Nehalem到底有没有集成内存控制器呢？哈哈哈哈？

Prescott

两年前我的预言是多么正确阿。No IMC on desktop!! :lol:


http://www.theinquirer.net/default.aspx?article=38232

Intel's Nehalem has integrated memory controllers

Codemappery

By Charlie Demerjian in Hanover: 星期四 15 三月 2007, 08:45
INTEL IS KEEPING quiet about its upcoming strategy for integrated memory controllers (IMC).

We all know they are coming, and even Kicking Pat Gelsinger confirmed it a week or two back. but what about details? The results will surprise you.

The one that is pretty well known about is the IMC for Stoutland, that part defintetly has an IMC. The config there is FB-DIMM 1 or 2 in a 4 + 1 channel setup. This is per CPU mind you, so a 4S Stoutland will have 20 FBD channels.

The volume part, Thurley, will have three DDR3 channels per socket. The memory is currently slated to run at 800/1066/1333, so bandwidth will not be a problem there. We understand Intel is adopting the micro-buffer strategy from AMD.

For laptops, things are quite the opposite, sources tell us that there will be no IMCs in laptop parts at all, quite a curious turn.

Desktops get murky quickly, but the vast majority of them will not have IMCs either. We hear the only candidate for IMCs on desktop parts are the EE/XE parts, extreme as in Jello.

When questioned why, Intel spokespeople say "we can not comment on future products", but other sources give us a range of answers. Some cite flexibility in memory architectures, others engineering bandwith, and that guy asking for quarters outside Moscone said it was sunspots interfering with the CIA mind control orbital lasers. No two will give you the same answer though, and that we find worrying.

Toss in that there are at least four sockets coming for each application, and you have the potential for a right royal mess. Add in Jasper, a year or so later, and you have the potential for a short lived mess to be supplanted by a slightly longer life short lived mess.

In the end, it works out to be IMCs on Xeons and EE/XE, and nothing else. IMCs come in two flavours, possibly more if the moon is in the seventh house and Intel does something odd with the EE/XEs. Add in sockets aplenty, and you have Intel circa Q4/2008 and on. µ

[ 本帖最后由 Prescott 于 2007-3-15 20:45 编辑 ]

ITANIUM2

把楼上拖出去狠狠地打 w00t)

itany

要是桌面不使用IMC，YY度立马就下降了一半～
鄙视啊！ :mad:

HeavenPR

and that guy asking for quarters outside Moscone said it was sunspots interfering with the CIA mind control orbital lasers.

我只才能说 ... W T F w00t) w00t) w00t)


CSI 提供的带宽足够取代 FSB 就好了，有没有 IMC 确实没啥不同，IMC 只有在 NUMA 的时候才起作用，试问哪个桌面用户玩 NUMA？？？ 站出来给我看看 w00t)

HeavenPR

你家里什么机子？ AMD 4X4?? 如果是，我拜你一个~ w00t)

HeavenPR

那叫 NUMA 嘛？！ 你瞎站出来干啥，赶快坐回去！:lol:

Prescott

原帖由 itany 于 2007-3-15 21:13 发表
要是桌面不使用IMC，YY度立马就下降了一半～
鄙视啊！ :mad:

想要YY还不简单，买Xeon的桌面版咯，不过性能不见得有多大提高阿。 :lol:

rage6

弱弱的温一下~ numa是什么啊?

HeavenPR

http://chlmx.spaces.live.com/blog/cns!404202A2C27CA6FB!408.entry

zaarath

原帖由 rage6 于 2007-3-15 22:46 发表
弱弱的温一下~ numa是什么啊?

Non-Uniform Memeory Acess

针对多处理器系统的技术。简单地说就是给处理器分配专有内存，访问内存的速度由内存离处理器的距离决定，每个处理器访问自己专有的内存是最快的，访问别的处理器的内存延迟会大大增加。

[ 本帖最后由 zaarath 于 2007-3-15 23:04 编辑 ]

zaarath

桌面没有IMC, 我觉得就不会有CSI。桌面可能还会继续沿用老将廉颇-FSB, 现在改名叫PSB了:p 。
我看这也挺好，桌面+移动这种commodity的东西用一套体系，xeon+itanium用一套体系，没必要用一个完全统一的架构，毕竟这两种应用环境完全不同。如果说server应用能从IMC+CSI得到100分的好处，桌面可能只有20分。现在的771xeon已经同775老死不相往来了，差别在大点也没什么。

如果Intel在桌面继续扛着FSB的大旗，一大部分也是从自身芯片组生意出发的考量。把MC和系统互联交给北桥可以有效降低cpu的生产cost, 增加chipset的售价。还可以提高chipset市场的门槛，VIA/SIS为什么萎靡不振? 因为内存控制器技术不行！

另外，charlie的文章。。。唉，我倒是希望能相信他一次。

89度热水

对扣肉这种缓存大内存预取出色的架构来说集成MC降低延迟的作用的确不会太明显

89度热水

活跃度还在，太好了:loveliness:

Prescott

原帖由 89度热水 于 2007-3-16 00:05 发表
对扣肉这种缓存大内存预取出色的架构来说集成MC降低延迟的作用的确不会太明显

也可以集成MC，省点缓存嘛 :p

不过集成MC那些地方，说不定都够做2M L2的。

Edison

这里的4+1通道应该怎么理解。

Prescott

原帖由 zaarath 于 2007-3-15 23:23 发表
桌面没有IMC, 我觉得就不会有CSI。桌面可能还会继续沿用老将廉颇-FSB, 现在改名叫PSB了:p 。
我看这也挺好，桌面+移动这种commodity的东西用一套体系，xeon+itanium用一套体系，没必要用一个完全统一的架构，毕 ...

据某位A Fan的说法，800Mhz的前端总线已经是一个奇迹了。w00t) 当然这位A Fan现在已经改名换姓，再也不踏足本版一步了。

不过，用CSI替代FSB其实还是不错的。

另外，charlie的文章，曾经是非常准确的。。。。

Prescott

原帖由 Edison 于 2007-3-16 00:09 发表
这里的4+1通道应该怎么理解。

memory raid 5 w00t)