Hacker News new | ask | show | jobs
by programminggeek 4310 days ago
I really don't keep up on this stuff much, but why is this still Haswell based? Why not just do this on Broadwell?
2 comments

Intel has a pattern of releasing variations of microarchitectures with higher core counts after their immediate successors have been announced/launched. Think Sandy Bridge-E after Ivy Bridge, Ivy Bridge-E after Haswell, and now Haswell-E.

They probably do this because it's probably to easier to manage the larger die size and greater complexity as the process has matured and yields have improved.

Also, these are essentially rebadged Xeon's. Xeon's have higher requirements for maturity, testing, stability, et cetera.
And a lot of the time, many of these CPUs are almost identical to one another. They will take a single CPU and brand it in 10 different ways depending on how it does during manufacturing testing.

For a given CPU that comes off the production line, the max frequency it can run at will vary from chip to chip. As well, if there are dead cores on that chip they can just disable them and sell the chip at a lower price (though, intel may not do this; nvidia has in the past with graphics chips).

"Extreme Edition" is normally what they've done at the end of a lifecycle. Broadwell's been delayed it appears, due to process difficulties possibly? This may fill the gap... a little.