First Nehalem Benchmarks :)

[Bojamijams]

Well the best part of all this, is that at the mid-end of November, we can build some very nice LGA775 systems for games, for very cheap

[jaredpace]

not that impressive... Probably next to zero improvement in games over yorkie

They would never show a gaming review like this pitting these chips (stock or Oc'd) against a 4.5ghz Penryn.

[SNiiPE_DoGG]

Well the best part of all this, is that at the mid-end of November, we can build some very nice LGA775 systems for games, for very cheap

yep, the quick movers will be very well rewarded, those who drag their feet will be stuck with low cache quads and crappy clockers on lga 775, watch those q9650 prices people

[Chad Boga]

but its not about the future its about all those that have a 65-45nm quad setup with most of them ddr2.

Why is it all about those with existing quads?

Nehalem will be a viable option for the majority who don't currently have a quad.

Intel just killed there own high end market share, not even AMD was required for that.

Nehalem enhances Intel's own high end and gives AMD even less places to hide with their lesser performance.

[Drwho?]

i am sure that you already tested that part, other wise you wouldn't post this . And indeed future wise it might be interesting, but its not about the future its about all those that have a 65-45nm quad setup with most of them ddr2. At this point in time switching would be waste of money for all of those.

Intel just killed there own high end market share, not even AMD was required for that. they should have focused on there q109 2s server en 2h09 4s server launch in stead of this desktop platform.

Little friend, with all the respect due, I did test Far Cry 2 and it does get faster.
Yorfield is in average 30% faster than any Phenoms on games, being 5% faster than Yorkfield is an archivement by itself, and if you think to 5% are easy to get, why don t you try to get it yourself? I mean, it is always easy to only concidere 5% easy, but if 5% was easy, AMD would do 5% + 5% +5% + 5% + 5% +5% and get back on track!

try to understand what you are taking about before doing big claims like this.
5% is super hard to get, and trust me, I mean really hard, especially on games, since some time is spent in the GPU too...

Just remember that Nehalem i7 is faster than the best super computer in the world in 1995 ... duh!

[Movieman]

i am sure that you already tested that part, other wise you wouldn't post this . And indeed future wise it might be interesting, but its not about the future its about all those that have a 65-45nm quad setup with most of them ddr2. At this point in time switching would be waste of money for all of those.

Intel just killed there own high end market share, not even AMD was required for that. they should have focused on there q109 2s server en 2h09 4s server launch in stead of this desktop platform.

Hello;
Can we first agree that I'm not your average bencher but someone that is only interested in computational power from a machine?
We'll also agree that I'm not a gamer so I can't speak to that area.
I've compared the Nehalem to both Q6600 based machines and to 8 core Harpertown(Yorkfield) based machines that I have here.
My Q6600 runs at 3600mhz at under 50C so it's what you'd consider a "good" one.
The Harpertown runs at 3758mhz with top X5470 chips so it also falls into the category of a 'good" one.
Comparing the Nehalem(Bloomfield) to those two systems made a beleiver out of me.
I've run benches all this last week in some DC apps and also in SuperPi and there is enough difference that it will "Shock and Amaze ya" to put it bluntly.
Had I seen as some have said a simple 5% increase I would have smiled and said that this was nice but thats not the case.
Comparing Nehalem to a Q6600 is like comparing a C2D to a high end P4 Netburst IMO..
Just blows it away in every test I've seen.
When the NDA lifts I'll show you the numbers and make a beleiver out of you!

[Drwho?]

let s add a little bit of knowlegde into the "only 5%" ...
based on Amdahl's law, The speedup of a program using multiple processors in parallel computing is limited by the time needed for the sequential fraction of the program. For example, if a program needs 20 hours using a single processor core, and a particular portion of 1 hour can not be parallelized, while the remaining promising portion of 19 hours (95%) can be parallelized, then regardless of how many processors we devote to a parallelized execution of this program, the minimal execution time can not be less than that critical 1 hour. Hence the speed up is limited up to 20x, as shown in the diagram on the right.
so, if a Game use 50% of the time the GPU, a speed up of 5% will be a processor improvement of 10%.
when you saw the gaming improvement of Core 2, you did not realized how big was the real improvement, but it was incredibly close to 100% on games like FEAR.
The improvement that Core 2 had are in Core i7.

Let me know if you need more help to understand the Amdahl's law, but every benchmark result should be analysed with this law in mind.

take a look here: Speed up

Francois

[Marvin_The_Martian]

Hello;
Can we first agree that I'm not your average bencher but someone that is only interested in computational power from a machine?
We'll also agree that I'm not a gamer so I can't speak to that area.
I've compared the Nehalem to both Q6600 based machines and to 8 core Harpertown(Yorkfield) based machines that I have here.
My Q6600 runs at 3600mhz at under 50C so it's what you'd consider a "good" one.
The Harpertown runs at 3758mhz with top X5470 chips so it also falls into the category of a 'good" one.
Comparing the Nehalem(Bloomfield) to those two systems made a beleiver out of me.
I've run benches all this last week in some DC apps and also in SuperPi and there is enough difference that it will "Shock and Amaze ya" to put it bluntly.
Had I seen as some have said a simple 5% increase I would have smiled and said that this was nice but thats not the case.
Comparing Nehalem to a Q6600 is like comparing a C2D to a high end P4 Netburst IMO..
Just blows it away in every test I've seen.
When the NDA lifts I'll show you the numbers and make a beleiver out of you!

Quit teasing us

[R101]

I think the teasing will stop on Monday .

[Marvin_The_Martian]

Oww idk, it can be on 8th as well as far as I'm concerned, would definetly spice up my birthday

[villa1n]

max stable BCLK On my boards:

P6T Deluxe - max stable QPI with i920 and i940 = 150MHz
Rampage II Extreme = 190MHz
Gigabyte X58-UD5 = 180MHz
MSI Eclipse = mobo dead after M-Flash BIOS upgrade ...

So i920's are looking at 3-3.4 max clock with good cooling on a good mobo it sounds like... I guess depending on the numbers that could be worth the upgrade

[Neuuubeh]

Mmm those rendering times look sexy. Not good enough to get me all worked up tho.

I'm kinda excited about the pricing on say q9450's a few weeks after nehalem release :>

[Caveman787]

Umm didn't they stop selling q9450's already at most places?

Since they were phased out by q9400 and q6600 is being phased out by q8200 and q9300.

There getting rid of all there good gaming processor's that were nice price/ performance and replacing them with ones that have a lot less cache and overclockability.

I find this very sad from a gamer/ poor man's perspective.

[Metroid]

So i920's are looking at 3-3.4 max clock with good cooling on a good mobo it sounds like... I guess depending on the numbers that could be worth the upgrade

I believe 4.0 is a reality if you get a good motherboard, 20 x 200 = 4000. The MSI claimed that anyone when using the Eclipse X58 can achieve that and no overclocking knowledge is needed, marketing was now some true reviews are needed.

Umm didn't they stop selling q9450's already at most places?

Since they were phased out by q9400 and q6600 is being phased out by q8200 and q9300.

There getting rid of all there good gaming processor's that were nice price/ performance and replacing them with ones that have a lot less cache and overclockability.

I find this very sad from a gamer/ poor man's perspective.

They will be replacing the higher MS segment with I7.

Metroid.

[justthefax]

Little friend, with all the respect due, I did test Far Cry 2 and it does get faster.
Yorfield is in average 30% faster than any Phenoms on games, being 5% faster than Yorkfield is an archivement by itself, and if you think to 5% are easy to get, why don t you try to get it yourself? I mean, it is always easy to only concidere 5% easy, but if 5% was easy, AMD would do 5% + 5% +5% + 5% + 5% +5% and get back on track!

try to understand what you are taking about before doing big claims like this.
5% is super hard to get, and trust me, I mean really hard, especially on games, since some time is spent in the GPU too...

Just remember that Nehalem i7 is faster than the best super computer in the world in 1995 ... duh!

Yorkfield 30% faster on avg? What are you, an Intel employee?

[jaredpace]

I believe 4.0 is a reality if you get a good motherboard, 20 x 200 = 4000. The MSI claimed that anyone when using the Eclipse X58 can achieve that and no overclocking knowledge is needed

[[XC] hipno650]

these are some pretty nice numbers. i plan on picking up an I7 920 and getting the best OCing board out there and hopefully i can hit at least 3.6ghz on air once new BIOS's come out. this is going to be one monster WCG CPU

[Metroid]

Toppc plans were uncovered too soon. That is what he wanted to show us when he got this motherboard but could not, in a away he did.

Metroid.

[Bellisimo]

Yorkfield 30% faster on avg? What are you, an Intel employee?

he is...

[AuDioFreaK39]

Wow, a 1074 point jump from QX9650 to Core i7 965.

[SAE]

^^Yeeeehaa! Lookin' quite good to me.
Wonder why multiplier and bus clocks were hidden from us though...

[Tonucci]

So i920's are looking at 3-3.4 max clock with good cooling on a good mobo it sounds like... I guess depending on the numbers that could be worth the upgrade

Someone may correct me but from what I've seen, the multi will go up a few notches even on non-extreme CPU's, due to the turbo mode. Extreme ones will go up beyond what turbo mode can do of course. Therefore we wont be limited by the QPI frequency alone.

[GAR]

since i will be buying an i920, i never in my life have got the xtreme cpu, that is for people who dont overclock, i will wit for 32nm, im sure clocks will be much better.

[AuDioFreaK39]

since i will be buying an i920, i never in my life have got the xtreme cpu, that is for people who dont overclock, i will wit for 32nm, im sure clocks will be much better.

I buy Extreme Edition Intel CPUs because of the unlocked multi.

In my opinion, once I bought my first Extreme Edition, I knew there was no going back. There is SO much unthinkable potential with an unlocked multiplier that I hadn't realized...all I could do was watch the great liquid nitrogen stories on sites like these. But now that I've had one, there's just no going back.

[Hornet331]

I buy Extreme Edition Intel CPUs because of the unlocked multi.

In my opinion, once I bought my first Extreme Edition, I knew there was no going back. There is SO much unthinkable potential with an unlocked multiplier that I hadn't realized...all I could do was watch the great liquid nitrogen stories on sites like these. But now that I've had one, there's just no going back.

hehe, same here. Once you tasted it you are not satisfied with other stuff.