Kensfield ES is being tested

[Cooler]

http://www.xtremesystems.org/forums/showthread.php?t=103982

8Mb Cache at 2.4 GHz with a 1066 FSB. It seems that this quad core monster may be coming sooner then expected. Also Now With OC of 3.2 Ghz at end of thread .

 

[d3lt4]

I was just about to post this.... Oh well, This looks awesome. The CPU market is really looking interesting. I wonder how much that will cost?

 

[CrystalBay]

Man , those are some monster numbers for 2.4 Ghz.

Thanks

 

[imported_Questar]

80% scaling from two to four cores. So much for the FSB holding things back.

 

[TraumaRN]

Not to sound like a fanboy in either direction, but unless AMD has a counter to this and quickly they'll be beaten like a red-headed stepchild....

 

[Munky]

Geez, why dont they just slap 2 gigs of cache on the thing and call it a day? :roll: Seems like their favorite solution to whatever and limitations the cpu and its platform might have - more cache.

 

[dmens]

Geez, why dont they just slap 2 gigs of cache on the thing and call it a day? Seems like their favorite solution to whatever and limitations the cpu and its platform might have - more cache.

Damn, if it were that easy, I might as well just go home at 2pm every day. All that crazy crap I have to do to eek out 1% extra int perf is just a waste of time anyways.

 

[HostofFun]

Originally posted by: munky
Geez, why dont they just slap 2 gigs of cache on the thing and call it a day? :roll: Seems like their favorite solution to whatever and limitations the cpu and its platform might have - more cache.

The issue with that being? If it gets the job done for them, what's it matter to you?

P.S. We can fire the engineers now!

 

[ZOXXO]

Originally posted by: munky
Geez, why dont they just slap 2 gigs of cache on the thing and call it a day? :roll: Seems like their favorite solution to whatever and limitations the cpu and its platform might have - more cache.

What's your solution?

 

[Tangerines]

Check out the CPU-Z screenshots. This beast apparently supports SSE4.

 

[Marmion]

All of the merom cpus (eg merom/conroe/allendale/woodcrest/kentsfield/cloverton) support SSE4. We've known this for quite sometime

 

[n7]

Originally posted by: dmens

Geez, why dont they just slap 2 gigs of cache on the thing and call it a day? Seems like their favorite solution to whatever and limitations the cpu and its platform might have - more cache.

Damn, if it were that easy, I might as well just go home at 2pm every day. All that crazy crap I have to do to eek out 1% extra int perf is just a waste of time anyways.

Finally you've seen the light!

Now go quit your job; they don't need you; cache is the answer to everything

 

[coldpower27]

Well what did you expect it's a 2x4MB configuration, it's basically a Presler but instead of using 2 Cedar Mill's it's 2 Conroe's Hey if it works, and cna get to market quick whos to complain?

 

[swtethan]

*looks around for amd* nowhere... lol

 

[FelixDeCat]

Originally posted by: DeathBUA
Not to sound like a fanboy in either direction, but unless AMD has a counter to this and quickly they'll be beaten like a red-headed stepchild....

:laugh: :laugh: :laugh:

 

[DrMrLordX]

AMD has that goofy 4x4 thing coming, not that buying two FX chips will be cost-effective versus Kentsfield. In any case, dual-core is nice but quad-core is a little difficult for most end-users to utilize (as indicated by some of Duvie's dual Opteron 270 benchmarks). Whoever rolls out "reverse" hyperthreading will probably be the first to make quad core on the desktop truly effective.

 

[Furen]

AMD's reverse hyperthreading thing is something for its next generation architecture, I believe, so this may be used (if it works as AMD hopes it will) sometime in 2008. Personally I think I'll go 4x4 IF regular X2s work on it, since being able to buy a single chip (for cheap, no less) and upgrade to quad core later (or an accelerator, if they ever come out) is quite a nice option. Of course the motherboard have to be cheap, too.

 

[dexvx]

Originally posted by: Furen
Personally I think I'll go 4x4 IF regular X2s work on it, since being able to buy a single chip (for cheap, no less) and upgrade to quad core later (or an accelerator, if they ever come out) is quite a nice option. Of course the motherboard have to be cheap, too.

If regular X2's work, that defeats the purpose of the inherent cost inflation of the 4x4.

 

[ahock]

I agree dexvx. Which they might as well scrap their FX line if that's the case.

 

[Furen]

Huh? People still bought FX-57s even though they were single-core and AMD offered dual-cores. The same can be said here, there will always be someone willing to pay a huge price premium for an extra speed bump even if there are superior alternatives. If 4x4 is FX only it'll die a slow and painful death, since it will be 2x as expensive as Intel's Conroe XE and the Opteron 200 series is already a viable alternative (to get 4 cores) that can end up being cheaper. AMD HAS to push 4x4 as a mainstream dual-socket platform, and the only way to do so is by keeping its price low enough. I dont see why you're assuming that there will be an inherent cost inflation to the 4x4 implementation besides the additional motherboard costs, AMD is likely releasing this platform to blunt the impact Conroe will have on enthusiasts and pricing it out of reach will only drive people to buying Conroes. AMD will make money regardless of which CPU people buy, and people who end up buying 2x 3800s are probably those who would buy a single mid-range chip anyway.

 

[evenmore1]

No fair!

Why can't I test new chips?

Yeah, Intel finally decided to step up to the plate. This is good stuff. Hopefully K8L will match the performance, so then I can get the cheaper one!
Yeah, AMD has to push 4 cores through first, or they might lose too much money. Not that I even see the need for 4 cores yet, I mean not that many programs use 2 cores anyway. Heck, I'm still using 130nm Athlon XP @ 1.725Ghz and even at stock it is sufficient for most tasks. If I wanted to play games, I would use my Playstation 2. Video encoding....well I'd watch a movie while it is encoding Don't do it to much anyway...

 

[sindows]

Originally posted by: munky
Geez, why dont they just slap 2 gigs of cache on the thing and call it a day? :roll: Seems like their favorite solution to whatever and limitations the cpu and its platform might have - more cache.

Right, so I'm getting the feeling that you know better than some of the world's top engineers working for one of the world's largest companies. So, what woulld've been the better choice for Intel to take?

 

[Hyperlite]

Originally posted by: sindows

Originally posted by: munky
Geez, why dont they just slap 2 gigs of cache on the thing and call it a day? :roll: Seems like their favorite solution to whatever and limitations the cpu and its platform might have - more cache.

Right, so I'm getting the feeling that you know better than some of the world's top engineers working for one of the world's largest companies. So, what woulld've been the better choice for Intel to take?

dude....chill....

What munky said does seem to be true, though it may be for other reasons. I'm sure someone around here is knowledgeable enough to tell us why cache sizes keep increasing. It could just be the evolution of the processor, or there could be a reason it has to keep increasing, or all of the above. I bet Viditor knows, ask him.

 

[Griswold]

Originally posted by: Questar
80% scaling from two to four cores. So much for the FSB holding things back.

Hmm where does scalability really starts to show its ugly face? Right, that was beyond 4 cores or even worse, 4 sockets. :disgust:

 

[n19htmare]

If it gets the job done, what is your problem with that.
Conroe has a 4MB share cache...the Kentsfield is like two conroe cores ....4 for each so 8MB... sounds logical. Cache helps Intels architecture.. more the marrier.

some of you guys are just whining, If AMD puts 8MB of cache on their K8L, you guys would be praising and kissing AMDs ass.

If doing something benefits you, why not do it? what is SO wrong with the cache? if it helps, it helps.

Now I can understand it if you are trying to say that cache helps inflate benchmark numbers. well, guess what genius, Conroe wiped the floor with any AMD chip thats currently out including the AM2 chips in nearly every benchmark...from synthetic, to encoding to gaming.....