Author Topic: Pentium III cpus use 1/2 to 1/3 the power of other processors (Jan 2002)  (Read 2117 times)

0 Members and 1 Guest are viewing this topic.

Offline chrisNova777

  • Administrator
  • Posts: 8334
  • Gender: Male
  • "Vintage MIDI Sequencing + Audio Production"
    • www.oldschooldaw.com | vintage audio production software + hardware info

http://ark.intel.com/products/27525/Intel-Pentium-III-Processor---S-1_40-GHz-512K-Cache-133-MHz-FSB
http://www.geek.com/chips/intel-releases-new-14ghz-pentium-iii-548434/

Quote
Yesterday Intel announced a new Pentium III release. It's a “Pentium III server processor” at 1.4GHz with 512 KB L2 cache, copper interconnects, 0.13 micron architecture, and a host of supporting hardware. The chip price is US$315 in 1K quantities. The press release says in part: “Intel Corporation introduced a 1.40 GHz version of its Intel� Pentium� III server processor with 512 kilobytes of L2 cache.” The press release does mention the chip is for front-end and general-purpose servers, so I'm thinking “revised Tualatin.”

Quote
If anything, they've shown that it was a pretty good design… because the 0.18 version scaled to 2000 MHz (which is far beyond the limit of the 0.18 micron P3, even 25% derated), and because it is scaling well at 0.13 micron. Simply put, they invested more silicon to “do it their way”, which brought additional pipeline stages, additional MHz, and additional performance.

We can argue until the goats come home as to how “perfect” or “garbage” this design is/was, but in truth, are any of us here remotely qualified to balance such ephermeral design methodologies? I think not. I certainly am not!

So, for all you P4 bigots, consider: the 0.18 P4 beat the squit out of the 0.18 P3, and was on par (!!!) with the 0.13 P3. Is there any reason to believe that the 0.13 P4 won't beat the squit out of the 0.13 P3? … NO! …

Finally, even if a chip is 25% poorer in per-clock performance over its competitors (and brother chips), if it can clock to 35% or 50% or 75% higher clock levels because of the design, doesn't this qualify as good engineering? I THINK SO.

If the AMD:INTEL war has shown nothing else, it is that neither company is getting radically better performance out of either design technology in the 0.18 micron process. I see no way to break that identity in the 0.13 and 0.10 arena until one of the two designs a radically new way TO compute.

Grrr.rrr.rrrrrr

TDP = thermal design power
rated at 32 watts!!!!!!!
but still provides decent performance! equal to that of a g4 MDD cpu
with way less heat produced

« Last Edit: November 07, 2014, 06:21:58 PM by chrisNova777 »