AMD delays Phenom 9700 into Q2 2008

Status
Not open for further replies.
since you are a Phenom-master do you know what I said about Phenom's TLB Fix is "to run the fix and overclock" it back is accurate (I heard about it on XS not sure if its true)
 
wow they are really killing the phenom, by that time you will wait a few more months an nahalem will be out, which is even faster, amd needs to pick up there socks, and i mean quick. they should have just released teh chips and doen a 9750 version and 9950 version like intel did with there e6000 series, and there is no way you will 100% your cpu, that will cause the crash, i have yet to go over 50% on my c2q, alltho i do not fold.
 
still has the bug
B3 stepping will definitely not have the bug

so what is the point of OCing it if it has a chance of crashing ur system at full load?
I don't think you've seen the thread where there have been people systematically hunting for the TLB bug
the only time the TLB ever showed up, is with:
* VMware
* 4 programs running, maxing the CPU load
* Cool'n'Quiet enabled

the TLB bug NEVER showed up with cool'n' quiet disabled, or without VMware running. no matter what load the CPU was under.

and overclocking does not in any way increase the chances of the TLB bug occurring.

The only reason AMD didn't release the higher-end chips, is because the OEM market wants those chips without the bug.
Plus, i hate it how the BIOS fix decreases performance.
So use a BIOS that doesn't have the fix.
If you really really want the TLB bug fixed, disable Cool'n'quiet. Or don't use VMware or any other virtualisation.

I've got BIOS 1.13 beta on my K9A2 Platinum, and I'll be sticking with it unless any of the newer versions can be modified to remove the TLB fix.
i can hold out until the 9700 n 9900 come out.
With an unlocked multiplier, who needs a higher model number?
you can just increase the multiplier, and you automatically have one that's as fast or faster without even touching the HTT speed.

wow they are really killing the phenom
I don't think they are. Phenom performs very well clock for clock, and scales better than C2Q when adding cores, and when increasing frequency.
The only think that's holding them back right now, is the frequencies they're running at.

by that time you will wait a few more months an nahalem will be out
Nehalem is due in mid 2009. B3 stepping is due in about a month.
AMD has Bulldozer planned for 2009.

they should have just released teh chips and doen a 9750 version and 9950 version like intel did with there e6000 series
Don't need to if you have unlocked multipliers.

and there is no way you will 100% your cpu, that will cause the crash, i have yet to go over 50% on my c2q, alltho i do not fold.
I max out my CPU a lot of the time. But there's an easy way to make sure the TLB bug won't happen, without a BIOS fix.
* disable Cool'n'Quiet
* don't use any virtualisation programs.

just one will do.
 
Nehalem is due in mid 2009. .

Nehalem is due in Q4 2008

Nehalem desktop for Q4 2008? - BeHardware

DailyTech - Intel Slates "Nehalem" for Q4 2008

Phenom performs very well clock for clock,

But around 10% less than the Kentsfield

and scales better than C2Q when adding cores

But tell me how you will add cores ?

They will not make more than 4 cores for desktop, and it seems than Core 2 performs better with 4 cores

and when increasing frequency.

I like to see a proof for that !

EDIT:
You need to check this
Phenom vs. Athlon Core Scaling Compared | Tom's Hardware

It seems that Phenom scales worse than its predecessor (Athlon64) in most cases
 
O RLY?
Intel Nehalem mainstream CPUs to be released in Q2 2009
that's more recent.

But around 10% less than the Kentsfield
when paired with DDR2-800 with pretty ordinary timings.
They perform very well when you put DDR2-1066 in with them, and with good timings. Then, you will see them pulling ahead in a lot of games.

Phenom at 2.6GHZ with DDR2-1066 gets 4300 CPU score in 3Dmark06, for example. Faster than a QX6800.

But tell me how you will add cores ?
multi-chip socket F/F+ systems

They will not make more than 4 cores for desktop, and it seems than Core 2 performs better with 4 cores
Doesn't scale as well when you increase frequencies.

I like to see a proof for that !
Search Xtremesystems
Phenom offers more than 100% clockscaling?! - XtremeSystems Forums

You need to check this
Phenom vs. Athlon Core Scaling Compared | Tom's Hardware

It seems that Phenom scales worse than its predecessor (Athlon64) in most cases
Great, let's test it in GAMES. Gee, that's a true test of frequency and core scaling!
and let's not give it DDR2-1066 RAM!
 

Read your own article

"Following the launch of the first Nehalem-based CPUs, the quad-core Bloomfield for high-end markets, in the fourth quarter this year, Intel is planning to launch a performance level Nehalem-based quad-core series, Lynnfield, before the second quarter of 2009, while dual-core Havendale CPUs will be launched in the second quarter of 2009 targeting markets from entry-level to mainstream, according to sources at motherboard makers"

High-end will come out on Q4 2008

The performance level will come before Q2 2009

The mainstream will come out on Q2 2009

You said mid 2009, and your own article proves that you are wrong


when paired with DDR2-800 with pretty ordinary timings.

No

Please check this
Phenom 9700, AMD's 1st Quad-Core CPU | Tom's Hardware
They were you using DDR2-1066, and it performed 10% worse than C2Q (Kentsfield) at the same clock speed

Phenom 9700, AMD's 1st Quad-Core CPU | Tom's Hardware

They perform very well when you put DDR2-1066 in with them, and with good timings. Then, you will see them pulling ahead in a lot of games.

Why you don't show us proof for your claims ?

Phenom at 2.6GHZ with DDR2-1066 gets 4300 CPU score in 3Dmark06, for example. Faster than a QX6800.

Links ? proof ?

Having higher 3dmark score doesn't always mean that processor is faster is real world applications, anyway.


multi-chip socket F/F+ systems

Is there a phenom processor that use socket F ?

As far as I know only AMD opteron use them and those are server processors



Are you kidding me ?

You didn't prove anything

It only scaled well in one application (cinebench) !!

Scaling well in one application doesn't mean that it will scale well in others

Great, let's test it in GAMES. Gee, that's a true test of frequency and core scaling!

It scaled worse than Athlon64 in gaming

Here is an example
prey.png



Athlon64 2.2GHz to 2.8GHz, there was %16.5 performance change.

Phenom 2.2GHz to 2.8GHz, there was only %12.7 performance change

And please before you reply to me, see this
Phenom vs. Athlon Core Scaling Compared | Tom's Hardware

They were using AMD 790FX motherboard, and they were using DDR2-800 with a good timing (4-4-4-12), yet Phenom scaled worse than Athlon64 with higher freq


and let's not give it DDR2-1066 RAM!

Do you know that DDR2-1066 cost almost twice as much as DDR2-800 ?

Most people here buy DDR2-800
 
Chill, people. There is no reason to act like this.

AMD pushed the chips back. that is nothing new in the industry. The reason wasn't a problem with the chips, which is the main point.

Enough with the sniping. If it continues, this (and all other threads of the same ilk) will be close without warning.
 
Read your own article

"The high-end quad-core Bloomfield Nehalem-based processors are slated for the fourth quarter of this year but for the performance level quad-core Lynnfield processors we'll have to wait until the second quarter of 2009.:

Nehalem will first come out on Q4 2008

However, the mainstream phenom will come on Q2 2009

You said mid 2009, which is neither Q4 2008 nor Q2 2009.
That's nice. But I really don't care.
Phenom performs better than you think it does, and I'm going to buy one. That's what matters to me.
No

Please check this
Phenom 9700, AMD's 1st Quad-Core CPU | Tom's Hardware
They were you using DDR2-1066, and it performed 10% worse than C2Q (Kentsfield) at the same clock speed

Phenom 9700, AMD's 1st Quad-Core CPU | Tom's Hardware
That review, they didn't use an AM2+ board. regular AM2 does not support DDR2-1066. Only DDR2-800.

Why you don't show us proof for your claims ?
I don't need to prove anything to you. Even if I do give you links, you're going to keep attacking the Phenom anyway.

I could give you examples. but really? I can't be bothered arguing with an intel fanboy.

Links ? proof ?
See above.

Having higher 3dmark score doesn't always mean that processor is faster is real world applications, anyway.
Oh, like Superpi? those kind of real-world applications?

It only scaled well in one application (cinebench) !!
I only gave you one link.
Besides, I really can't be bothered arguing with you.

From Athlon64 2.2GHz to 2.8GHz, there was %16.5 performance change.

From phenom 2.2GHz to 2.8GHz, there was only %12.7 performance change

And please before you reply to me, see this
Phenom vs. Athlon Core Scaling Compared | Tom's Hardware

They were using AMD 790FX motherboard, and they were using DDR2-800 with a good timing (4-4-4-12), yet Phenom scaled worse than Athlon64 with higher freq
Do you know that DDR2-1066 cost almost twice as much as DDR2-800 ?
Depends where you look.
RAM is cheap anyway.
 
That's nice. But I really don't care.

Then don't post incorrect information

Nehalem will come much earlier than Mid 2009

I don't need to prove anything to you. Even if I do give you links, you're going to keep attacking the Phenom anyway.

I could give you examples. but really? I can't be bothered arguing with an intel fanboy.

See above.

Thats what people always say when they can't prove their points

You are saying that I'm intel fanboy as if you are not an AMD fanboy !!

you're going to keep attacking the Phenom anyway.

And you were attacking Intel by posting incorrect information like "Nehalem will come out on Mid 2009" and "scales better than C2Q"


Oh, like Superpi? those kind of real-world applications?

Did I say that Superpi is a kind of real-world application ? The answer is no


I only gave you one link

So, you can post more links.

You only proved that phenom scales well only in cinebench !!

While Tom's Hardware benchmarks proves that phenom scales worse than Athlon64 with higher freq in most applications

Besides, I really can't be bothered arguing with you.

No one is forcing you to argue with me. If you don't want to argue then don't.
 
Status
Not open for further replies.
Back
Top Bottom