Recent content by Gianti

Yes, there will be desktop parts for 2.8GHz dual-core NGA (Next Generation Architecture, trying to avoid the K8L / K10 names) parts. But that should be Kuma with 2MB shared L3 cache. The binned clock is for eng samples. The part we have here might not even surface. At the current state Santa...

Wait till IDF, AMD will have to respond to it with something.

I know you are not modifying the data above. But as far as I know, the caching system you've tested is even worse than the AMD K8 architecture. So I doubt it. :? And what have you tested the "Rana" 2.8GHz CPU? And what is the Intel CPU you tested? I live on the intel's side of the things and...

I'm no tech guru, but those specs look right to me. Anyone else chime in? :oops: eng samples =) duals are complete, quads are cut. You will see from the pics tonight, need to get home to post them

That was from the runs, tested with differently binned CPUs.

Hey thx for backing me up. I have to say that I was a little misled, as I was working on the Tigertons for awhile now. The Barcelonas are binned 1.2 and 2.0 in eng samplesm that's for 4 cores, and 2.8 for 2 cores. They will hit the market with 2 core chips this summer. Quads will come in...

Not possible, Barcelonas will top at about 2.3GHz this year. In early summer you will see OEMs coming out with 4 cpu servers having 2.8 Barcelonas and up 128G of RAM on board. I have a few early builds in the lab that are still not ready for prime time, but in a month or so it will be taping...

BTW the Barcelonas used in testing are p1 2.8GHz samples

I mean that Santa Rosa (2c) and Barcelona (4c) are very unimpressive with engineering samples that were received lately, not talking about P0 versions (7 months ago). It's not the best practice to compare these vs C2D, as they are server CPU's. However, I've seen that there is no way you can...

There is no way that can be determined since there are no benches for the new cpu's from AMD so your statement is just utter crap. Crap? Don't think so. I've seen them and worked with them, so can't be really crap. Engineering versions of these CPUs were out a long time ago. Where did you...

There is no way that can be determined since there are no benches for the new cpu's from AMD so your statement is just utter crap. Crap? Don't think so. I've seen them and worked with them, so can't be really crap. Engineering versions of these CPUs were out a long time ago.

Don't wait for AMD... they wont have anything good in the upcoming months, just get the C2D after the price drop. From all the specs it looks like E6600 and higher will rip upcoming AMD's CPU apart.

Intel has it and you will see first 4x4 Intel based products arriving this fall. About AMD and Intel... well AMD s still stron in the server market due to the mem controller... however it wont last long. And Barcelona wont be able to hold against Tigerton and Dunington from the Intel, which...

Just out of interest, how did you benchmark the disks, because sequential writes and random access writes are 2 different things. Can't get into details but using server testing software, developed internally at the company I work for. Takes about 8 hours to run over 100 tests under variable...

FSB is quad pumped - four operations per clock (rising, high, falling, low) DDR is dual - two operations per clock (rising, falling) Stock settings should be for FSB 1066/4 = 266 and DDR 800/2 = 400. If you mess around with clocks they may change. When overclocking it is recommended to change...