AMD Shanghai is officially out

[informal]

At low speeds yes. At 2.8 and 3Ghz its already 125W TDP again. The new limit just moved abit upwards. But it behaves just as with 65nm. Try a lower clocked 65nm Phenom/Barcelona and you see alot lower power consumption.

Can we just wait for Deneb OCing(for overvolt tests) to see how it actually behaves?Did you miss the toppc report at coolaler forum on Deneb C1 OCing?The dude hit 4.2Ghz on air.For a "leaky" (in your words) part it was doing quite good.
Also if you missed the Deneb C0 early preview from August,you could see it practically halved the power draw while running at the same measly clock of 2.3Ghz.
I'm reserving my final judgement on power draw at both stock and OC for a reviews day.

Guys this might surprise you but according to Anandtech Deneb has 8MB OF L3 CACHE : http://www.anandtech.com/cpuchipsets...7&p=2#comments

WTF ; IF ITS TRUE THEN AMD fooled us all

No it doesn't... The slide says 8 MB of cache but a genius who wrote the article wrote 8MB of L3 cache.It's anandtech...

[Blaber]

At low speeds yes. At 2.8 and 3Ghz its already 125W TDP again. The new limit just moved abit upwards. But it behaves just as with 65nm. Try a lower clocked 65nm Phenom/Barcelona and you see alot lower power consumption.

To compare you can buy 3.4 and 3.5Ghz 45nm Xeons.

1.5V for 3.6Ghz...

Are you sure cpu-z is reading volts on cpu correctly ????

[Blaber]

8 MB of cache[/B] but a genius who wrote the article wrote 8MB of L3 cache.It's anandtech...

K , so its L2 of 2MB + L3 of 6MB == 8MB , correct ???

[Helmore]

K , so its L2 of 2MB + L3 of 6MB == 8MB , correct ???

Yup, just like Nehalem has 8MB of total usable cache.

[informal]

If you want to know how "hot"(or rather cold) the C0 was at very high overvolt and OC(a locked multi 2.2GHz chip that hit a wall due to HTT limit of the board) ,look here:


OC temp at above overclock(vcore is 1.56V,air cooled;orthos large in place FFTs test used,priority 9):

CPU超频，主板目前对CPU支持还不够好，表现为外频限制在265MHz左右，超过会开机不亮机，还好倍 频没锁能超倍频。CPU加压到主板上限1.568V，外频252倍频13主频3.277G，H T 2269MHz，待机温度只有28度。

and idle temp:


So at those volts and on air it's quite "hot".And it was C0,the very first part produced.Imagine what will retail C2 do or C3(AM3+DDR3) when it is released.From Shanghai reviews we can see that C2 is really doing awesome job.

[Shintai]

Here are some Deneb vs Agena benches btw. And an Oced Deneb at 3.2Ghz with numbers.

The IPC gains are very low as expected. Soit seems Deneb is finally on pair with 65nm Core 2s.

[informal]

Yeah he picked it up now and we have been discussing those since August.But he has been slow before,remember he just found out that there will be 2 versions of Deneb(even though it's know fact for a good while).
The tested chip is the very first 45nm part produced and was "healthy enough" to be sent for testing and validation.
And what's funny is that even with those high volts the first C0 step. chip is very cool.It halved the power draw practically while performing 8% better on average.The 2.3Ghz ES had a 1.8Ghz NB clock,lower than a 2.2GHz one the retail C2 Denebs will have.Not to mention AMD could improve the chip a bit since this one(C0) did perform oddly in number of tests under Vista SP1(very low multicore score under WInrar,around 550..The tester mentioned a "memory bug" a few times at itocp forum where the C0 was first tested).

[Shintai]

Yeah he picked it up now and we have been discussing those since August.But he has been slow before,remember he just found out that there will be 2 versions of Deneb(even though it's know fact for a good while).
The tested chip is the very first 45nm part produced and was "healthy enough" to be sent for testing and validation.
And what's funny is that even with those high volts the first C0 step. chip is very cool.It halved the power draw practically while performing 8% better on average.The 2.3Ghz ES had a 1.8Ghz NB clock,lower than a 2.2GHz one the retail C2 Denebs will have.Not to mention AMD could improve the chip a bit since this one(C0) did perform oddly in number of tests under Vista SP1(very low multicore score under WInrar,around 550..The tester mentioned a "memory bug" a few times at itocp forum where the C0 was first tested).

C1 Shanghai/Deneb is 2000Mhz NB...

Make some more fake news. We need it!

It was 2.2 and 2.3Ghz versions btw used at Coolaler. Oh...and 2Ghz NB

Keep betting on your babelfish/googletranslations. Easier just to read it right from the source

[informal]

The one you linked (those graphs) is C0.Don't look like a fool while trying to post most outdated stuff on Deneb.You looked like a fool numerous times in the past(yesterday was the last one).It's getting old you know.


Link to the graphs you posted:
http://en.hardspell.com/doc/showcont...58&pageid=3149

Yeah C0 was 2.3Ghz ES with 1.8Ghz NB(math is your friend since 2269/252=9 =>NB multi is 9 and hence NB runs at 1800MHz=>your claims are busted my friend).
Oh btw i never said toppc used C0 exclusively ..I said he used both C0 and C1.The pictures i linked are from C0 if you can't read or see properly:

[Shintai]

And lets just post it all.

[duploxxx]

C1 Shanghai/Deneb is 2000Mhz NB...

Make some more fake news. We need it!

no its not, all is depending on CPU speed even on C1, you will see 2,0-2,2-...(SE) on shanghai, he does not have to fake it.

proof: http://it.anandtech.com/IT/showdoc.aspx?i=3456

and deneb is a different nb again depending on speed.

btw what is the added value of the post with screenshots accept for thread filling and post count for you? its a thread started for shanghai and we have 2 pages full of deneb results.

[gallag]

@Shintel
Yeah so what?Look at core temp for ondie sensor readings if you prefer that way(28C is more reliable). You just can't come to terms Deneb/Shanghai is a great CPU ,can't you.
No,it's leaky,slower than Agena and can't 1Mhz over stock clock.You satisfied with that?

Are you basing you opinion of denebs success on unproven sources or have you seen some trusted reviews? To me it looks like deneb will have great power consumption but will only match Intel's third best 65nm cores. It looks like deneb will be further behind nehelam than Barcelona was behind penryn.

[Shintai]

@Shintel
Yeah so what?Look at core temp for ondie sensor readings if you prefer that way(28C idle is more reliable than 36C DMM shows?). You just can't come to terms Deneb/Shanghai is a great CPU ,can't you.
No,it's leaky,slower than Agena and can't OC 1Mhz over stock clock.You satisfied with that?

You dont put it FAR OUT to one side. So it doesnt block for your heatsink.

You cave it down instead in the middle of the IHS. Not pretty. But thats the way.

[informal]

How about a screen from HW monitor app?
http://detail.zol.com.cn/picture_ind...x2148368.shtml

Full load with orthos on C0 ES @ 2.3Ghz was showing a 30 degrees Celsius on air cooler.It's inline with what Coretemp shows and DMM measures.You can believe it or not,your choice really.

Are you basing you opinion of denebs success on unproven sources or have you seen some trusted reviews? To me it looks like deneb will have great power consumption but will only match Intel's third best 65nm cores. It looks like deneb will be further behind nehelam than Barcelona was behind penryn.

I disagree with most of your post so i won't go in detail since there is no point.We can just wait for reviews of C2 retail stepping.

[Shintai]

Do you honestly not believe he has a point about the temps, Look at the way they are reading them lol.

Forget it. Its a waste of time. They are in 100% denial. 45nm K10 only sips power, 20% higher IPC! and everything clocks 4ghz and being cold as the grave.

But ye, finally 65nm Core 2s got competition. And one with much greater power consumption than the Q6600

How about a screen from HW monitor app?
http://detail.zol.com.cn/picture_ind...x2148368.shtml

Full load with orthos on C0 ES @ 2.3Ghz was showing a 30 degrees Celsius on air cooler.It's inline with what Coretemp shows and DMM measures.You can believe it or not,your choice really.

I disagree with most of your post so i won't go in detail since there is no point.We can just wait for reviews of C2 retail stepping.

And you are sure the software reads correctly. Specially after Coretemp shows 28C always....

Not even to talk about the MEASURED temperature on the EDGE of the CPU was HIGHER?
http://forum.coolaler.com/showthread...191848&page=14

[gallag]

We can just wait for reviews of C2 retail stepping.

True that, And I will be happy to be wrong.

[informal]

Forget it. Its a waste of time. They are in 100% denial. 45nm K10 only sips power, 20% higher IPC! and everything clocks 4ghz and being cold as the grave.

No no,the Pentium D's are finally getting some stiff competition,being matched by IPC and power draw with Deneb!!!!! lol

And you are sure the software reads correctly. Specially after Coretemp shows 28C always....

Not even to talk about the MEASURED temperature on the EDGE of the CPU was HIGHER?
http://forum.coolaler.com/showthread...191848&page=14

28C is conspicuosly confirmed by the DDM reading..There is a delta between teh two,but HW monitor uses registry reading.So you can use more strawmen arguments.
BTW if you look closer you can see the wire going inside the thermal paste not just sitting at the edge...

[Bellisimo]

Do you honestly not believe he has a point about the temps, Look at the way they are reading them lol.

he may have some point about the temp reading, but check informal post above mine

secondly, he is trying to show deneb/shanghai in bad daylight (again btw)

Thats due to cache. And not that much faster after you remove the extra clock. Plus remember this is FBDIMMs. Not Regular.

Deneb is january 8th (Atleast embargo) on the last AMD slide.

slide from april 9th

Just a note...on desktop. a 9950 does 2930 in 1 CPU score at 2.6Ghz....

Was this a B2 TLB bugged one?

Q6700 does around 3100 and Q9450 3250 with DDR2.

All 1 CPU.

wrong info as stated by other members

At low speeds yes. At 2.8 and 3Ghz its already 125W TDP again. The new limit just moved abit upwards. But it behaves just as with 65nm. Try a lower clocked 65nm Phenom/Barcelona and you see alot lower power consumption.

To compare you can buy 3.4 and 3.5Ghz 45nm Xeons.

1.5V Deneb C1 3.6G Stable air. Deneb C1 uses 1.368V for 3Ghz

And 1.6V for 3.8Ghz.

making predictions about power consumption based on the C0 revision

Here are some Deneb vs Agena benches btw. And an Oced Deneb at 3.2Ghz with numbers.

The IPC gains are very low as expected. Soit seems Deneb is finally on pair with 65nm Core 2s.

really old benchmark with C0 revision, again he thinks the C0 is representative for final deneb revision

C1 Shanghai/Deneb is 2000Mhz NB...

Make some more fake news. We need it!

no its 2.2 GHz

Oh...informal...where did they measure that temp again? A hint is its not the CPU itself

After losing all battles, he turns on the temp monitoring, as a last attempt to prove that deneb sucks....



it's really sad to see...

[Katanai]

lol When in his first post Shintai actually said something nice about an AMD chip I knew what was to come. I just have one question for you guys, who maybe know more about this: does Intel even have a division in Denmark?

[Bellisimo]

lol When in his first post Shintai actually said something nice about an AMD chip I knew what was to come. I just have one question for you guys, who maybe know more about this: does Intel even have a division in Denmark?

http://www.intel.com/intel/location/Denmark.htm

[informal]

lol When in his first post Shintai actually said something nice about an AMD chip I knew what was to come. I just have one question for you guys, who maybe know more about this: does Intel even have a division in Denmark?

If they do(rhetorical q.),the million dollar question is who runs it^^

[Calmatory]

True that, And I will be happy to be wrong.

Frankly I honestly doubt that.

[gallag]

What I don't get is why the people that rag on shintia don't realise that they are just as biased and as negative to Intel. Pot, kettle, black.

[cegras]

What I don't get is why the people that rag on shintia don't realise that they are just as biased and as negative to Intel. Pot, kettle, black.

It's painful to read Shintai's posts. From page 3 to 4 he magically switched arguments. I assume it's because he finally got proved wrong and subconsciously moved on to something else to discuss.

[gallag]

Frankly I honestly doubt that.

Why?