socket 754, needs.
2800+ CH 1800 mhz 9x200 512Kb L2
3000+ CH 2000 mhz 10x200 512Kb L2
3200+ CH 2000 mhz 10x200 1024Kb L2
3400+ CH 2200 mhz 11x200 1024Kb L2
NC and CH 3400+ have differrent multiplier, same for 3200+
Printable View
socket 754, needs.
2800+ CH 1800 mhz 9x200 512Kb L2
3000+ CH 2000 mhz 10x200 512Kb L2
3200+ CH 2000 mhz 10x200 1024Kb L2
3400+ CH 2200 mhz 11x200 1024Kb L2
NC and CH 3400+ have differrent multiplier, same for 3200+
Isnt the 3000+ 1MB cache for clawhammer?
The only 1mb Ch are 3200+, 3400+, 3700+Quote:
Originally Posted by NiCKE^
The 2800+, 3000+ CH have half the L2 cache disabled.
See HERE
And HERE
Note the code of the chip.
ADA3000AEP4AP
AP is Clawhammer CO revision.
Many review sites got it wrong like anandtech it was a little confussing at the time. The real newcastles surfaced a little later. It was rumored that the cache crippled CH were had defective Cache and were made good by disabling half of it and marking them down as 3000+. I'm not sure about this theory, its possible.
Is this right:
ClawHammers:
2800+ 1.8ghz 512KB
3000+ 2ghz 512KB
3200+ 2.2ghz 1mb
3400+ 2.4ghz 1mb
3700+ 2.4ghz 1mb
NewCastle:
2800+ 1.8ghz 512kb
3000+ 2ghz 512kb
3200+ 2.2ghz 512kb
3400+ 2.4ghz 512kb
It doesn't make sence, why do a 3700+ CH when the 3400+ is the same and why make two diffrent cores which got the same spec in 3000+ and 3000+ ?
ClawHammers:Quote:
Originally Posted by NiCKE^
2800+ 1.8ghz 512KB
3000+ 2ghz 512KB
3200+ 2ghz 1mb
3400+ 2.2ghz 1mb
3700+ 2.4ghz 1mb
NewCastle:
2800+ 1.8ghz 512kb
3000+ 2ghz 512kb
3200+ 2.2ghz 512kb
3400+ 2.4ghz 512kb
;)
Thanks, I already figured that out by myself but thanks anyway :D
don't forget about the palermo
Some info on Palermos, using my chip as an example (a 2800+ 90nm Sempron) which runs happily at 320MHz btw:)
SDA2800AIO3BA
CBBID0501GPMW
S=Sempron
D=Desktop, M=Mobile
A=62W, S=25W
2800=speed
A=lidded (heatspreader), B=OEM unlidded
I=1.4V, O=1.25V
0=69°C, X=95°C, P=70°C
2=128k L2, 3=256k L2
BA=DO step, 90nm SOI (Palermo)
BO=EO step, 90nm SOI (Venice)
AX=CG step, 130nm
LB=DO step mobile
older versions of CPU-Z showed it as a Paris whereas v 1.28.6 show it as a Palermo
The table is done, we just need to fix some things with it so it will fit good.
http://home.no/nickeh/AMDcputablexs.orgedition.txt
This is how it should look but we get a huge between the tables and we don't know how to correct it.
those palermo's are .09 not .13 only the paris's are .13
OKi updated..
Good extensive work but AMD now do a quick desktop reference guide (hence unfortunately excluding mobiles) for all A64s including X2's and FX's
www.amdcompare.com
Dont forget this one.
The 2800 is
SDA2800AIO3BA
CBBID0503CPAW
1196078B50277
The 3300 is
SDA3300AIO2BO
LBBLY0518BPAW
Z645517E50339
http://home.bendbroadband.com/bypolar/Picture%20433.jpg
And don't forget teh Sempron 2500+, which is soon to be released on s754 too with E3 stepping (aka venice), 90nm, 1400mhz stock and 256kb cache.
link in case that sayaa starts whining about truth ;)
may i ask for more info on this 3300+ plz bypolar
Here is a review from Annandtech Sempron 3300
I my self have not even opened the box yet.
okay thank you very much, when you do open it could you send me a cpu-z screenie for my enjoyment???
Anandtech only got 2.4ghz out of the 3300? That's rather dissapointing.
Let see how far I get.
http://home.bendbroadband.com/bypolar/3300.jpg
could you verify with central brain identifier, thought 3300+ was palermo, tyvm bypolar
AMD Sempron 3100+ Model C Revision CG eg 130nm Dublin core @ 1800mhz 256k L2