sblantipodi
2[H]4U
- Joined
- Aug 29, 2010
- Messages
- 3,765
Hashwell will arrive soon.
Post here all rumors or news.
Post here all rumors or news.
Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
Yes... What is thsi hashwell you speak of?
Hashwell is code for Hookers&Blow I believe. At least I though so...well, TBH, this thread is terribly disappointing so far.
Also, do not think that there is a direct upgrade path from Sandy/Ivy Bridge to Haswell, there is a socket change over. Going from LGA1155 to LGA1150.
Well, I'll do a bit of paraphrasing
Significantly stronger iGPU.
Integrated VRMs. <--- VERY IMPORTANT CHANGE FOR US ENTHUSIASTS
Top model will be quad-core with hyperthreading.
Stronger CPU.
Rumored to have L4 cache for sharing between CPU and iGPU.
16 PCI-E 3.0 lanes
Dual channel DDR3
uses LGA1150.
I think that sums it up for what we know so far.
LGA1150? Where did you read it?
Is it a rumor or a confirmed info?
is there some news on the chipset who will drive the LGA1150?
It's all about Haswell-E !!!
LGA1150? Where did you read it?
Is it a rumor or a confirmed info?
I wish my life would go slower.
Intel internally calls Haswell its fourth generation Core processor micro architecture, and we believe that Intel counts only major steps in its development.
Its first Core architecture was called Conroe /Merom in 65nm, followed by Nehalem in 45nm, Sandy Bridge in 32nm and Haswell will be the second 22nm core with a new architecture.
Intel still plays its old tick-tock game and Haswell is a major step, or a tock, while Broadwell is 14nm shrunk version of the Haswell architecture. A more obvious example is Sandy Bridge in 32nm, a tock in Intels development cycle, while the most recent Core processor is a tick, and it is based on 22nm Ivy Bridge core.
Now Intel tells its partners to expect that Haswell should end up at least 10 percent faster than Ivy Bridge based cores at the same clock. These numbers are based on pre-silicon projections that Intel always does before it gets the working prototype back, but since we are some month away from the IDF 2012 in San Francisco, we are quite sure that we will see Haswell again, much closer to its final design, and we even expect to see Broadwell prototypes to hit at least one of the keynotes.
Intel also hints at enhanced based overclocking and end user tunability, which sounds promising to many enthusiasts. If all goes well for Intel, Haswell should be quite a nice piece of hardware once it ships in Q2 2013.
and that relatively small increase should give AMD some time to catch up.
I'm still rocking an 1156 setup. That sucks that there is going to be yet another socket change... I guess this is why I'm still using an 1156 setup and have not upgraded.
10% higher IPC. That's a bit disappointing, and that relatively small increase should give AMD some time to catch up. Looks like Intel really is focusing heavily on the iGPU side.
I'll reserve judgement until we see how well it clocks. If clocks are similar to current offerings, then it's very disappointing. If clocks are at least 10% higher in the same thermal envelop, resulting in a 20% increase at the same wattage, then it's a decent step forward. If clocks are 20% higher...
When you keep adding features to the CPU, it's impossible to keep the same socket.
I guess that's true. I miss the days of a simple CPU upgrade without having to replace the motherboard.
I do not think that gives AMD much of anything. I mean if AMD has a 15% IPC improvement that would be a net gain of around 5% however they are behind in IPC by so far its not like 5% would eat at all into the lead.
Intel has increased IPC with every new generation for the last many years. AMD has decreased it, with BD in particular. I'd like to know how this math works where AMD catches up...
Except AMD is also upping up their clocks quite a bit.
IPC isn't everything, how many times does that need to be said? AMD aimed for a high clocking design. As a result, they sacrificed some IPC for high clocks, but they failed to achieve their high clock target.
So they failed in clock speed and IPC, which means down on both multiples... Now you've got me wondering even more how this math is going to work out? Sorry, but AMD is not going to make any ground on Haswell.