I have a G5 quad 2.5 and as I always thought and said... it is so... slow specially when I was playing the other week with one of the new 15" Mac Book Pro. It was faster than my G5.
I hope intel realease chips twice a year only, they are coming with a new chip every other week and is frustrating as a costumer to feel that my so expensive Mac Book is already discontinued.
We don't need to know about the costumes you wear.
Intel releasing new chips all the time is nothing new, they have done this for ages. Would you rather have IBM or Freescale providing the chips? Then you will be lucky is you see an update every year!
I think this new processor from Intel was designed to compete with this from Sun::
http://www.sun.com/processors/UltraSPARC-T1/ But even with Intel's new chip Sun is still winning the the low power race
Sorry, but the Niagara series of processors from Sun are aimed at a different market. These are designed for high threads count situations and not per thread performance. An example would be web servers. Intel has nothing that can compete with the Niagara. It has encryption built-in so SSL can be done with virtually no penalty. No need for an offload card, this is done on the CPU. FP performance though is horrible, as the Niagara comes in 4, 6 and 8 cores, but only has one FP unit for all of the cores. Niagara2 will fix that issue and will have one FP unit for each core, more encryption schemes, more threads per core (from 4 to 8), higher clock speed and built-in 10-Gig-E x 2 and two-way support. It will still have four built-in memory controllers as does the current chip. Expect to see it in the next few months.
These processors are quite slow such that the performance per watt isn't really there for most things. If you benchmark it for databases and other apps that can exercise all 32 threads/8 cores you will find that a conventional 2x2 Intel or AMD rig will run rings around it such that it is not obvious why you would want to use it. Even when running 100 threads the Intel and AMD cores are capable of getting a lot more work done faster with no hardware thread support.
It's slow in some aspects and in other aspects it makes the Intel and AMD offerings look slow.
http://blogs.zdnet.com/Murphy/?p=524
“So if we blandly assume that the Xeon results would scale linearly, this suggests that it would take somewhat more than eight 3.2Ghz Intel Xeons to match one UltraSPARC T1 at 1.2Ghz. Similarly, it would take roughly four IBM Power5+ dual core machines.
So why is this interesting? For two reasons: first because that 2:1 ratio for Xeon to PPC crops up a lot in other benchmark results, and secondly because this illustrates the utter dominance of the "slow" CMT approach over higher megahertz on multi-threaded tasks.”
This is not the only benchmark (real world or otherwise) that shows this.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2657&p=4
Yep, these processors are quite slow. If you look at open databases, then this chip shows poor performance as the database is not tuned for the chip. Use another database and this chip does much better, and if the database is licensed per socket, this is a great chip.
http://www.opensparc.net/blogs/2006-03/colm-maccarthaigh-niagara-vs-ftp.heanet.ie-showdown.html
They're ruled out because there is no such thing as a single socket Xeon solution anymore.
Really, look here:
http://h10010.www1.hp.com/wwpc/us/en/ss/WF06a/15351-15351-3328412-241644-241475-3201178.html
Just because the chip support it, doesn't mean it has to be used.
As for the power the current Xeon's in the Mac Pro, here you go:
http://www.intel.com/cd/channel/res...ucts/server/processors/5100/feature/index.htm
