You have a middle range chip.

I don't think so


this site seems pretty useful HERE
the thing that struck my eye here was this

Electrical/Thermal parameters
1.15 - 1.475 (at rated frequency)
so I tested it

Mine can go as low as 1.15 volts at stock clocks stock vid is 1.275 for this chip
but I noticed it's horribly slow at cinebench 11.5r I only scored 4.62 at those voltages with everything else off too.
Normally I can score 5.66 in cinebench 11.5r at stock vid and clock.
overclock 4.2 at 1.55 volts score 7.45 in cinebench

(note nothing was running in background when running cinebench 11.5r was ran)

[chew*]

I don't think so


this site seems pretty useful HERE
the thing that struck my eye here was this

Electrical/Thermal parameters
1.15 - 1.475 (at rated frequency)
so I tested it

Mine can go as low as 1.15 volts at stock clocks stock vid is 1.275 for this chip
but I noticed it's horribly slow at cinebench 11.5r I only scored 4.62 at those voltages with everything else off too.
Normally I can score 5.66 in cinebench 11.5r at stock vid and clock.
overclock 4.2 at 1.55 volts score 7.45 in cinebench

(note nothing was running in background when running cinebench 11.5r was ran)

It's not how low your chip can go. It the hard VID that AMD gave the chip that matters in relation to tdp.

[stangracin3]

It's not how low your chip can go. It the hard VID that AMD gave the chip that matters.

any idea what the avaerage VID is?

[chew*]

any idea what the avaerage VID is?

To know that would require excessive binning, something i'm not fond of.


I know the lowest vid i have touched was 1.15......of couse HL ES that i got from amd c2 rev x4......

I think AMD knows how to deem a chip HL , in fact nothing HL i ever got had a vid over 1.25.

Of course i won't use these chips, they are just for "play"

[Tomasis]

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.....

it makes sense!

[quake6]

This info is so good Was looking for something like that, but I thought it isn't good to create thread like this...

And what about intel cpu? (I know it is AMD section, but just a few info about it. If it is completly wrong, I'll delete this post)

[BeepBeep2]

From what I've seen in the forum it seems for intel CPU's that the batch number (3009F988) carries most of the info. Certain batches seem to be better than others.

[chew*]

Quake vid can play a role in intel as well, same idea lower vid denotes higher leakage.

My q6600 is semi stellar, default VID is 1.225, I run undervolted at 1.20 at 3.0 24/7.

My 920 is also stellar I forget vid but its rather low on the scale.

[quake6]

From what I've seen in the forum it seems for intel CPU's that the batch number (3009F988) carries most of the info. Certain batches seem to be better than others.

Really? heh
I know that, but what I'm wondering is what 988 means? Is that also how far from the middle of the wafer the chip is or?

[AlleyViper]

At the time I got my c2 955 I went for a 0911epmw (iirc) with 00006 in the end. VID is quite low at 1.325V, but at 1.475V it's hot as heck to hold in a decent case, paired with good air cooling. I'd guess a high leakage chip would scale well with low added voltage due to tpd ramping, but this c2 needs 1.375V for 3.6, 1.425 for 3.7 and 1.475 for 3.8 that is the limit for my cooling to keep it under 55ºC in medium/cold weather while stress testing, so I must resort to running it at 3.6 for all year 24/7.

[[XC] Oj101]

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

A = first day of the week

[cab]

Interesting, in that case...

my 965 is 1023FPMW wafer #1074
my 555 is 1008FPMW wafer #0496

The 555 unlocks to a quad, performs identically to the 965, but runs cooler and needs less voltage for the same overclock. So is it possible that wafer 496 is always better than 1074?

[skier]

Interesting, in that case...

my 965 is 1023FPMW wafer #1074
my 555 is 1008FPMW wafer #0496

The 555 unlocks to a quad, performs identically to the 965, but runs cooler and needs less voltage for the same overclock. So is it possible that wafer 496 is always better than 1074?

your 965 is a much worse week (23 vs 8 for the 555), as mentioned before the week is more important than its place on the wafer, though the 965 is worse in both respects

[Tom128]

Just jumping in on the VID topic. Cant the CPU's vid be up to the motherboard as well? Or more correctly, the motherboard reading the VID correctly? I only ask because I used to have a 955BE and ran it in my M3A79-T Deluxe and it had a VID of 1.3v, and I then moved it over to a GA-MA790XT-UD4P and the VID was 1.35v. Same chip, two different VID's. It was consistently these VID's, they didn't move around, it was just different. Had C&Q/ACC/CE1/everything off in both.

[Zeus]

I have experienced something similar after i flashed my bios (!) can't remember which board/cpu, might as well have been an Intel setup.

[demonkevy666]

Just jumping in on the VID topic. Cant the CPU's vid be up to the motherboard as well? Or more correctly, the motherboard reading the VID correctly? I only ask because I used to have a 955BE and ran it in my M3A79-T Deluxe and it had a VID of 1.3v, and I then moved it over to a GA-MA790XT-UD4P and the VID was 1.35v. Same chip, two different VID's. It was consistently these VID's, they didn't move around, it was just different. Had C&Q/ACC/CE1/everything off in both.

Memory controller Change from DDR2 to DDR3 Different voltage for Ram and probably stability.
check NB Voltage then too it will probably be different too then, it still had to fit in the TDP with the change.
VID is also variable in Turbo mode form 1.40-1.450(from what I've seen) to stay in TDP, when that's happening it's not using all cores too.

[BeepBeep2]

Memory controller Change from DDR2 to DDR3 Different voltage for Ram and probably stability.
check NB Voltage then too it will probably be different too then, it still had to fit in the TDP with the change.
VID is also variable in Turbo mode form 1.40-1.450(from what I've seen) to stay in TDP, when that's happening it's not using all cores too.

I think it's just the bios misreading it.

CPU-NB volt is stock at 1.1v (at least they were on my CPU) regardless of board its in.

[cab]

your 965 is a much worse week (23 vs 8 for the 555), as mentioned before the week is more important than its place on the wafer, though the 965 is worse in both respects

Thanks... I find it hard to believe that AND made a good batch back in February and nothing worthwhile since then. How many CPUs are there in a wafer?

[skier]

Thanks... I find it hard to believe that AND made a good batch back in February and nothing worthwhile since then. How many CPUs are there in a wafer?

a guess would be around 1100 since i've not seen any above that, but i'm sure every wafer comes out slightly different in terms of the amount of usable dies

just helped a friend install a 945 and it was from week 35, with # 1045 i think

[stangracin3]

The ideal number would be 1095, but since a wafer is a circle some of the cores get cut off