Printable View
As promised, the 4 CPUs i have ordered. They all have the exact same stepping information on them, the only variable is the batch #. I thought this would be a good opportunity to do a batch # comparison.
Place your bets!
i will update thread shortly, still benching these
oh i threw my CABYE in there for the picture, not part of the comparison :P
hmm pr0n
http://img.photobucket.com/albums/v7...E/94c3bc3a.jpg
:kittenhat:
Thats just like pron!:) I would say the 0246 but you never know... I look forward to your results...
i'll go for the 0351 :D
My bet goes to the 133 one.;)
oooooh ..... where's the screenies ? :D
I'll let a few more people guess and then i'll reveal which are best.Quote:
Originally Posted by eva2000
I did this quick and dirty so instead of taking screens for all of them, i just wrote down the results.
Same week and stepping usually will yield same performance.
Comparo between these prbly more like it.....
http://img146.imageshack.us/img146/4702/z0fu.jpg
http://img146.imageshack.us/img146/5697/z19ly.jpg
oh we're meant to guess i see :)
#351 is the one!
I just wanted to compare batch #s, as we currently have an idea that many buy into that states a correlation between batch # and overclockability. The results will look into that relationship...Quote:
Originally Posted by Dumo
The CABYE is my phase chip, runs 3.3 Ghz easily on my SS, thanks again to Coop!
i did similar compare with my 144 and 146 cabye (in sig) on both instances highest batch # clocked the best for meQuote:
Originally Posted by Absolute_0
well it's getting late and i want this posted
All i have is a stock heatsink, so what's what i tested on.
I took each one to the max prime stable on stock volts with stock heatsink and recorded numbers.
So here it is
batch # max clocks volts load temp
0246 2.92 1.36 50C
0154 2.89 1.36 48C
0133 2.86 1.36 52C
0351 2.82 1.36 50C
Some fun
0154 doing 3 Ghz despite HOT temps, what a trooper
http://img.photobucket.com/albums/v7...E/866de740.jpg
The 0246 doing 2.92 Ghz on stock volts / stock cooler, what an animal
http://img.photobucket.com/albums/v7...E/823bbad3.jpg
Conclusion - inconclusive
I had been thinking that either 1 of 2 things would happen
1) lower is better
2) as you go from one wafer to the next, there will be a repeat point in the pattern of overclockability. I estimated around 300, which would make 0351 the best. Since it was teh worst, and 0246 (which would have been the worst) was the best, i am either way off in my number (~300) or the whole idea is off.
It's not what you wanted to hear, but from my results, batch # doesn't mean much about overclockability...
If anyone can find any interesting multiples of these numbers (as large as possible) it might be interesting to know.
guess that 154 is about same as my 148 CABYE 0543FPMW at 3ghz on water is 46-50C load
http://www.fileshosts.com/DFI/NF4_SL...prime_1_tn.jpg
http://www.fileshosts.com/DFI/NF4_SL...urnK7_1_tn.jpg
You going to try water cooled comparison ?
i got an Arctic Freezer 64 Pro coming in soon, that's the best it will be.
But, the 0246 is doing better than my Opteron 148 CAB2E 0547GPMW did. And when i put that 148 on watercooling it did 3.1 Ghz with 1.50 volts. And 3.2 Ghz with 1.60 volts. I'm expecting good things.
When you have all this same stepping cpu`s with different batches, remember that all the cpu`s IHS attachment may be different.. One will have better contact than another and so on.. So its alittle hard to tell that a higher or lover batch is the best by looking at the OC .. I may just be the contact between IHS and core that is the reason fore one going higher than other.
Yes you could see the variation in IHS contact easily by the load temp with stock volts. While they all ran hot, the 0154 was cooler than the others.
But, these all have concave IHSs. I noticed it with my 0133 and now i notice it with all of these. The corners seem almost turned upward just the slightest bit, and the stock heatsink has a rectangular outline from the raised edges of the CPUs scraping into it.
being the owner of this guy:Quote:
Originally Posted by Absolute_0
http://eclipseoc.com/phpwcms_filestorage/1/1_69.jpg
i can vouch for that ;)
lol, i take it the 0001 wasn't as good as anticipated?
lol 0001 :D
I also thought soQuote:
Originally Posted by Absolute_0
Maybe you need 10-20 cpus to make any conclusions? And IHS removed :)
:D Not for sure - at socket A times I had some XP-M 2600 and one of them was 0035.Quote:
Originally Posted by Absolute_0
Then the reseller told me he also has the 0001 of the same stepping. I bought it and.........----------->:( - the 0035 was the better clocker.
Don't ask for screenies or numbers - too long ago............;)
:toast:
IHS couldn't have made TOO much difference. The 0246 loaded 2 degrees hotter than the 0154 yet kept priming 30 Mhz further.Quote:
Originally Posted by kiwi
Good comparision, thanks for sharing the info.
mm, it's definitly not bad, does 3ghz with 1.5v, but it's rediculously sensitive to temps. i need to get a chiller in my loop or something ;)Quote:
Originally Posted by Absolute_0
Yeah, the higher you go, the more temp-sensitive these chips become. That's why i'm having a hard time priming past 3 Ghz on the stock heatsink, I'm dealing with 55C load temps so more voltage isn't an option.
My 148 is ridiculously sensitive to temps past about 47C, anything over 49C and it'll crash quickly. I can push about 1.56V before I hit a point where I can't cool the chip fast enough with a Big Typhoon, unfortunately I didn't hit 3.0g at 1.56 :(
What kinds of results did you get out of those chips with 1.55-1.6V?
Don't know, with 1.50 volts they load at almost 60C, so obviously that is counter-productive to the stability provided by the addditional voltage.
I thought the ' substrate ' ' batch ' was the 4th, 5th, and 6th numbers on the third line
http://www.xtremesystems.org/forums/...ad.php?t=88240