Talk:Nehalem (microarchitecture)

From Wikipedia, the free encyclopedia

Contents 1 Power 2 news story - added info on memory controller integration 3 Rename article? 4 Compatibility with LGA 775 sockets 5 Cache on Havendale and Auburndale

[edit] Power

"30% lower power usage for the same performance." - compared to what? Penryn? —Preceding unsigned comment added by 77.183.56.218 (talk) 16:27, 13 April 2008 (UTC)

[edit] news story - added info on memory controller integration

http://www.betanews.com/article/Intel_Officially_Confirms_Integrated_Memory_Controller_for_45nm_Nehalem/1175123788 —The preceding unsigned comment was added by 24.22.22.101 (talk) 16:10, 29 March 2007 (UTC).

I was wondering if anyone thinks that we should add a page on the original Nehalem (the NetBurst version). I also don't know where this page should be located. Imperator3733 19:41, 22 April 2007 (UTC)

[edit] Rename article?

Shouldn't the article be renamaed to "Nehalem (microarchitecture)" (together with the Intel Core (CPU architecture) and Gesher (CPU architecture) articles) to avoid using the word "architecture" for microarchitecture (hardware) and instruction set (software) possibly confusing a bit?

I think that maybe they should be renamed to "Nehalem (CPU microarchitecture", "Intel Core (CPU microarchitecture)" and "[[Sandy Bridge (CPU microarchitecture)" instead. Core, Nehalem, and Sandy Bridge/Gesher are microarchitectures, not architectures, but it should be made clear that they are CPUs. Imperator3733 15:36, 24 September 2007 (UTC)

I mentioned at Talk:Intel Core (CPU architecture) a new, revised idea for renaming these articles. The idea is to rename the articles to Intel Core (microarchitecture), Nehalem (microarchitecture), and Sandy Bridge (microarchitecture). Another possibility would be to have "Intel" in all the names (making the naming more consistent -- i.e. Intel Nehalem (microarchitecture) and Intel Sandy Bridge (microarchitecture). If we do that I would suggest making the change to other uArch pages (i.e. NetBurst to Intel NetBurst (microarchitecture), Intel P6 to Intel P6 (microarchitecture), etc). What do you think? I'll give this a week for comments. If the response is favorable, I (or someone else) will/can rename them. If there is no response, I'll probably wait a bit longer. If anyone sees this, please respond. Thank you. -- Imperator3733 00:22, 16 October 2007 (UTC)

The pages have now been renamed. See Intel Core (microarchitecture), Nehalem (microarchitecture), and Sandy Bridge (microarchitecture) -- Imperator3733 18:06, 24 October 2007 (UTC)

[edit] Compatibility with LGA 775 sockets

Will the Nehalem CPUs be compatible with current LGA 775 sockets and current chipsets (P35, G33, G35, etc.) in use today? If not, will there be any special versions that can run on today's computer motherboards? Hellcat fighter (talk) 14:39, 24 November 2007 (UTC)

I do not believe so, since Nehalem uses an integrated memory controller and QuickPath. This requires a new socket, since there is a new bus and the system is set up differently -- if they used the same socket, a system could end up having two memory controllers. At least the server Nehalems will be using LGA1366 as a socket -- I don't know what the desktop chips will use. Because of the bus and memory controllers, the current chipsets won't work with these either. I think there were rumors of cut down/budget/legacy versions to run on LGA775, but I personally hope that that doesn't happen. -- Imperator3733 (talk) 01:11, 25 November 2007 (UTC)

If what you mentioned is true, current LGA775 board owners will have to replace their motherboards to upgrade to Nehalem processors. An Intel chipset less than two years old may not be able to support a Nehalem processor. This will be financially costly to the PC upgrader. Intel revises (or changes) their chipsets quite frequently, so upgrading the CPU of an Intel based PC can be expensive thing to do. AMD based PCs do not suffer from as high CPU upgrading costs (since socket AM2). Hellcat fighter (talk) 18:39, 26 November 2007 (UTC)

This isn't the first time that something like this has happened, and it certainly won't be the last. At least this time it is really necessary because of all the changes that are happening (IMC, QuickPath, etc.) -- Imperator3733 (talk) 18:45, 26 November 2007 (UTC)

[edit] Cache on Havendale and Auburndale

Does anyone know why the variants table lists Havendale and Auburndale as having 4MB L2 cache with no L3 while all the other chips have and L3 cache? The other Nehalems all have 256KB L2 per core, so changing to a shared L2 in the low-end chips seems to be too much work. Is there any place that specifically said that Havendale and Auburndale have L2 cache? -- Imperator3733 (talk) 00:44, 12 February 2008 (UTC)

I've updated the cache column, I think you'll find it makes sense now. It looks like (if it is the most recent information) the mainstream LGA1160 form of Nehalem has been pushed back to H2 2009. Is this reliable data? I thought they were going to be released about 3 months after the LGA1366 Nahalems. (Skilltim (talk) 08:30, 7 June 2008 (UTC))

Thanks for updating the table. The LGA1160 delay seems pretty reliable (X-bit Labs had an article), although I'm still quite mad at Intel about it (I had been thinking of doing an upgrade when those came out). The X-bit Labs article says "second half of 2009".

I also noticed that the tables are getting a bit cramped, at least when I'm viewing it. Do you have any ideas on how to make it easier to read? -- Imperator3733 (talk) 04:08, 9 June 2008 (UTC)