3500+ San Diego?

[Koyanisquatsi]

lists a 3500+ with 1MB L2 cache and a stock speed of 2.0ghz. Same price as the 3500+ Venice. This is getting confusing.

http://www.xtremesystems.org/forums/archive/index.php/t-59210.html

 

[bunnyfubbles]

How is it confusing? If its true its great news. San Diego should be the best silicon AMD is using for their single core stuff, which should mean higher potential maxes, as well as more tollerance for cold (great news for the extreme OCers). The extra cache should be a nice little bonus as well.

People need to stop looking just at the code name, and start looking at the rest of the chip, such as the core speed, the L2 cache, the fab size...

 

[SynthDude2001]

That would be pretty sweet if you could get a $250 chip with 1MB cache to >2.8GHz...

 

[n yusef]

I might have to wait a little bit and but one of those instead of a 3200+. I wonder how much 1MB L2 affects real gaming. Chilly1 has a thread over there where he took a 3800+ Venice to 3.8GHz, but he got like 3000 less 3DMarks (01) than an FX-55 would have. I just hope 1MB matters as much in real games as 3DMark01SE.

 

[cjsketchy]

Impact of L2 cache on games (at 1.6ghz): http://www.xbitlabs.com/articles/cpu/display/sempron-2600_5.html
makes for a nice little boost

 

[bunnyfubbles]

It is next to negligable when you reach high speeds (2.4GHz+, especially if we're looking at 2.8-3GHz with these Venice/San Diego chips). As I said before, the advantage to SanDiego should be the cream of the crop silicon moreso than the cache. The average high will probably be greater with the San Diego vs. the Venice chips. The 2.2GHz 512KB 3500+ is going to be faster than a 2GHz 1MB 3500+, there's no doubt in my mind about that, but when each chip is overclocked to their maximum the San Diego core should be faster in the end than the Venice, although it is going to be hard to beat the value of the 3200+ and 3000+ Venice chips if they can do 2.8GHz consistently on stock voltage.

 

[CaBoOse999]

Isn't San diego the fx series?? Why would it be listed as a 3500 then?

 

[imported_SLIM]

San Diego = strained silicon + silicon on insulator + 90nm fab process + 1MB L2 cache
Venice= strained silicon + silicon on insulator + 90nm fab process + 512KB L2 cache
Palermo= strained silicon + silicon on insulator + 90nm fab process + 256KB L2 cache

AMD will use san diego for their higher end A64s and FXs, venice for most of their A64s, and palermo for semprons.