David Mosberger, developer of the initial GCC port to IA-64 and lead kernel architect for Linux on IA-64, tells us why you should care about Intel’s new 64-bit chip.David joined an Internet research group within Hewlett Packard in 1997, and a few years later began to contribute to the Itanium port, where he is now lead kernel architect. Since 1988, he has written the first IA-64 back end for the GNU Compiler Collection (GCC), wrote much of the IA-64 toolchain support, and implemented much of the IA-64-specific parts of the Linux kernel.
The Future of Linux on IA-64
Submitted by Koki 2002-10-04 Linux 13 Comments
I didn’t get much from that interview…did I miss something? Sounded more like chatting about things than really giving any strong reason why one should care.
Itanium 2 is actually a very formidable CPU, and I don’t think people quite appreciate it — partly because there aren’t enough machines around yet. But I think as people start to realize just how powerful a CPU it is, you will see attitudes change a bit.
I’ve yet to see a single attack on Itanium 2 based on bad performance. Universally its recognized as a very powerful processor at least every report I’ve seen. Does anyone else’s experience differ?
Anyway the part about a desktop version for 2003 sounded great.
to get the sale price on one of these:
Okay, so I don’t really have $4000 cash to spend on hardware. Any guesses as to what prices will be like a year from now?
If you compare prices to P4s with the same amount of cache Itanium 2’s run about the same. Right now they are only available with a ton of cache so they are expensive. Also the motherboards and everything else is pricey. There isn’t any good reason though that Intel couldn’t cut the cache and have these in $2k computers next year.
If this this were real pricing, I would still skip on this offer, note the zx6000 is 2x the price of zx2000, yet are basically same except 1 more cpu chip. In other words HP must be losing $ big time on the zx2000 or the extra Itanium chip is $4k.
Yes it all comes down to cache size v prices, if Itaniums are cut down in on chip cache size they will surely lose what ever scores they shine on. If any cpu running at 1G-3G had that kind of cache it would rock, even a model T pentium.
I only wish I had the option of having Level3 cache on mobo again to boost cpu performance, I could upgrade that instead of cpu every yr or so, wouldn’t be as fast as on chip cache but way cheaper & way faster than DRAM.
Will never get in place with suchj prices my guess is that the future of the desktop on x86 is stuck to IA-32 for price reasons. Many manufacturer have problem generating revenue on their IA-32 desktop lines and make money on their server line (read branded machine, not those home build twaiwanese machines). So servers get IA-64 because anyway the client does not really look the price and the end users get stuck in iA32 land where price continue to drop and inovation take years to happen.
I went and looked at the HP’s that were mentioned above (why not go and price out an Itanium developer solution). I was shocked that the machines are being offered with DDR RAM instead of RDRAM. Would this not be the final nail in the coffin of RDRAM? Let’s hope so!
The final nail is already in the coffin. We’re just waiting for the body to be interred.
was IA-64 already being actively developed at that time? Or was that a typo?
(Read More section of the osnews article)
I suppose that’s just for server.
As a long time assembly language programmer I have an interest in any potentially successful chips, especially anything that’s better than the technically limited X86 architecture. A while back I started studying the Itanium in depth and have found that it’s an awesome technicial solution. However, the markets will decide it’s success in the end as is the case with most, if not all, new gagets. Intel’s biggest competition is it’s own invention, the X86 line.
As Intel takes the Itanium design to smaller chip processes and improves on the implementation the results are likely to be impressive. Why? The instruction set is designed so that a larger percentage of the chip circuits are devoted to performing real work, rather than figuring out the best way to execute the instructions as is the case for the Pentium design. This is a critical factor in the chips “promise” of more “headroom”. With more circuits devoted to actual calculation, and less to “figuring out instructions and an optimal sequence” of execution it’s easy to see how the Itanium could win out in the long run. Yes, the load is shifted to the compiler, but maybe that’s where it should be since there are many more “eyes” outside of a chip company that can devote time to optimizing code generated by compilers. By shifting the responsibility for optimzation to the compiler Intel has invited these eyes to work on innovative solutions.
Intel’s Itanium 2 pricing is a huge barrier to mass market acceptance. This could give AMD the time it needs to get their foot in the 64 bit door.
It’s certainly an interesting time with both Intel and AMD assulting the Pentium Success Mountain.
In the long run, if Intel can hack it that long and if they lower their prices, Itanium is likely to be the winner over the AMD X86-64 bit architecture. Why? Five, ten or twenty years from now who wants to be using a 64 bit architecture hobbled by it’s 32 bit past? Especially when there are cpus running at 10Ghz or faster, with truely massive stores of RAM and disk.
The reality is that right now the Pentium IA-32 architecture is the king of the market place. It’s hard to see how Intel can make the transition to Itanium being king smoothly expecially with AMD nipping at it’s heals.
For those who’ve wondered, here’s the current list of intel CPU prices : http://www.intel.com/intel/finance/pricelist/ . Notice that high-end Itaniums cost about $4200 , a bit more than high-end Xeons (which are $3700).
I’ve read that HP receives significant discounts on the Itanium chips since they are a co-designer of it with Intel.