ES of Kentsfield out in the wild....

I don't recognize the bench used so I can't say anything about performance... but, if it works on a 975 board, does that mean the retail Kentsfield will likely run on the same boards and Conroe? If so that makes my Conroe upgrade choice easy if I can pop in a quadcore when they come out.
 
NulloModo said:
I don't recognize the bench used so I can't say anything about performance... but, if it works on a 975 board, does that mean the retail Kentsfield will likely run on the same boards and Conroe? If so that makes my Conroe upgrade choice easy if I can pop in a quadcore when they come out.


I believe I heard at one point that they are chipset-compatible (in addition to being the same socket). The "bench" isn't a bench, it is a different Prime95 client for easier stress-testing on multicore CPU's.
 
NulloModo said:
I don't recognize the bench used so I can't say anything about performance... but, if it works on a 975 board, does that mean the retail Kentsfield will likely run on the same boards and Conroe? If so that makes my Conroe upgrade choice easy if I can pop in a quadcore when they come out.
Yeh the timeframe is Q1 2007, so were looking at just before the introduction of the Bearlake and Weybridge ICH9 chipsets in Q2 2007.

It will most likely work with 975X, I hope Intel didn't do some voltage thing again.
 
coldpower27 said:
Yeh the timeframe is Q1 2007, so were looking at just before the introduction of the Bearlake and Weybridge ICH9 chipsets in Q2 2007.

It will most likely work with 975X, I hope Intel didn't do some voltage thing again.

I will most likely go for a 965 board that offers extra SATA, SATA-E, and Firewire ports anyway, but still good to know there will be an upgrade path. With the P4 there were two different sockets within the same generation, nice to know things have changed.
 
NulloModo said:
I don't recognize the bench used so I can't say anything about performance... but, if it works on a 975 board, does that mean the retail Kentsfield will likely run on the same boards and Conroe? If so that makes my Conroe upgrade choice easy if I can pop in a quadcore when they come out.

Ya but you will probably want DDR3 memory and Intel hasn't released a DDR2 memory controller yet that works with DDR3 memory
 
$BangforThe$ said:
Ya but you will probably want DDR3 memory and Intel hasn't released a DDR2 memory controller yet that works with DDR3 memory

Memory bandwidth isn't as important as CPU performance anymore really. Look at Conroe vs. X2s, the X2s have a far better memory bandwidth, yet still fall far behind.
 
NulloModo said:
Memory bandwidth isn't as important as CPU performance anymore really. Look at Conroe vs. X2s, the X2s have a far better memory bandwidth, yet still fall far behind.

Were talking about 4 cores here memory bandwidth will be important factor.
 
I'll start drooling when he posts some benchmarks. It's pretty impressive that his sample is a 2.4GHz version so early.
 
$BangforThe$ said:
Were talking about 4 cores here memory bandwidth will be important factor.

the bus bandwidth will probably be more of a limiting factor long before the memory bandwith becomes to much of an issue.
 
pxc said:
I'll start drooling when he posts some benchmarks. It's pretty impressive that his sample is a 2.4GHz version so early.

It could be a good sign that Intel is right on tracks with it's time table :)

Terra - Wonder how long they have been spitting retail Conroes out then ;)
 
Terra said:
It could be a good sign that Intel is right on tracks with it's time table :)

Terra - Wonder how long they have been spitting retail Conroes out then ;)

Hehe it's sweet isn't it? Intel has a dual core CPU that can kill AMDs best and ES samples of quad-core chips way before they even have working silicon.
 
NulloModo said:
Hehe it's sweet isn't it? Intel has a dual core CPU that can kill AMDs best and ES samples of quad-core chips way before they even have working silicon.

The blue gigant has awoken with one hell of a war-cry ;)

Terra - And it will only benefit end-users...that would be US!!! :D
 
Something is not as it should be with the forum database...
I have seen several posts now that appear in the wrong order..all across the forum?

Terra...
 
JetUsafMech said:
Obviously it's fake...

Just wanted to chime in before duby229...heh.
Don't forget vis. I'm sure he'll be here with FUD soon.
 
Terra said:
Something is not as it should be with the forum database...
I have seen several posts now that appear in the wrong order..all across the forum?

Terra...

I think the intarweb is broken. I keep getting 'refused by host' errors over the past several days.

The god of the internets must be in a foul mood.
 
NulloModo said:
I think the intarweb is broken. I keep getting 'refused by host' errors over the past several days.

The god of the internets must be in a foul mood.

So far the only place it have seen anything out of the ordinary is the post-switch-problem here on [H]ard.

Terra...
 
Terra said:
So far the only place it have seen anything out of the ordinary is the post-switch-problem here on [H]ard.

Terra...

post switches I have only seen here, I have seen the 'client refused connection' message a lot of places.
 
Coolaler posted benchmarks:

Cinebench CPU Render Test
1 thread: 63.1s
4 threads: 19.5s

Sandra Multi-Media
Integer x8 264482 it/s
FP x4 142883 it/s

Sandra cache/memory
41502 MB/s

TMPG encoding (m:s)
Kentsfield 2.4GHz 10:55 and 11:11
Conroe @ 2.95GHz 15:34 and 15:32
Merom @ 2.96GHz 16:39 and 16:38

Everest
CPU 7627, 52298 KB/(s? memory subsystem?)

and that's at stock speed with DDR2-667 memory. :eek:

He overclocked the Kentsfield to 3.2GHz on stock voltage: http://www.xtremesystems.org/forums/showpost.php?p=1528175&postcount=72
 
NulloModo said:
Why did Merom take longer than Conroe in TMPG ?
Lower FSB speed. The Conroe starts at 1066MHz FSB and the Merom starts at 667MHz FSB. At 2.95GHz the Conroe he used has a 368MHz FSB and at 2.96GHz the Merom he used has a 228MHz FSB. The details are posted on page 3.

Another benchmark:
3DMark06 CPU test 3519 @ stock 2.4GHz
 
I found this tidbit interesting:

The scaling is almost perfectly linear:

Single CPU: 63 seconds
Quad Core: 19 seconds

63/(19*4) = 0.82

So there is about an 18% decrease from theoretical max.


The 8 core AMD HT enabled system shows a 4.73x (best case) speed up with 8 cores.

4.73 / 8 = 0.59

Which is a 41% decrease from theoretical max.

Which one do you think is scaling better?

Terra - That new architecture sure is well made :D
 
So anyone besides me think that the 4 core Woodcrest edition running Dual FSB's in a DP motherboard is going to give AMD fits at the MP 4 socket level for AMD servers.

I see equal performance with way less power usage.
 
Terra said:
I found this tidbit interesting:



Terra - That new architecture sure is well made :D


Lets see a two-way Kensfield before we compare it to an 8 core Opteron. (or see a 4 core opteron's scaling)
 
robberbaron said:
Lets see a two-way Kensfield before we compare it to an 8 core Opteron. (or see a 4 core opteron's scaling)
It will be a long time to see a 4 core opty scaling . 08. to long to wait on that.
 
$BangforThe$ said:
It will be a long time to see a 4 core opty scaling . 08. to long to wait on that.

Prove it :rolleyes:

Show me the 4 core opteron bench, please.
 
Kentsfield will work with Conroe capable mobos.

Anyway , 2.4GHz is more than I expected at first , and it is probably using a 80w TDP.There is the performance version ( XE ) of Kentsfield running at a TDP of 120w ( I expect it to run at 2.66GHz )

:eek:
 
HOLY SHIT. That thing ran 4x 1mb PI in 25secs, my PD [email protected] can do 1x 1mb in like 26 secs. Wow, the Kentsfeild can do 4 times as much work as my PD 950 in less time than my CPU :eek: .
 
cupholder2.0 said:
HOLY SHIT. That thing ran 4x 1mb PI in 25secs, my PD [email protected] can do 1x 1mb in like 26 secs. Wow, the Kentsfeild can do 4 times as much work as my PD 950 in less time than my CPU :eek: .
It is a newer CPU generation. What did you expect? It should be slower?
robberbaron said:
Prove it :rolleyes:

Show me the 4 core opteron bench, please.
that is the nice thing talking about future chips and technology, you don't need to prove anything. I do believe Coolaler's benches, but anything else I am taking with a huge grain of salt.
 
JetUsafMech said:
Obviously it's fake...


Just wanted to chime in before duby229...heh.


What does that have anything to do with me? I never said anything was fake. I said I dontr trust quite a few benches out there, thank god for it too, I'm sure you guys dont believe every bench you see either.... If you do, then something wrong
 
pxc said:
Don't forget vis. I'm sure he'll be here with FUD soon.


What fud? Asking questions is fud? WOW! You may need to see a doctor about that. If you believe everything you see, then it could cause some serious mental defects later in your life.
 
duby229 said:
What fud? Asking questions is fud? WOW! You may need to see a doctor about that. If you believe everything you see, then it could cause some serious mental defects later in your life.
Like your post. I'll take advice from the expert on "mental defects." :D

Dismissing something because you don't want to believe it is a sure sign! Obviously you didn't even look at the thread. There are plenty of pictures to make it easy to read!
 
He still posting more stuff. When apps are written for multi- cores its going to be just crazy fast. The Amd -!!!!!!s were just getting use to seeing conroe's name blasted all over the web . Now they have Kentsfield screaming at them . SHOW ME WHAT YOU GOT!
 
This will be a nice upgrade path. Buy a Conroe ready mobo.
And you have your choice of a cheap 900 series to hold you over OR, any Conroe by the end of July, OR Kentsfield by the beginning of next year.

Too bad most of that extra power of 4 cores won't be used any time soon. Though I will get one as soon as they come out any way so I won't have e-penis envy. :)
 
Back
Top