Talk:POWER7

Latest comment: 8 years ago by Pkunk in topic C1 core?

What about a comparison to Intel Core/Xeon Processors. —Preceding unsigned comment added by 212.114.159.142 (talk) 12:52, 27 January 2011 (UTC)Reply

Done --89.120.104.138 (talk) 15:37, 21 August 2014 (UTC)Reply

power? edit

how many watts are this processors? I can't find it anywhere... thanks 213.221.17.130 (talk) 14:20, 20 June 2013 (UTC)Reply

About 200W per chip (@ 3.9 GHz) according to this german article at Heise.de. -- Henriok (talk) 21:17, 20 June 2013 (UTC)Reply

you don't say... edit

There's a maximum of four chips per quad-chip module? How presciently named! —Preceding unsigned comment added by 69.41.137.46 (talk) 23:40, 7 July 2010 (UTC)Reply

missing this edit

Power7 have quad-chip multi-chip module at Max..

that's mean 8 Memory Controller (Each chip had 2) and 32 core for a single Socket. and Max 1034.4 GFLOPS per module Salem F (talk) 21:39, 27 August 2009 (UTC)Reply

References IBM's POWER7 and Sun's Rainbow Falls, Unveiled PC Magazine

SC09 edit

Power7 was presented at SC09, so there are going to be conference proceedings. The Wikipedia article should include any such citations. —Preceding unsigned comment added by 76.232.38.36 (talk) 15:42, 2 December 2009 (UTC)Reply

I know. Still gathering documents and plannign an update. -- Henriok (talk) 16:10, 2 December 2009 (UTC)Reply

PS4 edit

Is there any speculations (or something more reliabile) known about possible usage of Power7 in PlayStation 4 by Sony? I bet they would use 4 or 6 core version (which are defective 8 core units), similary like with Cell. —Preceding unsigned comment added by 83.18.219.252 (talk) 00:39, 30 December 2009 (UTC)Reply

There are speculations, but Wikipedia is not a place for that. Personally I think the PlayStation fanboys are jumping on any impressive PowerPC coming from IBM. There are no signs from either Sony not IBM that POWER7 would be anywhere near a game console, and frankly it's not suitable at all and POWER7 bears very little similarity to cell, besides them both being 64-bit PowerPCs. -- Henriok (talk) 09:47, 30 December 2009 (UTC)Reply
The Wii U is believed to be based on POWER7. IBM has said the Wii U CPU is POWER but never specified which generation though they've alluded to Watson, 45 nm process, manufactured at the East Fishkill, NY plant and lots of embedded DRAM...which are all aspects of POWER7. But it's not a 100% confirmation. However, it's much closer to conformation than anything I've ever heard regarding the PS4.24.129.98.70 (talk) 19:59, 10 February 2012 (UTC)Reply
These aspects also applies to PowerEN and Blue Gene/Q, both using PowerPC A2 cores which are much more suitable for a game console. Time will tell. -- 21:16, 10 February 2012 (UTC)
I may be mistaken but I don't believe the A2 was used in Watson. Nil Einne (talk) 13:04, 26 April 2012 (UTC)Reply
You are not mistaken. However. Neither IBM nor Nintendo did say that Wii U is based on POWER7. They said its based on the same technology as Watson. Technology is a pretty broad term, and Watson is based on Power Architecture technology. That applies to POWER7 and A2, and a myriad of other technologies. POWER7 is exceedingly unsuited for a cheap gaming console. A2 is not. -- Henriok (talk) 17:34, 26 April 2012 (UTC)Reply

Why External Link to Intel Xeon? edit

Any reason why there is an external link to: SGI Altix UV 1000 SPECint_rate_base2006 result ? This is not even POWER-based, let alone POWER7. I'll remove this in 2 weeks unless notified. Gareth.randall (talk) 17:49, 11 December 2010 (UTC)Reply

Done. Gareth.randall (talk) 18:52, 24 December 2010 (UTC)Reply

Endian? edit

Big endian or little endian? — Preceding unsigned comment added by 71.222.114.1 (talk) 21:04, 26 October 2012 (UTC)Reply

Bigendian noq (talk) 10:21, 27 October 2012 (UTC)Reply

Missing some trivial technical details edit

This article, as well as a few other POWER architecture ones, appear to be missing some very basic information: big endian byte order, 64-bit, number of general purpose integer and floating point registers and their size, address space size. Possibly that the infobox for all processor architectures need to hold this as well? Is there a place to better address this issue for a consensus and possibly the development of a better infobox template? Thanks, 76.10.128.192 (talk) 02:43, 3 June 2013 (UTC)Reply

This isn't a page for an instruction set architecture, it's a page for a processor, so its infobox should give microarchitecture etc. details, not instruction set architecture details. One field in the processor infobox is "Instruction Set", which should link to a page giving instruction set architecture details; however, the Power Architecture page doesn't give that. Part of the problem is that the "Power Architecture" includes more than the instruction set architecture; the instruction set architecture is essentially "the instruction set architecture formerly known as PowerPC". Perhaps there should be a "Power Instruction Set Architecture" page (with a hatnote pointing people to the IBM POWER Instruction Set Architecture page, that ISA being the original basis of the PowerPC ISA), which gives the PowerPC/Power ISA details, with "PowerPC" perhaps redirecting back there - or perhaps with "PowerPC Instruction Set Architecture" redirecting there, and with "PowerPC" mentioning and linking to the PowerPC/Power ISA as well a mentioning all the various flavors of PowerPC processors.
In any case, the right place to discuss this is probably the talk page for Power Architecture. Guy Harris (talk) 06:08, 3 June 2013 (UTC)Reply

C1 core? edit

What means that "C1 core"? Is it a stepping marking or some other abbreviation? --Pkunk (talk) 22:13, 15 May 2015 (UTC)Reply