so this is my theory on why e7200 dies are different than e8200 dies and are not just disabled cores.
normal wolfdales are c0 or c1 stepping.
these are 45nm m0 stepping.
m0 stepping is a different die. if you go on intel's site, and buried in spec documents, you can see that wolfdale-3M cores for mobile chips are classified into 3 different types.
6mb cache
3mb cache on 6mb cache die
or 3mb cache
intel doesnt usually give a different stepping number for cores that have disabled cache (hence an L2 stepping allendale 2mb e4300 is called the same thing as an L2 stepping allendale 1mb e2160 with half cache turned off, same applies with B2 stepping e6420 vs B2 stepping e6400) so i'd assume if there were any chips with a 6mb die disabled down to 3mb it'd still be called c0 or c1 stepping. so i'm assuming q9300 and e7200 cores with m0 or m1 stepping are this "3m cache" die. i think its a fairly safe assumption since the e7200 has less capacitors and fuses on the bottom than a e8200. at least from looking at other dies intel has made (such as say e4300 and e2160 based on the same L2 core) the number of caps and fuses stays the same even if they disable cache on the same core. so the bottom of a e6300 and e6320 looks the same if they are both using B2 dies (the L2 stepping e6300 has a different looking bottom). for example check out this picture.
http://www.xbitlabs.com/misc/p...m-e2160/cpus-2.jpg&1=1
4mb e6320 on the left, 2mb e4300 in the middle, 1mb e2160 on the right. the e4300 and 2160 are both L2 allendales and have the same looking bottom.
http://www.xbitlabs.com/images...re2duo-e4300/photo.jpg
that picture onthe left is a e6300 b2 core (disabled cache) , and L2 core e4300. notice the e6300 looks the same as the e6320 from the first pic.
now here is the e7200 vs. an e8400
http://www.xbitlabs.com/images...re2duo-e7200/photo.jpg
notice they are totally different.
anyhow, if you go through the mobile cpu documentation only it actually gives the measurements for the die in physical specifications since mobile chips dont have a heatspreader . on the desktop one you dont get anything since it has a heatspreader so they dont give the measurements (i think they give these out for heatsink manufacturers and such). anyhow, the e7200 M0 core die (and i'd assume the same as the m1 core on the q9300 and of the wolfdale-3m mobile chips) have a die size of 81mm^2 (at least from the mobile specs) . These die are not the same as the c1 dies, so i think its possible that is the reason for what seem to be low overclocks in most of the reviews i've read. The L2 dies on the first allendales also did not oc as well as the initial B2 or lower dies of the first core 2s either. It wasnt until the m0 core allendale dies that it got better.
anyhow thats something to be aware of. i think pretty much every site on the internet has it wrong except a few japanese ones that had some presentation charts. wikipedia seems to say the e7200 and e8200/e8400 have the same 107 mm die.
normal wolfdales are c0 or c1 stepping.
these are 45nm m0 stepping.
m0 stepping is a different die. if you go on intel's site, and buried in spec documents, you can see that wolfdale-3M cores for mobile chips are classified into 3 different types.
6mb cache
3mb cache on 6mb cache die
or 3mb cache
intel doesnt usually give a different stepping number for cores that have disabled cache (hence an L2 stepping allendale 2mb e4300 is called the same thing as an L2 stepping allendale 1mb e2160 with half cache turned off, same applies with B2 stepping e6420 vs B2 stepping e6400) so i'd assume if there were any chips with a 6mb die disabled down to 3mb it'd still be called c0 or c1 stepping. so i'm assuming q9300 and e7200 cores with m0 or m1 stepping are this "3m cache" die. i think its a fairly safe assumption since the e7200 has less capacitors and fuses on the bottom than a e8200. at least from looking at other dies intel has made (such as say e4300 and e2160 based on the same L2 core) the number of caps and fuses stays the same even if they disable cache on the same core. so the bottom of a e6300 and e6320 looks the same if they are both using B2 dies (the L2 stepping e6300 has a different looking bottom). for example check out this picture.
http://www.xbitlabs.com/misc/p...m-e2160/cpus-2.jpg&1=1
4mb e6320 on the left, 2mb e4300 in the middle, 1mb e2160 on the right. the e4300 and 2160 are both L2 allendales and have the same looking bottom.
http://www.xbitlabs.com/images...re2duo-e4300/photo.jpg
that picture onthe left is a e6300 b2 core (disabled cache) , and L2 core e4300. notice the e6300 looks the same as the e6320 from the first pic.
now here is the e7200 vs. an e8400
http://www.xbitlabs.com/images...re2duo-e7200/photo.jpg
notice they are totally different.
anyhow, if you go through the mobile cpu documentation only it actually gives the measurements for the die in physical specifications since mobile chips dont have a heatspreader . on the desktop one you dont get anything since it has a heatspreader so they dont give the measurements (i think they give these out for heatsink manufacturers and such). anyhow, the e7200 M0 core die (and i'd assume the same as the m1 core on the q9300 and of the wolfdale-3m mobile chips) have a die size of 81mm^2 (at least from the mobile specs) . These die are not the same as the c1 dies, so i think its possible that is the reason for what seem to be low overclocks in most of the reviews i've read. The L2 dies on the first allendales also did not oc as well as the initial B2 or lower dies of the first core 2s either. It wasnt until the m0 core allendale dies that it got better.
anyhow thats something to be aware of. i think pretty much every site on the internet has it wrong except a few japanese ones that had some presentation charts. wikipedia seems to say the e7200 and e8200/e8400 have the same 107 mm die.
Facebook
Twitter