the coming of itanium?
- Thread starter her34
- Start date
He glibly glosses over the fundamental issue with IA-64 and pure RISC designs in general. IA64 isn't a compiler writer's dream, it's more like a nightmare. It doesn't matter that most people won't be writing machine code, the compiler writer's still have to. It turns out that it is extremely nontrivial to write a compiler that can spit out highly optimized IA64 machine code. They have been working on the problem for over 10 years now and it still isn't there yet.
That quote is like Bhagdad Bob stating that the Americans are being soundly defeated as American tanks are seen in the background driving by. Ridiculous. Itanium definitely has its nitch but total takeover of X86 .... No.
Edit: Also, everything Intel wanted? X86-64 ... no. Dualcore ... no. Being pushed hard ... no. Being sued for monopolistic practices ... no.
Capt Caveman
Lifer
- Jan 30, 2005
- 34,543
- 651
- 126
with multi cores, will intel try to push for itanium?
one problem with itanium was x86 performance, but in future they could just add a conroe core for that
one problem with itanium was x86 performance, but in future they could just add a conroe core for that
To some people, the idea - of basing a processor's hardware architecture and ISA on that a compiler, in advance, should make all kinds of very hardware involved, very low level but complex, detailed decisions about exactly how code should be executed, - seemed like a good idea a long time ago. Like 20 years ago, ...or longer.
Tom Yager seems to have caught up with this idea only just recently, and is apparently thunderstruck with the sheer *brilliance* of the concept.
Good software utilization and performance mainly relies on good algorithms. Algorithms as in program design and source, not optimization of binary. Secondly, it relies on hardware performance - on the ability of already mature software to run on the next generation of hardware and make full use of it.
I'm an "old cow" so what do I know? - But this is why I think multithreaded and multicore is going to make mincemeat out of any stillborn concept trying to get a compiler to extract parallelism out of a source.
Besides that, I have some confidence in the time honored principle that: 'Tiers of computing are successively conquered by systems from the tier below.'
That is exactly why Microsoft are spending so much money to mess up Sony and their Playstation. They are defending the Windows PC's position as the household computing platform from a future threat.
-------
I'm not 100% sure the article isn't a joke, satire in the renowned style of "SubtleIntelfreak".
Just look at some quotes:
"The time is right for IA-64 because, except for the AMD hitch, the future played out precisely as Intel planned."
- Really? Did it now? In terms of Itanics's performance? In terms of timeline? In terms of Microsoft and HP dropping out?
Or maybe he means the business deals to dismantle the competition, HP-PA and Alpha?
"The time is right for IA-64"??? Taste that phrase again.
What is it I don't know? :Q Have I been in a coma for years? Who is the US president now? Arnold?
- No guys, he can't be serious. It's gotta be some kind of joke.
"The compiler is the processor for the next 20 years."
Again, I can't help sensing some irony (or ridicule?) in that exclamation.
-Uhu? Does Sun and SPARC ring a bell? Excuse me but I have a far easier time seeing how things like Java, C# and dot.net fits into current and future computing, than endlessly recompiled, very hardware dependant software, with lots of expensive and persistent bugs.
"I'm bullish on IA-64 because a dream world of compilers that take their sweet time to build and optimize but that produce mind-blowing code will surface there first."
Like - Wow?
I don't think "mind-blowing" binary code really exists. Does he seriously believe there are huge gains to be had beyond some basic, robust, optimization?Doesn't he realize that they are forced to these vulgar compiler excesses on the Itanic just to make it run at all?
"dream world of compilers" - He must be joking! Right?
But let's step back and consider the current challenges for CPUs, To see how the Itanic fits in. ...In the spirit of SubtleIntelfreak of course:
1: Making use of an increasing, very large amount of available transistors. - Yes! Itanic truly excels here. It uses many times more transistors than a similar performing Opteron.
2: Moving the work into and out of the processing. - Yes! Itanium solves this problem with gigantic caches. This solution is not available to the competition, since those CPUs are not expensive enough to have such large caches.
3: Execution speed. No, Itanic is a dog, even though this was exactly the target for the entire EPIC (IA-64) concept, but this is not really a problem any longer since the bottlenecks are elsewhere. Also 2 out of 3 makes a winner, doesn't it? And who knows, with "mindblowing code"...
I think "mind" and "blowing" might indeed be the operative words here.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter