Socket 940 - 939 -754 future?
Rumor has it that Socket 754 processors may see 1 or 2 more speed bumps then follow the path that Athlon XP is now on...to retirement. Socket 940 is the current king but OEM consumers may hesitate to purchase because of the high price combined with the processor requiring the more expensive ECC Registered memory.
Combine that with the coming launch of Socket 939 which operates with the cheaper (and what all of us already have) unbuffered non-registered memory and the epitaph for Socket 940 seems to be writing itself.
Or is it?
Blue sky assumption #1: Could the upcoming Socket 939 follow the same build path as Socket 754. IE: a single channel memory controller vs. Socket 940's dual channel memory controller? This would keep Socket 940 just above Socket 939 for features and, perhaps, performance. Socket 940's marketable life has been given a second chance. This would make sense because it would be odd for AMD to make a processor (Socket 940) obsolete only two models into its life.
Socket 940's naming convention for the FX series jumps in odd number increments, FX-51, FX-53...a FX-55?...so it seems apparent that there is a planned future path for Socket 940.
Blue Sky assumption #2: Could it be that AMD will just phase out 32-bit processors altogether over the next 2-3 years. AMD sits on the poop or get off the pot stage between 32-bit and 64-bit. We all know that the AMD 64-bit processor is backwards compatible to 32-bit. Since they have a "forwards" and backwards compatible chip...why make plain old 32-bit processors anymore?
But this is a five year plan. Longhorn MAY make it to market by 2006 but that will definitely be 32-bit. Who knows when a fully fledged 64-bit only version will come out and if it will be able to run 32-bit software in a virtual desktop...
You now start to see the factors involved.
So what are your blue sky assumptions/prophecies?
Combine that with the coming launch of Socket 939 which operates with the cheaper (and what all of us already have) unbuffered non-registered memory and the epitaph for Socket 940 seems to be writing itself.
Or is it?
Blue sky assumption #1: Could the upcoming Socket 939 follow the same build path as Socket 754. IE: a single channel memory controller vs. Socket 940's dual channel memory controller? This would keep Socket 940 just above Socket 939 for features and, perhaps, performance. Socket 940's marketable life has been given a second chance. This would make sense because it would be odd for AMD to make a processor (Socket 940) obsolete only two models into its life.
Socket 940's naming convention for the FX series jumps in odd number increments, FX-51, FX-53...a FX-55?...so it seems apparent that there is a planned future path for Socket 940.
Blue Sky assumption #2: Could it be that AMD will just phase out 32-bit processors altogether over the next 2-3 years. AMD sits on the poop or get off the pot stage between 32-bit and 64-bit. We all know that the AMD 64-bit processor is backwards compatible to 32-bit. Since they have a "forwards" and backwards compatible chip...why make plain old 32-bit processors anymore?
But this is a five year plan. Longhorn MAY make it to market by 2006 but that will definitely be 32-bit. Who knows when a fully fledged 64-bit only version will come out and if it will be able to run 32-bit software in a virtual desktop...
You now start to see the factors involved.
So what are your blue sky assumptions/prophecies?
0
Comments
Socket 940: Opterons, registered mem
Socket 939: FX CPUs making the move to regular RAM, dual channel
Socket 754: The Athlon 64, single channel, keep XP rating. Maybe also move the XP line to this socket as well but have the 64 bit part disabled.
Perhaps 939 will come in two design types? Presently there's Athlon 64 and there's Athlon FX. Athlon 64 is the less expensive and FX is the performance king (more expensive).
It'd make motherboard manufacturer's happier if they could come out with one socket type that would support both rather than a 754 pin version and 939 pin version of the same motherboard.
Maybe they know more than us and are not allowed to say. I would hope so anyway.