- Nov 18, 2001
- 13,234
- 2
- 81
I just got mine setup. Running at stock speed/voltage with a Tuniq Tower at about 27C idle. Going to stress it overnight to see how hot it will get, then I'm going to start overclocking. Hoping to get 3.6GHz.
- Sep 28, 2005
- 21,052
- 3,533
- 126
You guys now see why i kept pushing people to get the E6420?
This is the exact reason i kept telling people. All the NEW E6600, are a disapointment.
A lot of people think intel tweeked their output, so quality is no longer as good as it used to be. Also, all the good E6600's are being fed to the Q6600's. Unless your uber lucky and get a B3 Q6600. I hear its not hard to get one right now, however come august you'll be SOL.
And that guy who cant get his Q6600 to clock faster then 333. Have you up'd your NB voltage? Also the 680i NB cooler is very inadaquit. I noticed a hugh increase in stability once my NB and SB were under water.
This is the exact reason i kept telling people. All the NEW E6600, are a disapointment.
A lot of people think intel tweeked their output, so quality is no longer as good as it used to be. Also, all the good E6600's are being fed to the Q6600's. Unless your uber lucky and get a B3 Q6600. I hear its not hard to get one right now, however come august you'll be SOL.
And that guy who cant get his Q6600 to clock faster then 333. Have you up'd your NB voltage? Also the 680i NB cooler is very inadaquit. I noticed a hugh increase in stability once my NB and SB were under water.
lopri
Elite Member
- Jul 27, 2002
- 13,310
- 687
- 126
Ah.. this theory makes sense indeed. I was wondering what's up with all the moans about new E6600s.
One more thing that I've been very curious about: Where are the year 2007 chips? If my understanding is correct, the batch number beginning with L6xx refers to year 2006 week xx? I have never seen anyone talking about L7xx batch? It's strange because 4 months have passed since the beginning of 2007.
magreen
Golden Member
- Dec 27, 2006
- 1,309
- 1
- 81
I was thinking maybe that in the beginning of C2D, Intel purposely didn't bin their chips so that even lower end chips would OC to insane levels. That way they'd get all the enthusiasts over to their side and away from AMD. And only later (i.e. now) do they start binning.
Let's look at it like this: I'm assuming that the max speed to which silicon from a given process will clock will be distributed on a bell curve like most other things in life. So let's say their Conroe silicon is distributed on a bell curve centered at 3.4 GHz, with a standard deviation of 0.4 GHz, i.e. a significant number of chips go up to 3.8 and a significant part only reach 3.0 GHz. So in the beginning they don't bin, they just take some of them and package them as E6300s, some as E6400s, etc. That way, a significant number of people who only buy a E6300 or E6400 or E6600 will get silicon that clocks up to 3.6+ GHz, and will share their amazing OC stories on forums. And Intel sends hand-picked engineering samples to tech sites like AT who overclock them to great heights. All the enthusiasts start clamoring for these things and AMD loses its luster to enthusiasts. Even though a lot of E6600s don't clock to 3.6+ (remember the curve is centered at 3.4), the people who own the 40% or so that clock well are a lot louder.
Then a year later once all the enthusiasts have ditched AMD, they start binning, since binning is best for profits. Silicon that clocks to 3.6GHz goes to high-end E6700 and X6800 and QX6700 chips, so they have plenty of headroom -- or it's even saved for later use in yet-to-be-released 3GHz chips. Silicon that only clocks to 3.2GHz goes to E6600s, since it still leaves 33% headroom over 2.4GHz stock. And they know that enthusasts only get their info based on individual customer testimonials and rumors and hearsay on forums (I got 3.8GHz on air on my E6600, this thing rocks!!!), not on solid statistics or solid insider information on changes in manufacturing and marketing.
The only thing left is what about the E6320 and E6420? They might want to start by using unbinned silicon. That way the enthusiast crowd will be excited by them since a good number will clock to 3.6+ GHz, and they'll stay loyal to Intel, even though AMD's price cuts make their chips look somewhat decent right now. Once the E6320 and E6420 get a really good reputation for OCing, Intel might bin those, too.
I'm just guessing here, people. What do you think of my theory?
Let's look at it like this: I'm assuming that the max speed to which silicon from a given process will clock will be distributed on a bell curve like most other things in life. So let's say their Conroe silicon is distributed on a bell curve centered at 3.4 GHz, with a standard deviation of 0.4 GHz, i.e. a significant number of chips go up to 3.8 and a significant part only reach 3.0 GHz. So in the beginning they don't bin, they just take some of them and package them as E6300s, some as E6400s, etc. That way, a significant number of people who only buy a E6300 or E6400 or E6600 will get silicon that clocks up to 3.6+ GHz, and will share their amazing OC stories on forums. And Intel sends hand-picked engineering samples to tech sites like AT who overclock them to great heights. All the enthusiasts start clamoring for these things and AMD loses its luster to enthusiasts. Even though a lot of E6600s don't clock to 3.6+ (remember the curve is centered at 3.4), the people who own the 40% or so that clock well are a lot louder.
Then a year later once all the enthusiasts have ditched AMD, they start binning, since binning is best for profits. Silicon that clocks to 3.6GHz goes to high-end E6700 and X6800 and QX6700 chips, so they have plenty of headroom -- or it's even saved for later use in yet-to-be-released 3GHz chips. Silicon that only clocks to 3.2GHz goes to E6600s, since it still leaves 33% headroom over 2.4GHz stock. And they know that enthusasts only get their info based on individual customer testimonials and rumors and hearsay on forums (I got 3.8GHz on air on my E6600, this thing rocks!!!), not on solid statistics or solid insider information on changes in manufacturing and marketing.
The only thing left is what about the E6320 and E6420? They might want to start by using unbinned silicon. That way the enthusiast crowd will be excited by them since a good number will clock to 3.6+ GHz, and they'll stay loyal to Intel, even though AMD's price cuts make their chips look somewhat decent right now. Once the E6320 and E6420 get a really good reputation for OCing, Intel might bin those, too.
I'm just guessing here, people. What do you think of my theory?
The e6320/6420 are no different than the e6600's that have been produced since week 35-40+. The overclocks for all of the later week chips from all batches has been all over the map - some early reports of 6320/6420 O/C's were equally dismal to the late model e6600's. Inversely I've also heard of G core e6600's pushing the 3.8 ghz barrier over at HardOCP... that doesn't mean that all e6600's are better than the e6420's, etc... It's simply a "blip" example just like the monster 6420 overclocks that some have reported. Until someone can produce a significant amount of statistical OC result data that actually indicates a trend one way or another it's irresponsible for anyone to spout-off that either is a better overclocker.
The first 25 weeks of Conroe cores were over-engineered and produced in a single fab while they got their methods dialed in for the entire Conroe core C2D and Quad line of chips. This is why the early 6300/6400's were produced on the 4-meg die's and had 2 megs of cache disabled in assembly, all of the cores could be produced in a single fab from the same wafers.
Once Conroe production methods and tolerances were established they developed the Allendale line for their 2 meg parts (cutting costs) and undoubtedly applied some of their lessons learned to the 4-meg Conroe line in the process. This corresponds with their preparation for the current price war with AMD and the change in model numbers for the e6600/e6700 on Intel's site around the same time the F and G production codes started showing up on e6600's.
The current generation 6320/6420/6600/6700 are cut from the same wafers produced in one of Intel's four 65nm fabs. They all get sent to Malaysia where they are tested, binned and packaged depending on economic demand for each model. It's possible that a particularly good batch of chips will get binned "down" to meet demand on occasion, but the days of almost always getting an over-engineered Conroe core are over. There is no "overclocking conspiracy," Intel found ways to produce Conroe chips less expensively which reduces the price for consumers but consequently limits overclocking capability for the enthusiast crowd.
You get what you pay for...
All Conroes I have seen now, and before clock about the same.
Some are going to be better than others, but if you can't get over 3.0 ghz you are simply doing something wrong. I have had F, G, B, A. There is not a huge difference between all of them. 47G clocks like hell actually. Went to 3.5 ghz on 1.37v actual.
The different letters means that Intel is changing the drive of the cpu. IF they are lowering drive to alleviate heat, and lower wattage you need to do one of two things.
A.) Buy a motherboard that lets you change drive (dfi)
B.) Raise Voltage to compensate. If drive is lower you can use a higher voltage and still have lower temperatures. Drive is going to wreck your cpu a lot faster than voltage.
Some are going to be better than others, but if you can't get over 3.0 ghz you are simply doing something wrong. I have had F, G, B, A. There is not a huge difference between all of them. 47G clocks like hell actually. Went to 3.5 ghz on 1.37v actual.
The different letters means that Intel is changing the drive of the cpu. IF they are lowering drive to alleviate heat, and lower wattage you need to do one of two things.
A.) Buy a motherboard that lets you change drive (dfi)
B.) Raise Voltage to compensate. If drive is lower you can use a higher voltage and still have lower temperatures. Drive is going to wreck your cpu a lot faster than voltage.
You people think you know everything because you have yet to receive a bad 6600 in hand to play with.
f4phantom2500
Platinum Member
- Dec 3, 2006
- 2,284
- 1
- 0
Yeah; they aren't saying that EVERY one in that batch sucks, just that they tend to not overclock very well. So, you guys probably just got lucky. Considering how many processors Intel makes, it is not surprising that we would hear success and horror stories from a lot of different people on the same batches, we would just hear more of one or the other depending on how the batch actually is (which is why we can presume that most of them are good or bad in a given batch).
As I said in the other thread, my L645G does 3.2GHz at 1.37V (actual, 1.4 BIOS) and 3.4 at 1.47V (actual, 1.5 BIOS). It's surprising to see users having trouble getting over 3GHz even with 1.5V. Are you bumping your VDimm, VFSB, VNorthbridge etc? What about setting PCI-e freq to 110 vs 100?
No, thats not what we are saying at all. We are saying most of the people that are having troubles, have no idea what they are doing
.It is only a miracle that others seem to not have problems.
Send me your chip and I will show you a 3.6 ghz stable
.
- Nov 18, 2001
- 13,234
- 2
- 81
I'm at 3.0 right now with default voltage. Load temp is bouncing around between 45 and 49. Tried 3.6 at 1.5v earlier and it wouldn't boot up. 3.3 at default boots but dies in Windows after about 10 minutes of Orthos. I have a feeling there are some other settings I could change that would help stability, but I've been out of the game for 2+ years so I don't know all of the tricks.
- Sep 28, 2005
- 21,052
- 3,533
- 126
You guys:
Incase you wanted to know. Ive been getting a lot of pm's lately on Chip recomendations:
The reason why i recomended the 6420, is because your chancs at getting a earily week was their.
Now people are concerned about the Q6600. I would be, at 266 its a great deal, however, if it cant break 3.4ghz, to me its worthless. I should of just waited for the penryn's to come out.
However, if you were to get a good overclocking Q, then getting a penryn would make no senes. It doesnt show that much of an increase.
The guy that got the F batch. Dont worry if its a xeon. DDT says almost always a xeon is usually higher bin'd.
So i did some research for you guys, as well as ask some quad pro's over at XS Forums
This is what DDTung recomended to me a Xeon 3220, and this is your best chance at a high overclocking processor:
If you can find this tho:
QX6600 <--- most of these i believe are ES chips rebin'd to Quads. If you can get one of these, you'll be good to go. However if your USA, good luck finding one. I did find 1 in Canada, but no sales to US. <---- type in pricegrabber.
If your US:
Your best shot at a decient overclocking Quadcore is the Xeon 3220.
Currently TheNerds.net
http://www.thenerds.net/index.php?page=productpage&pn=BX80562X3220 11 total
That xeon is a kentsfield, and xeon's are usually higher bin'd then Q's So if the 40 dollar extra is something you dont mind at a higher chance of getting a better chip, thats my 2 cents.
And im not pushing anyone to buy anything. This is for people that mearly want my advice. And yes i got a xeon, and it should come in the mail sometime this week. I'll be sure to let you guys know how well she does so the 11 thats left can be up for a turkey shootout.
Incase you wanted to know. Ive been getting a lot of pm's lately on Chip recomendations:
The reason why i recomended the 6420, is because your chancs at getting a earily week was their.
Now people are concerned about the Q6600. I would be, at 266 its a great deal, however, if it cant break 3.4ghz, to me its worthless. I should of just waited for the penryn's to come out.
However, if you were to get a good overclocking Q, then getting a penryn would make no senes. It doesnt show that much of an increase.
The guy that got the F batch. Dont worry if its a xeon. DDT says almost always a xeon is usually higher bin'd.
So i did some research for you guys, as well as ask some quad pro's over at XS Forums
This is what DDTung recomended to me a Xeon 3220, and this is your best chance at a high overclocking processor:
If you can find this tho:
QX6600 <--- most of these i believe are ES chips rebin'd to Quads. If you can get one of these, you'll be good to go. However if your USA, good luck finding one. I did find 1 in Canada, but no sales to US. <---- type in pricegrabber.
If your US:
Your best shot at a decient overclocking Quadcore is the Xeon 3220.
Currently TheNerds.net
http://www.thenerds.net/index.php?page=productpage&pn=BX80562X3220 11 total
That xeon is a kentsfield, and xeon's are usually higher bin'd then Q's So if the 40 dollar extra is something you dont mind at a higher chance of getting a better chip, thats my 2 cents.
And im not pushing anyone to buy anything. This is for people that mearly want my advice. And yes i got a xeon, and it should come in the mail sometime this week. I'll be sure to let you guys know how well she does so the 11 thats left can be up for a turkey shootout.
Originally posted by: aigomorla
You guys:
Incase you wanted to know. Ive been getting a lot of pm's lately on Chip recomendations:
The reason why i recomended the 6420, is because your chancs at getting a earily week was their.
Now people are concerned about the Q6600. I would be, at 266 its a great deal, however, if it cant break 3.4ghz, to me its worthless. I should of just waited for the penryn's to come out.
However, if you were to get a good overclocking Q, then getting a penryn would make no senes. It doesnt show that much of an increase.
The guy that got the F batch. Dont worry if its a xeon. DDT says almost always a xeon is usually higher bin'd.
So i did some research for you guys, as well as ask some quad pro's over at XS Forums
This is what DDTung recomended to me a Xeon 3220, and this is your best chance at a high overclocking processor:
If you can find this tho:
QX6600 <--- most of these i believe are ES chips rebin'd to Quads. If you can get one of these, you'll be good to go. However if your USA, good luck finding one. I did find 1 in Canada, but no sales to US. <---- type in pricegrabber.
If your US:
Your best shot at a decient overclocking Quadcore is the Xeon 3220.
Currently TheNerds.net
http://www.thenerds.net/index.php?page=productpage&pn=BX80562X3220 11 total
That xeon is a kentsfield, and xeon's are usually higher bin'd then Q's So if the 40 dollar extra is something you dont mind at a higher chance of getting a better chip, thats my 2 cents.
And im not pushing anyone to buy anything. This is for people that mearly want my advice. And yes i got a xeon, and it should come in the mail sometime this week. I'll be sure to let you guys know how well she does so the 11 thats left can be up for a turkey shootout.
The only higher binned Xeons are Clovertown low voltage.
The Xeon 775 is the same exact chip only with a different warranty. Intel has publicly stated this...as servers may run 5 years they offer a 5 year warranty...
Unless you buy a 50w quad, you are getting the same chips.
Got a E6600 from ZZF last week (L644G425). Not much trouble in OC'ing to 3.4GHz @ 1.43V. Getting it to 3.6GHz took longer. Had to use 1.52V plus upping all the other voltage slightly. Gonna let it burn in at that level, and then step down voltage gradually from there. I'd say it's a moderate batch at best.
Forgot to mention Coretemp is 50C at full load.
Forgot to mention Coretemp is 50C at full load.
I got my x3220 from buy.com. $570 and free shipping.
Haven't got it set up yet due to ZZF mixing up ram part numbers on their website. Will report what I can do with it though, probably this weekend. Mostly I'll be concentrating on running silent, and I don't think I have the cooling for anything much over stock. we'll see though.
Haven't got it set up yet due to ZZF mixing up ram part numbers on their website. Will report what I can do with it though, probably this weekend. Mostly I'll be concentrating on running silent, and I don't think I have the cooling for anything much over stock. we'll see though.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 21K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter