Stories
Slash Boxes
Comments

SoylentNews is people

posted by martyb on Saturday August 08 2015, @11:38PM   Printer-friendly
from the nice-summer-beach-reading dept.

Raymond Chen recently posted a ten-part introduction to the ia64 architecture. Rapidly teaching me that while I might be able to write a brainfuck to perl compiler in a few minutes, there's no way in a million years that I'll ever be able to write a good compiler that targets ia64.

The Itanium is a 64-bit EPIC architecture. EPIC stands for Explicitly Parallel Instruction Computing, a design in which work is offloaded from the processor to the compiler. For example, the compiler decides which operations can be safely performed in parallel and which memory fetches can be productively speculated. This relieves the processor from having to make these decisions on the fly, thereby allowing it to focus on the real work of processing.


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 3, Interesting) by Grishnakh on Sunday August 09 2015, @01:42AM

    by Grishnakh (2831) on Sunday August 09 2015, @01:42AM (#220091)

    x86-32 bit the living crap out of Itanic when running x86-32 code, that's definitely true. Itanic had an x86-32 unit to make it backwards-compatible, but it wasn't very fast.

    Starting Score:    1  point
    Moderation   +1  
       Interesting=1, Total=1
    Extra 'Interesting' Modifier   0  
    Karma-Bonus Modifier   +1  

    Total Score:   3  
  • (Score: 0) by Anonymous Coward on Sunday August 09 2015, @01:58AM

    by Anonymous Coward on Sunday August 09 2015, @01:58AM (#220098)

    Indeed, it was a token compatibility unit and IIRC they deleted it from the 2nd or 3rd generation chips.

  • (Score: 3, Interesting) by Subsentient on Sunday August 09 2015, @02:12AM

    by Subsentient (1111) on Sunday August 09 2015, @02:12AM (#220103) Homepage Journal

    From Wikipedia:

    When first released in 2001, Itanium's performance, compared to better-established RISC and CISC processors, was disappointing. Emulation to run existing x86 applications and operating systems was particularly poor, with one benchmark in 2001, reporting that it was equivalent at best to a 100 MHz Pentium in this mode (1.1 GHz Pentiums were on the market at that time).

    So it appears that Itanium performance was pretty bad, and 32-bit x86 emulation was hilariously bad.

    --
    "It is no measure of health to be well adjusted to a profoundly sick society." -Jiddu Krishnamurti
    • (Score: 2, Insightful) by Anonymous Coward on Sunday August 09 2015, @02:29AM

      by Anonymous Coward on Sunday August 09 2015, @02:29AM (#220112)

      Don't make too much of that. The 1st gen chip was late to market, if it had come out on schedule it would have been competitive at the time. The 2nd gen chip was out just over a year later and it was a benchmark leader. [theregister.co.uk] Being that fast to market means development was already well under way before the 1st gen chip started shipping.

      • (Score: 0) by Anonymous Coward on Sunday August 09 2015, @11:35AM

        by Anonymous Coward on Sunday August 09 2015, @11:35AM (#220206)

        On the SPECint2000 integer performance benchmark, HP said that the 1GHz McKinley, presumably the one with 3MB of L3 cache, had 1.1 times the integer performance of a 1.67GHz AMD Athlon XP 2100+ processor, 1.2 times the integer performance of Intel's own 1.6GHz Pentium 4 Xeon MP processor, and 1.3 times the integer performance of the 1.05GHz UltraSparc-III processor from Sun Microsystems.

        OK so in July 2002 the Itanium 2 is 1.1 times faster than a 1.67GHz AMD in SPECInt.

        Too bad that in _June_ 2002 the AMD Thoroughbred was released: https://en.wikipedia.org/wiki/Athlon_XP#Thoroughbred_.28T-Bred.29 [wikipedia.org]

        The fourth-generation Athlon Thoroughbred was released on June 10, 2002 at 1.8 GHz (Athlon XP PR2200+).

        Which I daresay was more than 1.1 times faster than the "1.67GHz AMD Athlon XP 2100+ processor" and thus faster than the Itanium 2 before it even came out.

        Yes the Itanium 2 was better at SPECfpu however if you look closer at those FPU tests it tends to do very well at the easily parallelizable stuff and not so good on the other tests. With such tasks you can use clusters of AMD64/x86 machines instead which can work out to still be cheaper. Your own link says:

        HP warns that this configuration includes what it calls a "large cash discount" available through HP Direct, but does not specify what this discount is. These are great performance numbers, but these are slippery pricing disclosure practices that will undermine the credibility of the TPC-C tests if they are allowed to continue.

        So much for better price vs performance. Those who couldn't get the "large cash discount" for whatever reasons, found it made a lot more sense to buy AMD64 machines.

        Especially when they could still run their old apps without recompiling, rewriting etc. And so that's what they did.

        • (Score: 0) by Anonymous Coward on Sunday August 09 2015, @05:33PM

          by Anonymous Coward on Sunday August 09 2015, @05:33PM (#220324)

          > Which I daresay was more than 1.1 times faster than the "1.67GHz AMD Athlon XP 2100+ processor" and thus faster than the Itanium 2 before it even came out.

          I really don't care enough about the topic to get involved in a benchmark peen measurement contest for something over a decade old. I'm just going to say that you've substantially moved the goalposts from "performance was pretty bad" and getting "the living crap beat out of it on benchmarks" to performance was neck-and-neck with the fastest competitors.

      • (Score: 0) by Anonymous Coward on Sunday August 09 2015, @02:31PM

        by Anonymous Coward on Sunday August 09 2015, @02:31PM (#220253)

        Benchmark leader only for specfpu not leader for specint.
        https://www.spec.org/cpu2000/results/res2002q3/ [spec.org]

        Intel x86 and AMD were battling for leadership in "real world performance" leaving Itanium 2 behind before it even launched.