Slashdot Mirror
Oracle Ordered To Pay $3B Damages To HP (bbc.com)
Oracle has been ordered to pay HP $3 billion in damages by a California jury over HP's claim that Oracle reneged on a deal to support HP computer servers running on Itanium chips from Intel. Oracle said it will appeal. BBC reports:The court battle over the contract was settled in 2012 but the damages HPE was due have only now been agreed. HP was split into two in 2015 with HPE taking over the running of its servers and services business. In court, HPE argued that although the 2012 legal judgement meant Oracle had resumed making software for the powerful chips, its business had suffered harm. It argued that Oracle took the decision in 2011 to stop supporting Itanium in a bid to get customers to move to hardware made by Sun -- a hardware firm owned by Oracle. Oracle said that its decision in 2011 was driven by a realisation that Itanium was coming to the end of its life. It also argued that the contract it signed never obliged it to keep producing software in perpetuity. Intel stopped making Itanium chips in late 2012 and many companies that used servers built around them have now moved to more powerful processors.
Exactly my thought. In Oracle's defense, it was clear in 2001 that Itanium was dying. You could hear the Monty Python dialog, "I'm not dead yet!" but despite the protestations the writing was on the wall. AMD proved you didn't need a completely incompatible system in order to move forward to a 64-bit architecture.
You almost wonder about the choice of name as Itanium is close to Titanic, which is about the size of its failure.
The only thing AMD proved is that you could glue 64-bit operands onto a 32-bit processor and that people cared more about running their old software slightly faster than running future software much faster. Thanks to AMD, we're still stuck with one of the first and worst CPU architectures imaginable. The only reason x86/x64 can outperform competitors is due to huge research budgets pushed into eaking out increasingly smaller improvements. It required an entirely new market to give ARM the inroads it needed to start throwing around large research budgets too. I suspect at some point ARM will move onto the desktop (Microsoft-willing) and then Intel can finally start retiring their x86-limited CPU.
Slashdot social media options: AIM, ICQ, Yahoo, Jabber and Mobile Text. Why no MySpace?
Come on now, this is not the days of the 80486 anymore. There is no x86 architecture, only an x86 instruction set which although has some warts really isn't bad, its main problem, insufficient registers is addressed completely by x64 variant (thanks AMD). All modern x86/x64 CPUs use a very RISC like architecture, or micro architecture if you prefer, and basically translate the CISC x86 instructions to multiple micro instructions on the backend, the decoder being simpler and faster than most of the rest of chip means there is basically no penalty here, it just isn't bottleneck not in terms of die space, not in terms of clock speed capability.
Rest assured when you a single ARM core approaching anything near the performance of its top drawer x86/x64 counter parts it will be because the that chip has become more like today top line x86/x64 parts in just about every way except perhaps the exposed ISA.
TL;DR - x64 has nothing to do with that first and worst architecture, other than some legacy but more or less virtual instructions.
Repeal the 17th Amendment TODAY! Also Please Read http://www.gnu.org/philosophy/right-to-read.html
The same is true if Apple, Atari, and Commodore had use the 6809 but the 6502 was also cheaper.
I know this is off-topic: I programmed assembler for the 6809, and that thing was a workhorse for its era. It's instruction set was well thought out, and its indexing modes were awesome.
Then I entered the Intel world. Blech!