AMD Deneb ES 4.0GHz this time no ****

[chaohsiangchen]

http://forum.coolaler.com/showthread.php?t=191848&page=5

No benchmark provided.

On the other news, same bunch of Taiwanese got E8600 and QX9650 to 6GHz.......

http://www.overclock.net/hardware-news/396697-coolaler-e8600-6ghz.html

 

[yomamafor1]

Already posted:

http://www.tomshardware.com/forum/254253-28-deneb

You might want to check the forum for existing threads before double posting. Thank you.

 

[chaohsiangchen]

That thread is becoming another fanboy battle front. Unlike most around here, I do read Chinese and I am not reliant on google translator to figure out what these people said on that particular forum. Also I posted current Intel OC result as comparison.

Oh, BTW, the tester seems to be a contract board tester, and he has to add CPU codes into the BIOS to do the test. So don't bother check voltage. It's his job to figure out the calibration curve.

 

[chaohsiangchen]

Also, according to the tester, not all Deneb ES goes to 4.0GHz. He got one that stopped at 3.8GHz.

 

[jimmysmitty]

^So you are saying that CPU-Zs voltage is incorrect?

And also that it may be that Deneb has the same thing as Agena and not every CPU OCs the same or at all? If thats true then that means that its possibly the architecture limiting the OCing, not the process.

 

[yomamafor1]

...and why wouldn't this thread become another fanboy thread?

To be honest, you can save THG a lot of storage space by posting in the same thread. Before you post, please check this:

http://www.youtube.com/watch?v=B9q2jNjOPdk

 

[yomamafor1]

Oh, BTW, the tester seems to be a contract board tester, and he has to add CPU codes into the BIOS to do the test. So don't bother check voltage. It's his job to figure out the calibration curve.

Actually your translation is incorrect. The board he added the code to was P35 Neo2-FR, which is a P35 board.

So the voltage is somewhat closer to what we will be seeing in the future.

By the way, you're not the only one to have the ability to read Chinese on this forum.

 

[chaohsiangchen]

Voltage and temperature of CPUs were both read from a small ADC on the board. If the resistance changes, you will have the reading way off the real value. That happens with new process is introduced.

I think it's too early to say. AMD is now churning Shanghai processors, and retail Deneb is still not in production yet. These ES were sent to MSI (all leaked Deneb ES chips were from MSI employee/contractors) probably for them to test board compatibility to 45nm process.

ES chips were made on small scale production inside design facility or a separate test line in a fab. It only tells us that some Deneb will OC to 4.0GHz.

 

[megamanx00]

Yeah but it takes a 1.6v core to get there. I don't know about that. I'll reserve judgment until you can actually buy the CPU.

 

[jimmysmitty]

Well thats all the more reason to not get our hopes up with this.

1. its a ES chip.
2. even with a CPU-Z shot, since we don't know what the main reason that Agena cannot OC very high (process,arch) we cannot say if this would even be remotely standard for Deneb at all, even the 3GHz model.

So in all its nice to see it but we will have to wait until the retail chips come out.

 

[chaohsiangchen]

...and why wouldn't this thread become another fanboy thread?

To be honest, you can save THG a lot of storage space by posting in the same thread. Before you post, please check this:

http://www.youtube.com/watch?v=B9q2jNjOPdk

To save space, that's why I don't use sig or avatar.

This thread might become another fanboy thread, but, yet, it might not. Deneb ES at 4.0GHz is now as much fact as Intel QX9650 at 6GHz. If you have nothing more worth to say, stop wasting Tom's hard drive space and find something else to grow your e-peen.

And I don't think the voltage is even right. Take a look at the next page where the other guy shows what happened at his work: AM3 test board with dual channel DDR3-800. 2.6GHz with 1.65V? That's just not right.

I brought up the CPU code thing to high light what he might be doing for life. Go ahead an be picky all the way you want. I may not be the only one who can read Chinese, but can you?

 

[yomamafor1]

To save space, that's why I don't use sig or avatar.

This thread might become another fanboy thread, but, yet, it might not. Deneb ES at 4.0GHz is now as much fact as Intel QX9650 at 6GHz. If you have nothing more worth to say, stop wasting Tom's hard drive space and find something else to grow your e-peen.

And I don't think the voltage is even right. Take a look at the next page where the other guy shows what happened at his work: AM3 test board with dual channel DDR3-800. 2.6GHz with 1.65V? That's just not right.

I brought up the CPU code thing to high light what he might be doing for life. Go ahead an be picky all the way you want. I may not be the only one who can read Chinese, but can you?

You're entirely missing the point.

You're wasting space because you could've post the same thing in a thread specifically created for this. Why create a new thread talking about the exact same thing? You guys even have the same source for Christ's sake. At the moment, you're just an attention seeker to me.

The voltage number remains to be seen. It could be the BIOS set the Vcore at 1.648 when he ran the test. It is very plausible for Deneb to get to 4.0Ghz with 1.6V based on its process node. When Intel introduced Penryn, it never encountered any problem in regards to voltage. So why would AMD have this problem?

You'll be surprised as to who can read Chinese here, mainland Chinese "friend".

 

[zenmaster]

Well thats all the more reason to not get our hopes up with this.

1. its a ES chip.
2. even with a CPU-Z shot, since we don't know what the main reason that Agena cannot OC very high (process,arch) we cannot say if this would even be remotely standard for Deneb at all, even the 3GHz model.

So in all its nice to see it but we will have to wait until the retail chips come out.

#3 - It's 1.6v. Not something you can safely run your chip at.
#4 - It's CPU-Z, which does not meant stable. If that is the max CPU-Z, then "Stable" would likely be noticably lower.
#5 - Until we get some IPC and relative numbers, it does not mean much. Sort of like trying to compare a PentiumD at 4.0 vs an E500 at 4.0Ghz.

Now, Denab could come and be great, but until they show significantly increased IPC, significantly lowered Power Usage I will remain doubtful. Especially w/o moving to the improved materials that Intel and IBM are using to reduce leakage at 45nm. (AMD will be moving to this material in 2H '09 last I read.)

 

[jimmysmitty]

Well here is something interesting:

1.45

05/10/2008 •Intel Atom Z500/Z510/Z520/Z530/Z540 processors.
•Intel Core 2 Duo E7x00 processor.
•AMD 45 nm Phenom (Dened & Shanghai) processors.
•Intel P45 chipset.
Exe file properties
•MD5 (cpuz.exe 1.45) : DD4E540C04297CF93DC721166C13446E

CPU-Z has had support for Deneb since version 1.45 and even before Core i7 even though Core i7 will probably hit retail before Deneb will.

So I can easily say that more than likely that voltage should be correct. But my only question, is why would they be testing out AM3 models when AM3 wont be out until what 2Q/3Q 09? The AM2 Denebs will hit first so shouldn't that be what we want to see for now?

 

[someguy7]

What in the world. So now a 4ghz denab=6ghz quad core penryn. I wonder where that figure came from. And all this voltage crap needs to stop. Just wait and see when the retail chips or ES chips are properly tested. All this bios code speculation and all these pipes dreams are just sad. WAIT AND SEE.

Oh and the other thread there is not a flame war going on. If anything this is the fanboy thread with all that junk.

 

[spathotan]

Stop crying about saving space, who cares. Go buy THG some webspace and a rack if it bothers you that much.

Back on topic, as the OP says the voltage probably isnt final and the OC's themselves are hit or miss as far as 4ghz is concerned. 1.6v is preety damn high, but if its stable and the chip is designed to handle it then its fine. But any of that remains to be seen. Speculating about this is preety worthless, take the CPUz screenshot and the provided link for what it is.

 

[chaohsiangchen]

You're entirely missing the point.

You're wasting space because you could've post the same thing in a thread specifically created for this. Why create a new thread talking about the exact same thing? You guys even have the same source for Christ's sake. At the moment, you're just an attention seeker to me.

The voltage number remains to be seen. It could be the BIOS set the Vcore at 1.648 when he ran the test. It is very plausible for Deneb to get to 4.0Ghz with 1.6V based on its process node. When Intel introduced Penryn, it never encountered any problem in regards to voltage. So why would AMD have this problem?

You'll be surprised as to who can read Chinese here, mainland Chinese "friend".

To be honest, anything I said in a fanboy battle front will be labelled. Why would I want to post in a fanboy battle front? I started this with Deneb ES at 4.0GHz and Intel quads at 6.0GHz for comparison to avoid being labeled.

I do whatever I wanted to post within Tom's allowance, and I have every reason not to post in fanboy battle front. As long as admin don't bother a few kilo bytes on their hard drive, I don't need to bog down to your internet moral intimidation. In the end, you just don't want to see another AMD-overclock-well thread. There are people asking dumb questions on this forum over and over and over again. And I don't see you acting like forum-double-post Nazi over and over and over again on these threads. Why only on AMD-overclock-well thread like this? Feeling sand in your eyes every time you read good about AMD? Missing which point? There is no point in your post other than trying to dissuade people trying to post anything good from AMD.

AMD's CPU design has been very different from Intel since K6. I've burnt AMD CPU but never Intel ones. I won't be surprised with these difference. Also, Intel always have better fab, while AMD's fabs are usually one generation behind. The fact that AMD worked with IBM to develop 45nm process might be the prelude for the spin-off, which rid of themselves with possible lawsuit with Intel on fab patents. You said more than once that Intel is on HKMG while AMD is still on SOI. OK, do you know the exact physical property of the chip? How it changes with shrink? Do you know the exact process and design? The thickness and crystal orientation of substrate? Time and depth of lithography process? How about diffusion, annealing and doping concentration? AMD doesn't even use Hafnium. How would Intel processors even comparable to AMD ones in regards to these engineering detail is out of my mind. And how can you even put "it didn't happen to Intel, so will be AMD."? Nonsense.

Also, 1.65V might be whatever the test board send to onboard ADC for voltage monitor, while the true voltage might be connected to a Keithley Digital Multimeter. The fact is, we don't know. If you ever use AMD product, you'll know that many retail onboard sensors may not even be right in the first place. It only reflects readings from ADC, which gives you a digitized number of the voltage measured there. If you don't have the extensive knowledge about how the thing is hooked and what is the effective impedance, the value is useless to you.

I don't see any of your Chinese friends around here. You either can or you either can't.

 

[chaohsiangchen]

Well here is something interesting:

1.45

So I can easily say that more than likely that voltage should be correct. But my only question, is why would they be testing out AM3 models when AM3 wont be out until what 2Q/3Q 09? The AM2 Denebs will hit first so shouldn't that be what we want to see for now?

Why wouldn't they? Motherboard market is very competitive, and you want to get products into the market as soon as possible. They need to do design/validation before going into designing products. They need time for prototyping, testing, EMC test and regulation related tests. After all that was done, they need to prepare production lines and make first batch as pre-production samples. That's the ones you'll see from reviews. Once the pre-production is done, they then start mass-production. The development cycles is usually half year long for motherboards. According to current leaked schedule, AMD should be launching AM3 by March 09. This should be the time Asus, MSI and Gigabyte preparing tech validation boards. If we still don't see AM3 boards by next June, then we know AMD screw up again.

I remember that I've seen prototype X58 boards back in June, and some review sites already have their Ci7 running benchs.

 

[evilshuriken]

To be honest, anything I said in a fanboy battle front will be labelled. Why would I want to post in a fanboy battle front? I started this with Deneb ES at 4.0GHz and Intel quads at 6.0GHz for comparison to avoid being labeled.

I do whatever I wanted to post within Tom's allowance, and I have every reason not to post in fanboy battle front. As long as admin don't bother a few kilo bytes on their hard drive, I don't need to bog down to your internet moral intimidation. In the end, you just don't want to see another AMD-overclock-well thread. There are people asking dumb questions on this forum over and over and over again. And I don't see you acting like forum-double-post Nazi over and over and over again on these threads. Why only on AMD-overclock-well thread like this? Feeling sand in your eyes every time you read good about AMD? Missing which point? There is no point in your post other than trying to dissuade people trying to post anything good from AMD.

AMD's CPU design has been very different from Intel since K6. I've burnt AMD CPU but never Intel ones. I won't be surprised with these difference. Also, Intel always have better fab, while AMD's fabs are usually one generation behind. The fact that AMD worked with IBM to develop 45nm process might be the prelude for the spin-off, which rid of themselves with possible lawsuit with Intel on fab patents. You said more than once that Intel is on HKMG while AMD is still on SOI. OK, do you know the exact physical property of the chip? How it changes with shrink? Do you know the exact process and design? The thickness and crystal orientation of substrate? Time and depth of lithography process? How about diffusion, annealing and doping concentration? AMD doesn't even use Hafnium. How would Intel processors even comparable to AMD ones in regards to these engineering detail is out of my mind. And how can you even put "it didn't happen to Intel, so will be AMD."? Nonsense.

Also, 1.65V might be whatever the test board send to onboard ADC for voltage monitor, while the true voltage might be connected to a Keithley Digital Multimeter. The fact is, we don't know. If you ever use AMD product, you'll know that many retail onboard sensors may not even be right in the first place. It only reflects readings from ADC, which gives you a digitized number of the voltage measured there. If you don't have the extensive knowledge about how the thing is hooked and what is the effective impedance, the value is useless to you.

I don't see any of your Chinese friends around here. You either can or you either can't.

Ever notice, that the amount of text in someone's post seems to have a very peculiar relationship with how defensive the poster is being... could almost plot that relationship in some sort of graph. This needs to be studied at once!

 

[chaohsiangchen]

Ever notice, that the amount of text in someone's post seems to have a very peculiar relationship with how defensive the poster is being... could almost plot that relationship in some sort of graph. This needs to be studied at once!

Yeah, people are always busy anyway.

 

[teh_boxzor]

he's right about one thing. all those
'please help me with a problem EVERYONE ELSE HAD and POSTED A HELP ME THREAD ON THE SAME SUBJECT....'


go be a nazi to those newbs (or help them out or create a "problems" category with the most used compnents as sub categories)

i swear every other thread i see on the front of thg's site in the forums box is a bunch of "Help overclock my intel" or "my memory won't run its timings.."

sorry just venting at the lack of newsworthy stories THG has been putting out as of late.


edit: (THG as in the people in the forums, not the website which has been very informative in their up to date news and videos.)

 

[chaohsiangchen]

Please, no voting war in this thread.

 

[apache_lives]

http://forum.coolaler.com/showthread.php?t=191848&page=5

http://www.coolaler.com.tw/toppc/AMD45NM/4G.JPG

No benchmark provided.

On the other news, same bunch of Taiwanese got E8600 and QX9650 to 6GHz.......

http://www.overclock.net/hardware-news/396697-coolaler-e8600-6ghz.html

1) My old 65nm Q6600 does 3.5ghz at ~1.43v (poor sample compared to some)
2) 1.6v will will that 45nm chip in no time, i wouldnt put my chip to even that voltage
3) The last fanboi posted that it will perform similar clock for clock to the last, so its out to loose
4) Intels have already pulled 6ghz quads, well actually dual quads - http://www.cdrinfo.com/sections/news/Details.aspx?NewsId=23057

on the other hand, its good to see AMD has something new out, 45nm will lower costs and most likely fix those internal issues with dead cores we hear about etc (well poor batches of decent high clocked quads), and AMD has always been great at taking bit by bit away from Intel etc - no big leaps but a steady gain.

 

[jimmysmitty]

he's right about one thing. all those
'please help me with a problem EVERYONE ELSE HAD and POSTED A HELP ME THREAD ON THE SAME SUBJECT....'


go be a nazi to those newbs (or help them out or create a "problems" category with the most used compnents as sub categories)

i swear every other thread i see on the front of thg's site in the forums box is a bunch of "Help overclock my intel" or "my memory won't run its timings.."

sorry just venting at the lack of newsworthy stories THG has been putting out as of late.


edit: (THG as in the people in the forums, not the website which has been very informative in their up to date news and videos.)

The biggest one I tend to see is "CPU not running at rated speed". Or as I like to call it, Speedstep Strikes Back. But for some reason its only with newbies with Intel systems and not people with AMD systems. Probably because they were used to the old dtyle P4s that ran full power all the time.

 

[yipsl]

Now, Denab could come and be great, but until they show significantly increased IPC, significantly lowered Power Usage I will remain doubtful. Especially w/o moving to the improved materials that Intel and IBM are using to reduce leakage at 45nm. (AMD will be moving to this material in 2H '09 last I read.)

Deneb. It should have increased IPC from B3, but Nehalem will probably still have better IPC. I've always found these "CPU goes to X gigahertz" tests to be as much marketing as anything else. Aside from cherry picked processors used by Intel and AMD in different generations, they're usually overclocked in ways that would limit their lifespan if continued for the period of time that real overclocker's want.

It's very much like the Futuremark benchies done with massively overclocked rigs using liquid nitrogen. I'm one AMD fan who wants real world overclocks and benchmarks that are repeatable by other sites. It's called the scientific method. Anything else is magical thinking.

That said, I've been accused of magical thinking because of my religious beliefs, but I don't think it's magical thinking to say that 45nm, even with SOI, should bring improvements to Deneb/Heka and Propus/Rana over B3 quads and triples. That's what it's all about; an improvement over the same company's prior generation. It's not a war between companies except in marketshare and the somewhat fickle hearts of enthusiasts.

Perhaps when AMD and Intel are at 32nm with new architectures, and both are using high k processes, that we'll have an apples to apples comparison, but until then, it's more like apples to pears. Intel wins when everything is averaged out, but AMD does well with certain CPU's at particular price points.

Another thing I can't wait to see is Intel trying to take on AMD and Nvidia with Larrabee. That should be either a surprise or an amusing diversion. I expect Huang's response will be a hoot any which way it pans out.