Nehalem-EP......BLOOMFIELD

[Jupiler]

If I created offense, it was not intended. But I apologize nonetheless, because that's not what any of us are here for.

Don't worry, it wasn't your post I was refering to.
The rude comments that were made by some have already been removed.

[Hornet331]

Before jumped to the conclusion, re-run with no crap running in the background, just for fair comparison ...
Other else, just wait for the author comment

doesn't really matter its only a few mb/s.

http://www.computerbase.de/news/soft..._60_benchmark/

Q9450 @ 3,3 scores 416mb/s in ASE. my QX9650 @ 3ghz scores 419mb/s (i think its the faster ram on my system, thats why it beats a Q9450 @ 3,3ghz )

[gOJDO]

1) Don't forget your NDA.

What NDA you are talking about? Intel are letting people to publish Nehalem results. A month ago Anadtech published a lot of results:http://www.anandtech.com/cpuchipsets...spx?i=3326&p=2
I wonder if they were under NDA too.

2) This is an old silicon.

Cool, show us the new one.

Thanks for not informing our competitors, you are hurting us when you publish bench results, you all need to understand that.

Your competitor is well informed about your upcoming Nehalem CPUs a long time ago. Even if they had no sources to know the results, they could estimate the performance of Nehalem much better then most of us here(some XS members were very accurate in their predictions about Nehalem).

About the "hurting" part, the results can only hurt your competitor more since it has nothing to compete against Nehalem.

I know it is fun to publish your proto numbers, just remember that you are as well hurting the competitiveness of my employer.

In the past 3 years your employer used to publish lots of results of their upcoming products. If you haven't forgot, there were thousands of Conroe ES's floating around months before it was officially released, so your competitor was very well informed. Pity, they were unable to do anything, just like they are unable to do anything about Nehalem now.

I am speaking for myself only here, not for my employer here.

Nope, you were speaking about how the pre-release informations are going to hurt your employer.

It makes my job very hard , when my competitor can know what I am planning, before it is even finish.

Well, we are not your competitor, we are your customers. Your competitor have an informations from insiders, just like you have.

There are no personal archivement to release confidential information, except taking the risk to be in trouble.

Well, there are. Those who are publishing results and informations of unreleased products have high reputation here on XS. In that regard we must give a credit to JC.

Just think about it as "you work on something very cool, a good cake for your Daughter birth day, but a dude show the picture of the cake to your daughter before the D day." That kind of suck, isn't it?

Oh. What can you say about Intel publishing results of unreleased Yorkfield/Wolfdale CPUs, and it's unavailability months after it is being released. Only QX9650 were available, which are affordable for a very few rich customers. It is like a lot of kids watching other kid eating an ice-cream. That kind of sucks more, isn't it?

[Movieman]

What NDA you are talking about? Intel are letting people to publish Nehalem results. A month ago Anadtech published a lot of results:http://www.anandtech.com/cpuchipsets...spx?i=3326&p=2
I wonder if they were under NDA too.

Cool, show us the new one.

Your competitor is well informed about your upcoming Nehalem CPUs a long time ago. Even if they had no sources to know the results, they could estimate the performance of Nehalem much better then most of us here(some XS members were very accurate in their predictions about Nehalem).

About the "hurting" part, the results can only hurt your competitor more since it has nothing to compete against Nehalem.

In the past 3 years your employer used to publish lots of results of their upcoming products. If you haven't forgot, there were thousands of Conroe ES's floating around months before it was officially released, so your competitor was very well informed. Pity, they were unable to do anything, just like they are unable to do anything about Nehalem now.

Nope, you were speaking about how the pre-release informations are going to hurt your employer.

Well, we are not your competitor, we are your customers. Your competitor have an informations from insiders, just like you have.

Well, there are. Those who are publishing results and informations of unreleased products have high reputation here on XS. In that regard we must give a credit to JC.

Oh. What can you say about Intel publishing results of unreleased Yorkfield/Wolfdale CPUs, and it's unavailability months after it is being released. Only QX9650 were available, which are affordable for a very few rich customers. It is like a lot of kids watching other kid eating an ice-cream. That kind of sucks more, isn't it?

Puts on mod hat hopefully for the last time in this thread.
Did you read my post? If not please read it again.
I said the man is owed our respect even if we disagree with certain points he made. I do NOT want to see this turn into a "Lets bait and bash the Intel guy"
I mean that and am VERY serious.
Let it go please.
Thanks for reading.

[Kingcarcas]

Ok guys, how long do you think until we can buy a $200 Nehalem? I'm thinking of just jumping to a Q6600 and leaving it at that for 1-2 years....

[karl_eller]

Q1 '09? That's when the mainstream stuff is supposed to be coming out, so probably around then. That'll be the LGA1066 platforms.

Eller

[Shintai]

Q1 '09? That's when the mainstream stuff is supposed to be coming out, so probably around then. That'll be the LGA1066 platforms.

Eller

LGA1160

[davewolfs]

A few questions.

1. When should we see the first round of Bloomsfield chips?
2. How much will they cost?
3. What type of DD3 memory is a safe bet to buy now?

[Donnie27]

Ok guys, how long do you think until we can buy a $200 Nehalem? I'm thinking of just jumping to a Q6600 and leaving it at that for 1-2 years....

Isn't $200 a little ridiculous? Intel's RM shows 325 to 400 us dollars. For what folks will be getting, hell, that is a Budget processor. That puts it in the range of the very hot selling Q6600 and the cheaper than the 35 and 3800's Short memories? One VAR told me more realistic price is about the same as the Q9300's first shipping price of $339 (from their shop)

For $200 one in this forum should expect more than Dual Channel memory controllers instead of tri-Channel IMC as well. Yes, I'd love to see a Tri-Ch $200 processor but I'd be surprised as hell to see.

Yes Movieman, I laughed at the guy talking about Gimp RAM.

[virtualrain]

A few questions.

1. When should we see the first round of Bloomsfield chips?
2. How much will they cost?
3. What type of DD3 memory is a safe bet to buy now?

1. The only official date we have is Q4. Having said that, the CPU's are rumored to be good to go. We are now only waiting on motherboard validation and tuning. Tom's has a Foxconn X58 board on hand and they say the boards are half-way through validation. Given that the first boards appeared at Computex, one might conclude that Nehalem systems could be ready in a couple of months. Conroe did make an early appearance so perhaps Nehalem will as well.

2. The rumored costs are $316 (2.66GHz) to $999 (3.2GHz).

3. Unless you intend to overclock the max stock DDR3 RAM speeds supported will be DDR3-1333 for the 3.2GHz part and DDR3-1066 for the 2.66 and 2.93 GHz parts. However, overclocking will likely allow any speed up to the max supported by your memory. Keep in mind that Bloomfields IMC and tri-channel capability means there will be a lot less need for super high speeds (after all, DDR3-1333 in tri-channel mode has a peak theoretical bandwidth of 32GB/s which is more than C2D L2 cache!). I would focus on a kit that has lower latencies at 1333 speeds than a kit that blows the doors off at 2000MHz. More info on RAM potential with Nehalem here.

Cheers,
-Chris.

[savantu]

...

3. Unless you intend to overclock the max stock DDR3 RAM speeds supported will be DDR3-1333 for the 3.2GHz part and DDR3-1066 for the 2.66 and 2.93 GHz parts. However, overclocking will likely allow any speed up to the max supported by your memory. Keep in mind that Bloomfields IMC and tri-channel capability means there will be a lot less need for super high speeds (after all, DDR3-1333 in tri-channel mode has a peak theoretical bandwidth of 32GB/s which is more than C2D L2 cache!). I would focus on a kit that has lower latencies at 1333 speeds than a kit that blows the doors off at 2000MHz. More info on RAM potential with Nehalem here.

Cheers,
-Chris.

Umh..the L2 of C2D does 96GBs at 3GHz.

Your point stands however.I don't think that OC memory will give a worthy performance benefit at over 30GBs.Lower latency might be a winner.

On a similar note , I'd expect the Nehalem IMC to be more sensitive to memory and especially voltage.Because it is incorporated into the CPU , it is made with 45nm.That causes problems when linked to higher voltages needed for RAM links ( 1.5V ).Being in the chipset alleviated this problem since chipsets are usually on n-1 or n-2 process tech which matches better the voltages used in RAM.

[Hornet331]

Umh..the L2 of C2D does 96GBs at 3GHz.

Your point stands however.I don't think that OC memory will give a worthy performance benefit at over 30GBs.Lower latency might be a winner.

On a similar note , I'd expect the Nehalem IMC to be more sensitive to memory and especially voltage.Because it is incorporated into the CPU , it is made with 45nm.That causes problems when linked to higher voltages needed for RAM links ( 1.5V ).Being in the chipset alleviated this problem since chipsets are usually on n-1 or n-2 process tech which matches better the voltages used in RAM.

yeah i have thought about that, i really would be disappointed if i could us, lets say 1.9V DDR3 2000 on nehalem if that voltage would fry the imc or other thigs that are related to it.

[virtualrain]

Umh..the L2 of C2D does 96GBs at 3GHz.

Your point stands however.I don't think that OC memory will give a worthy performance benefit at over 30GBs.Lower latency might be a winner.

On a similar note , I'd expect the Nehalem IMC to be more sensitive to memory and especially voltage.Because it is incorporated into the CPU , it is made with 45nm.That causes problems when linked to higher voltages needed for RAM links ( 1.5V ).Being in the chipset alleviated this problem since chipsets are usually on n-1 or n-2 process tech which matches better the voltages used in RAM.

I was perhaps inappropriately comparing Nehalem's theoretical memory bandwidth to C2D's L2 real-world measured bandwidth (which from what I've seen in Everest is under 30GB/s). So my appologies for an apples to oranges comparison.

Is it the RAM modules or the RAM bus that needs to operate at 1.5-1.9V?

At any rate, it's been suggested that the IMC will be on it's own voltage/clock domain which will hopefully allow tweakers to optimize it for stability/performance.

[Donnie27]

I was perhaps inappropriately comparing Nehalem's theoretical memory bandwidth to C2D's L2 real-world measured bandwidth (which from what I've seen in Everest is under 30GB/s). So my appologies for an apples to oranges comparison.

Is it the RAM modules or the RAM bus that needs to operate at 1.5-1.9V?

At any rate, it's been suggested that the IMC will be on it's own voltage/clock domain which will hopefully allow tweakers to optimize it for stability/performance.

That's what I'm thinking as well. This came up in the Intel will not allow overclocking thread.

[Kingcarcas]

1. The only official date we have is Q4. Having said that, the CPU's are rumored to be good to go. We are now only waiting on motherboard validation and tuning. Tom's has a Foxconn X58 board on hand and they say the boards are half-way through validation. Given that the first boards appeared at Computex, one might conclude that Nehalem systems could be ready in a couple of months. Conroe did make an early appearance so perhaps Nehalem will as well.

2. The rumored costs are $316 (2.66GHz) to $999 (3.2GHz).

3. Unless you intend to overclock the max stock DDR3 RAM speeds supported will be DDR3-1333 for the 3.2GHz part and DDR3-1066 for the 2.66 and 2.93 GHz parts. However, overclocking will likely allow any speed up to the max supported by your memory. Keep in mind that Bloomfields IMC and tri-channel capability means there will be a lot less need for super high speeds (after all, DDR3-1333 in tri-channel mode has a peak theoretical bandwidth of 32GB/s which is more than C2D L2 cache!). I would focus on a kit that has lower latencies at 1333 speeds than a kit that blows the doors off at 2000MHz. More info on RAM potential with Nehalem here.

Cheers,
-Chris.

Thank you very much! 1333 is the most realistic anyway seeing how expensive 1600+ are

[Proka]

What NDA you are talking about? Intel are letting people to publish Nehalem results. A month ago Anadtech published a lot of results:http://www.anandtech.com/cpuchipsets...spx?i=3326&p=2
I wonder if they were under NDA too.

Cool, show us the new one.

Your competitor is well informed about your upcoming Nehalem CPUs a long time ago. Even if they had no sources to know the results, they could estimate the performance of Nehalem much better then most of us here(some XS members were very accurate in their predictions about Nehalem).

About the "hurting" part, the results can only hurt your competitor more since it has nothing to compete against Nehalem.

In the past 3 years your employer used to publish lots of results of their upcoming products. If you haven't forgot, there were thousands of Conroe ES's floating around months before it was officially released, so your competitor was very well informed. Pity, they were unable to do anything, just like they are unable to do anything about Nehalem now.

Nope, you were speaking about how the pre-release informations are going to hurt your employer.

Well, we are not your competitor, we are your customers. Your competitor have an informations from insiders, just like you have.

Well, there are. Those who are publishing results and informations of unreleased products have high reputation here on XS. In that regard we must give a credit to JC.

Oh. What can you say about Intel publishing results of unreleased Yorkfield/Wolfdale CPUs, and it's unavailability months after it is being released. Only QX9650 were available, which are affordable for a very few rich customers. It is like a lot of kids watching other kid eating an ice-cream. That kind of sucks more, isn't it?

Puts on mod hat hopefully for the last time in this thread.
Did you read my post? If not please read it again.
I said the man is owed our respect even if we disagree with certain points he made. I do NOT want to see this turn into a "Lets bait and bash the Intel guy"
I mean that and am VERY serious.
Let it go please.
Thanks for reading.

I didn't see his comment as disrestpectful. Are we forbidden from even disagreeing with company employees now? There's a fine line between some disagreements and rudeness.

There's also a fine line between being fair and outright butt-licking.

We're getting a bit tired of the condescending "I'm putting on my mod hat..." schtick frankly. Sure go ahead and delete this post, ban me if you want, but more than a few people are noticing a disturbing trend by mods on Xtremesystems lately, and a lack of even-handedness.

[Anemone]

1333 and low latency is certainly nice. Does this mean the nice 1600 ram that everyone has is going to be less than useful?

[Aberration]

I didn't see his comment as disrestpectful. Are we forbidden from even disagreeing with company employees now? There's a fine line between some disagreements and rudeness.

There's also a fine line between being fair and outright butt-licking.

We're getting a bit tired of the condescending "I'm putting on my mod hat..." schtick frankly. Sure go ahead and delete this post, ban me if you want, but more than a few people are noticing a disturbing trend by mods on Xtremesystems lately, and a lack of even-handedness.

I couldnt disagree more.

[Movieman]

I didn't see his comment as disrestpectful. Are we forbidden from even disagreeing with company employees now? There's a fine line between some disagreements and rudeness.

There's also a fine line between being fair and outright butt-licking.

We're getting a bit tired of the condescending "I'm putting on my mod hat..." schtick frankly. Sure go ahead and delete this post, ban me if you want, but more than a few people are noticing a disturbing trend by mods on Xtremesystems lately, and a lack of even-handedness.

You probably also didn't see the two posts I deleted from this thread addressed to Francois that had more dancing banannas in them than Jone's in a NYC telephone book.
If asking people not to flame the project coordinator of the SkullTrail system for Intel who was good enough to come here and try to help then yes, I am guilty of that.
In my opinion good manners is always the best way to proceed.
By the way, I don't appreciate the ass-licking comment.
If you'd opened your eyes and actually read through all of this thread you'd have seen that I posted disagreeing with Francois but I did it in a polite manner.
Now if you think that Xs is just a place to come and kick sand in peoples faces perhaps you should think twice before you post.
As to the putting on the mod hat, something I've used to discern between my posts as a member and my posts as a mod.
Your insinuation of "more than a few people" means nothing to me as that is what people say when they are insecure about their own position and want to make it look more legitimate by saying others think the same.
When others post their feelings then we will know but certainly not on just your say so.
So, is there anything else you'd care to discuss while you still have an account here?
I forgot to add this:
Anytime you feel that things here are just so bad that you can't handle it feel free to leave.
No one is forcing you to stay here..

[Proka]

You probably also didn't see the two posts I deleted from this thread addressed to Francois that had more dancing banannas in them than Jone's in a NYC telephone book.
If asking people not to flame the project coordinator of the SkullTrail system for Intel who was good enough to come here and try to help then yes, I am guilty of that.
In my opinion good manners is always the best way to proceed.
By the way, I don't appreciate the ass-licking comment.
If you'd opened your eyes and actually read through all of this thread you'd have seen that I posted disagreeing with Francois but I did it in a polite manner.
Now if you think that Xs is just a place to come and kick sand in peoples faces perhaps you should think twice before you post.
As to the putting on the mod hat, something I've used to discern between my posts as a member and my posts as a mod.
Your insinuation of "more than a few people" means nothing to me as that is what people say when they are insecure about their own position and want to make it look more legitimate by saying others think the same.
When others post their feelings then we will know but certainly not on just your say so.
So, is there anything else you'd care to discuss while you still have an account here?
I forgot to add this:
Anytime you feel that things here are just so bad that you can't handle it feel free to leave.
No one is forcing you to stay here..

It's ok, you can let SHintai and DrWho service you. Friends with benefits indeed

[Movieman]

It's ok, you can let SHintai and DrWho service you. Friends with benefits indeed

Shintai I know and get along with but we've never met.
Dr.Who I have spoken with exactly twice and I doubt he even know's who I am..
As to your sarcastic "service you" comment.
Consider it your last. Good bye and don't let the door slam you in the ass on your way out.

[virtualrain]

1333 and low latency is certainly nice. Does this mean the nice 1600 ram that everyone has is going to be less than useful?

Of course not... just run your DDR3-1600 at lower clocks and tighter timings!

[Movieman]

Of course not... just run your DDR3-1600 at lower clocks and tighter timings!

Nice to see a post here. I was getting the feeling my banning that guy ran everyone off and that wasn't my intent..

[MomijiTMO]

Nice to see a post here. I was getting the feeling my banning that guy ran everyone off and that wasn't my intent..

All I can say is that this new member who joined in June wouldn't be tired of you saying "puts on a mod hat" unless they had really been here before and say. . . were banned. .

No new posts since there isn't any new benches.

PS : I wonder if the new rumors of SLI on X58 will be true.

PPS : I've been here for a short time and have only seen you say the said line twice.

[Movieman]

All I can say is that this new member who joined in June wouldn't be tired of you saying "puts on a mod hat" unless they had really been here before and say. . . were banned. .

No new posts since there isn't any new benches.

PS : I wonder if the new rumors of SLI on X58 will be true.

PPS : I've been here for a short time and have only seen you say the said line twice.

I will say publicly how I view the threads here.
If I see an issue I try and calm it down.
Gently but firmly and usually, but not always without naming individuals.
If that doesn't work and someone challenges me then yes, I will say it straight to the individual.
If that doesn't work and the guy comes back at me as this one did then yes, I will take action.
This is the first member I believe that I have permabanned and only the third ban to any member(barring the spammers) that I have given out.
It is not something I enjoy at all but without some sort of order you have caos.
I try to view all the members as friends that I have not met yet and treat them in that way but I am human, there are buttons that can be pushed that will bring out my temper but rarely and someone has to be really looking for trouble and push hard before I react.
That ass-licking comment and the last one about Shintai and Dr. Who "servicing me" said that this guy wasn't going to back down and wanted a ban,The martyr syndrome".. He asked for it and he got what he wanted.