These are more or less guinea pig chips. Allows Intel to develop some early DDR4 R&D for later use in server and mainstream chips.

No way am I spending huge extra bucks for last-gen cores.

MikeMurphyThese are more or less guinea pig chips. Allows Intel to develop some early DDR4 R&D for later use in server and mainstream chips.

No way am I spending huge extra bucks for last-gen cores.

DDR4 was available in server way before X99 came around.

Steven BDDR4 was available in server way before X99 came around.

DDR4 is still in its very early expensive, relatively slow compared to the potential of the technology days, wondering whether this is exactly what we need... now...

I have seen a review and honestly speaking it wasn't very convincing, yeah, kind of promises some improvements but only this.

My 2600k has 8 MB L3, and the Broadwell-E would have 20 MB L3, I havent seen any comments regarding this fact, will this make a difference in perfomance?

OctaveanI would tend to agree with your assessment.

Past actions are the best indication of future actions IMO. The one thing that Intel has done "consistently" with the entry level HEDT processor offering throughout the generations is be "inconsistent".

For example:

Core i7 3820 - didn't receive the "K" designation and was often referred to as partially unlocked. while it had the 40 PCIe lanes it was only quad core and the other two Sandy Bridge-E processors were 6 cores.

Core i7 4820K - The Ivy Bridge-E entry level chip was much the same as the 3820 but this time presumably actually deserving of the "K" suffix and thus fully unlocked.

Core i7 5820K - Finally 6 core / 12 threads and fully unlocked but a somewhat truncated 28 PCIe lanes.

What is clear is that Intel wants to limit, impede or gimp the entry level processor in a significant way but they just seem undecided as to how to do it. So they keep experimenting with the variables.

I expect some change in the entry level Broadwell-E processor. Although it would be nice if they upped the PCIe limit to something like 32 lanes at least,....assuming they don't go for the full 40.

It's hard to tell with them because as you said they are constantly adjusting between offering something worthwhile at that price point without butchering sales above it or making something not worth the purchase (I felt that the 3820 not being fully unlocked made it a hard to swallow buy even with the additional lanes on top of the 3930X not really making a compelling counter offer over the 3930k). The 4820k alleviated that problem and made a compelling offer but was overshadowed by the fact that Haswell was already available making it only compelling if you want more lanes over the improved single threaded performance and chipset improvements (on top of the same issue being present on the 4960X vs 4930K argument). This round Intel gave us a different compelling buy basically saying you can have the cores and threads but your not getting as many lanes (still a higher than normal amount). Threw me off and honestly added something to consider when buying (especially with boards that contain PLX chips) on top of making all the chips a separate deal with the 5960X being 8 cores this round and the 5930K containing the 40 lanes.

Not sure honestly what to think yet of the entire broadwell lineup will be but I have a feeling it will justify itself as much as the recent Haswell-E has. It won't be completely necessary for everyone of course but it comes down to your needs and your current systems. Plus we never know as Intel could decide to make the middle an 8 core which would shake things up again.

Posts purged for being off topic.

Stay on topic.

Thank you.

GhostRyderIt's hard to tell with them because as you said they are constantly adjusting between offering something worthwhile at that price point without butchering sales above it or making something not worth the purchase (I felt that the 3820 not being fully unlocked made it a hard to swallow buy even with the additional lanes on top of the 3930X not really making a compelling counter offer over the 3930k). The 4820k alleviated that problem and made a compelling offer but was overshadowed by the fact that Haswell was already available making it only compelling if you want more lanes over the improved single threaded performance and chipset improvements (on top of the same issue being present on the 4960X vs 4930K argument). This round Intel gave us a different compelling buy basically saying you can have the cores and threads but your not getting as many lanes (still a higher than normal amount). Threw me off and honestly added something to consider when buying (especially with boards that contain PLX chips) on top of making all the chips a separate deal with the 5960X being 8 cores this round and the 5930K containing the 40 lanes.

Not sure honestly what to think yet of the entire broadwell lineup will be but I have a feeling it will justify itself as much as the recent Haswell-E has. It won't be completely necessary for everyone of course but it comes down to your needs and your current systems. Plus we never know as Intel could decide to make the middle an 8 core which would shake things up again.

Quite right,...

For all we know the tweaking will continue with Broadwell-E and a more compelling configuration may present itself.

However, the mainstream line should become more compelling too if Intel doesn't slow down their cadence too much.

One thing I am almost certain of is that we will have a repeat of the transition from Sandy Bridge-E to Ivy Bridge-E. I recall a number of people questioning why Intel would even bother to release such an (paraphrasing) incremental upgrade on the LGA2011 / X79 / C600 platform. And I remember thinking, something like "so you would prefer one round of processors for a given platform and no more before Intel moves on to a different socket / chipset?".

I think it's worth noting that when SB-E was released, there were up to 8c/16t Xeons available. After IVB-E was released 12c/24t variants were available. Now with Haswell-E we have 14c/28t Xeon parts. So I think it would be safe to assume that Broadwell-E will either bring 16c/32t or 18c/36t Xeons to the table. It's worth remembering that skt2011(-3) were both intended to be server platforms and that Haswell-E was the first Xeon lineup to offer entry-level Xeons as 6c/6t parts. No quads on skt2011-3.

Currently, I`m happy with my i7-4770k ( can`t figure a reason to upgrade ) :)