Servers
More 'Son of Nehalem' details leaked
A few more details emerged Wednesday about processors to be based on Intel's next-generation Sandy Bridge microarchitecture - and if you've recently invested in a socket-LGA1156 motherboard, it appears that it won't accept Sandy silicon. The Sandy Bridge news doesn't come from Intel itself - and the company didn't respond to our …
This topic is closed for new posts.
Who told them to change the socket?
Management to engineering, "Change the socket so they'll have to buy our motherboard chipsets all over again."
What a slap in the face.
Many years ago
RAID is software (sometimes buried in firmware). If that software is soldered to your motherboard you will have hassle accessing your data when the motherboard dies.
PS in this decade we use solid state disks.
Never
"""When will AMD Release a decent on board RAID solution?"""
Nobody will ever make a 'decent' onboard RAID. It just won't be done. Intel's Matrix is a load of crap and is to be avoided.
Your options are (Ordered by cheapest first):
- Software raid in a decent OS
- Real hardware raid in a pci-express card, with real cache
- SAN
Most of these onboard 'raid' devices require specific Windows drivers to work, which means that the raid is implemented in the driver, not the hardware. And you have no idea what those drivers are doing. Use software raid built into your OS, then at least it's probably gone through some QA procedures.
Is anybody else reminded
Of when AMD couldn't decide on one socket, and seemingly overnight jumped from socket A, to 754, to 939, to AM2/AM2+ seemingly overnight?
No I am not reminded at all in fact.
After all socket A was for the K7 (athlon) 32bit chips (and duron). That was around the time intel was making the P3 (using the socket 370 as far as I recall).
The Athlon 64 on the other hand (K8) came with either socket 939 (dual channel memory) or socket 754 (single channel, for cheap systems). The opteron had a socket 940 and used dual channel registered memory (servers often do).
A few years later, then moved to socket AM2 when they changed form DDR to DDR2 memory, and they dropped the single channel option entirely. AM2+ is just an update that supports AM2, AM2+ and AM3 chips (AM3 chips are DDR3 and DDR2 compatible as far as I understand things).
During that time, intel has made the socket 423 for the P4, the socket 478 for the P4, the socket LGA775 for p4, the socket LGA1366 for core i7 (tripple channel), the socket LGA1156 (for core i7/i5/i3 dual channel), and are apparently now considering an LGA1155 (for the same line as the 1156 except it's new and hence different). Not onyl that, there have been cases where even if you had the right socket, your voltage regulator wasn't the right kind for some new chips, so you needed a new board with the same socket but a new voltage regulator.
Clearly intel is totally incompetent at planning ahead and designing future proof sockets. Or they in fact don't care about customers and simply want to force you to buy all new stuff rather than incrementally upgrade. Given the P4 was clearly spec'd out by marketing and not engineering, that last option is probably the case unfortunately.
Sandy Bridge
Better a Sandy Bridge than a Castle Built on Sand, I suppose.
just have to....
AMD r so much bettr then smelly intel rawr rawr AMD wooyay
FLAME ON!!
more intel BS
No USB 3.0 and that damned LightPeak instead... this is insult to injury on top of the socket change.
So much for giving the customer what they want, eh?
The pins don't look more numerous
What's interesting is I understand this change is required because the gpu is integrated. I had heard this resulted in increasing the number of pins. While I didn't count them, I don't really see any increase in the number of pins so I do question the need for this change. Does changing the voltage really help them that much?
Did you read the article?
Because there's one less pin. LGA 1156 vs 1155.
This topic is closed for new posts.
