notes on new power mac (was Re: clusters v SMP)
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
Mark Hahn hahn at physics.mcmaster.caMon Jan 28 15:38:03 PST 2002
- Previous message: clusters v SMP
- Next message: sram, not sdram.
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
> The 4GB memory restriction of an Intel node would be too > restrictive for some of our users. The majority I suspect are reading this triggered me to look at the specs of the new Mac. after all, the PPC is a reasonably nice architecture, and afaik isn't maimed by 32b-isms. anyway, ignoring Apple's inane marketing-twit-speak, the specs are: dual PPC G4/1000 256K onchip L2 2M per-CPU DDR sdram L3 cache 133 MHz FSB, 3 slots of PC133 (non-ddr) sdram 4 64x33 PCI slots, builtin 1000bT 360W max power. personally, 2M seems strangely small for DDR, since it's not either fast or wide. probably a single chip. they'd obviously be much better off (at least from a beowulf perspective) by biting the bullet and finally engineering a decent dram bus. 3 dimms also means it has no ram-size advantage over commodity (ia32) parts. builting gigE is interesting. in short, a nice workstation, but not well-suited to cluster nodes.
- Previous message: clusters v SMP
- Next message: sram, not sdram.
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
