core i7 or deneb

[DrMrLordX]

Originally posted by: keysplayr2003
Interesting.. Well, I'm a potential Deneb candidate. I'd like to pop something better into this 780a mobo. This Phenom 9550 X4 just doesn't cut it. And I can't even get this thing to 2.4GHz without it crashing. I get better scores from my 2.13GHz e6420 than the Phenom with both systems equipped with 9800GTX's and same drivers. Phenom isn't "bad", but it just needs to be soooo much better in light of it's competition. I hope Deneb delivers performance at least close or equal to the Core 2's (Conroes) at the same clock speed. At least that would be something.

Deneb is something I'd be willing to examine if they can show significantly reduced power consumption versus Agena. That has to be their #1 starting point. After reading several i7 reviews, one of the most embarrassing (and damning) displays of weakness on Agena's part has to be power consumption, particularly at idle. X4s consume far too much power.

 

[IntelUser2000]

Don't you guys read the leaked benchmarks??

Deneb is 6-7% faster than 65nm Phenom. The bigger L3 cache will mostly benefit server benchmarks, which will be needed when Nehalem will vastly outperform Penryn server chips and AMD will be playing catchup in all fronts. That's where the 20% comes from.

Kentsfield is 10% faster than Phenom

Yorkfield is faster than Kentsfield as Deneb is to Phenom.

The only thing that the tripled L3 cache did a lot is on SuperPi, the one that Core 2 has an insurmountable lead.

Yorkfield will still have 10% lead over 45nm Phenom.

 

[aigomorla]

Originally posted by: Ocguy31
You just made Deneb not only = Yorkfield, but = i7. I think you are taking a couple leaps there. I would wait for more information on Deneb. We have seen leaked benchies on Deneb, and noone was blown away.

ROFLROFL!!

agree'd

you just made inferior AMD TOCK cycle, = to intels new top the of the line TIC cycle.

Does this make any sense anyone else?

No, you amd lovers, DENEB as much as you want to praise has no hope in performance crown. Sorry guys, i love AMD too, but not when it comes to bleeding edge.

Can you guys just accept that?

Deneb will be nice, phenoms are nice, but there not best, and no where near it either.

id like to also say intel is coming out with a 65W yorkfield class quadcore to gobble deneb. LGA775. [yes i'll be getting lucky to sample one].

What does AMD have to say about that now? I think there in deep trouble because the 65W phenoms was what i thought was hot.

 

[Zstream]

Originally posted by: aigomorla

Originally posted by: Ocguy31
You just made Deneb not only = Yorkfield, but = i7. I think you are taking a couple leaps there. I would wait for more information on Deneb. We have seen leaked benchies on Deneb, and noone was blown away.

ROFLROFL!!

agree'd

you just made inferior AMD TOCK cycle, = to intels new top the of the line TIC cycle.

Does this make any sense anyone else?

No, you amd lovers, DENEB as much as you want to praise has no hope in performance crown. Sorry guys, i love AMD too, but not when it comes to bleeding edge.

Can you guys just accept that?

Deneb will be nice, phenoms are nice, but there not best, and no where near it either.

id like to also say intel is coming out with a 65W yorkfield class quadcore to gobble deneb. LGA775. [yes i'll be getting lucky to sample one].

What does AMD have to say about that now? I think there in deep trouble because the 65W phenoms was what i thought was hot.

First you say you are under a NDA but all other review sites published information. You stated you had the CPU for a while which I still doubt is the truth. You can not even run a test because your said "mobo" was bad from the start.

Take the above guys opinion with a grain of salt. That is all yours is worth.

Anyways, the larger, faster L3 Cache will allow AMD to still beat up Intel in the 8-16 way server benchmarks. Most servers do not even have HT enabled as it causes enormous problems with software containing multiple active threads at once. Such as PeopleSoft, SQL databases, lighthouse and email archive software such as Zantaz.

Deneb will compete with C2Q on the lower retail benchmarks, such as games and single threaded apps.

 

[aigomorla]

Originally posted by: Zstream

First you say you are under a NDA but all other review sites published information. You stated you had the CPU for a while which I still doubt is the truth. You can not even run a test because your said "mobo" was bad from the start.

mmmm okey....

is that suposed to make me feel bad and post?

You need to try a bit harder then that.

I lived though our OT section.

 

[IntelUser2000]

Anyways, the larger, faster L3 Cache will allow AMD to still beat up Intel in the 8-16 way server benchmarks. Most servers do not even have HT enabled as it causes enormous problems with software containing multiple active threads at once. Such as PeopleSoft, SQL databases, lighthouse and email archive software such as Zantaz.

What are you talking about?? Deneb doesn't have faster and larger L3 cache than Nehalem. Deneb has 6MB L3 and Nehalem has 8MB L3.

You don't even know how HT will affect Nehalem. The theory of HT on Pentium 4 and Core i7 is same sure, but the architecture is not. There are flaws in P4 that causes specific losses with HT. Nehalem doesn't. You might want to know Power 6 has HT that allows big gains in server apps.

 

[Idontcare]

Originally posted by: Zstream
First you say you are under a NDA but all other review sites published information.

Not speaking for Aigo, but I will kindly remind that not 100% of all Nehalem NDA's are related to Bloomfield (920/940/965).

I know nothing, but consider for example that perhaps Aigo is getting a preview rig for gainestown or beckton...if he had such and was under NDA for such then perhaps he might be able to talk much about such...and such.

 

[Zstream]

Originally posted by: IntelUser2000

Anyways, the larger, faster L3 Cache will allow AMD to still beat up Intel in the 8-16 way server benchmarks. Most servers do not even have HT enabled as it causes enormous problems with software containing multiple active threads at once. Such as PeopleSoft, SQL databases, lighthouse and email archive software such as Zantaz.

What are you talking about?? Deneb doesn't have faster and larger L3 cache than Nehalem. Deneb has 6MB L3 and Nehalem has 8MB L3.

You don't even know how HT will affect Nehalem. The theory of HT on Pentium 4 and Core i7 is same sure, but the architecture is not. There are flaws in P4 that causes specific losses with HT. Nehalem doesn't. You might want to know Power 6 has HT that allows big gains in server apps.

Please read my post again, I do not mean any rudeness but I never said it had more cache then Nehalem at least that is not what I meant and you know it. AMD's increased cache and speed will increase the performance for AMD chips. AMD already does well in the server market with the L3 cache it improves upon what was one of the limiting factors.

Based on the reviews and architecture handouts that have been released the HT is very similar to what it was on the P4. Power 6 chips are meant to run in parallelism and based off a complete different architecture.

 

[Zstream]

Originally posted by: Idontcare

Originally posted by: Zstream
First you say you are under a NDA but all other review sites published information.

Not speaking for Aigo, but I will kindly remind that not 100% of all Nehalem NDA's are related to Bloomfield (920/940/965).

I know nothing, but consider for example that perhaps Aigo is getting a preview rig for gainestown or beckton...if he had such and was under NDA for such then perhaps he might be able to talk much about such...and such.

Fair enough but in all of his post he mentions none of that.

 

[aigomorla]

Originally posted by: Zstream

Originally posted by: Idontcare

Originally posted by: Zstream
First you say you are under a NDA but all other review sites published information.

Not speaking for Aigo, but I will kindly remind that not 100% of all Nehalem NDA's are related to Bloomfield (920/940/965).

I know nothing, but consider for example that perhaps Aigo is getting a preview rig for gainestown or beckton...if he had such and was under NDA for such then perhaps he might be able to talk much about such...and such.

Fair enough but in all of his post he mentions none of that.

2 same cpu.



with 2 x QPI.

ive showed this picture 100000 times since sunday..

http://i125.photobucket.com/al...aigomorla/IMG_0092.jpg

 

[garritynet]

I have a 140watt AM2+ board. All I want to know is how much will the AM2+ Deneb cost and when will it come out?

 

[Idontcare]

Originally posted by: aigomorla
ive showed this picture 100000 times since sunday..

http://i125.photobucket.com/al...aigomorla/IMG_0092.jpg

That lap job is nasty dude

Why not go a few more minutes on the 400 grit and go to all copper across the IHS?

Question that may not be publicly answerable - what happens if you lap those loaner CPU's? You get charged retail for them when you go to return them or you just get some pissy emails followed by years of never getting a preview sample again? (probably the same answer to the question of what happens if you blow them up OC'ing on LN2 with 1.7 Vcc)

 

[IntelUser2000]

Based on the reviews and architecture handouts that have been released the HT is very similar to what it was on the P4. Power 6 chips are meant to run in parallelism and based off a complete different architecture.

The Nehalem and P4 architecture is VERY different. Sure, the implementation of multi-threading might look similar. I'll give you credit there.

The Pentium 4 had replay feature. When there are extra threads and it had to replay, it can take up precious bandwidth and hamper performance. Although there should be situations where replay should have an advantage, most of the times it didn't, and definitely not with multi-threading.

-Nehalem has no replay
-Nehalem has significantly higher bandwidth, even more than 4 cores alone require. It'll have extra for multi-threading.
-Unlike P4 SMT vs. P4 no-SMT, Nehalem has more out of order resources to accomodate for SMT

http://www.techreport.com/articles.x/15818/5
"The results without Hyper-Threading are curious: higher performance in the L1/L2 cache ranges, but lower performance in the L3 range."

In servers, L3 would be more important.

With substantially more bandwidth and more resources with wider core without caveats(like replay), the behavior on servers are gonna be completely different on Nehalem.