IBM Power6 press release numbers don't add for iTunes statement
Thursday May 24, 2007
Why does IBM's iTunes POWER6 comparison looks totally misleading? Did they add things that can't be added to create a big number?
Let's look at the IBM math, it looks wrong doesn't it? One can't just add peak rates to say it is the bandwidth. It is like saying I can run at 10 miles per hour, my car has gone 100 miles per hour, and the plane I take to LA can got at 650 miles per hour. So let's see that means the speed I went to LA is 650+100+10, so I traveled at 760 miles per hour. Nope it just doesn't work that way. And no, I didn't break the speed of sound on a commercial jet.
-
IBM Press Release said:
"Even more impressive, the processor bandwidth of the POWER6 chip -- 300 gigabytes per second -- could download the entire iTunes catalog in about 60 seconds -- 30 times faster than HP's Itanium."
OK so let's see how IBM did a "very funny" kind of math:
-
IBM MIXED maximum theoretical peaks and added them together, all of which are
(guaranteed) not to reached by real applications (and certainly not
iTunes
- L3 cache: 16B wr + 16B rd, for a total of 32B x 2.5GHz = 80GB/s
this is a max of 40GB/s in each direction - memory: 8B wr + 16B rd, for a total of 24B x 2.5GHz = 60GB/s
(75GB/s when even adding in address lines)
no more than 40GB/s memory read bandwidth
no more than 20GB/s copy bandwidth (limited by write) - IO: 4B wr + 4B rd, for a total 8B x 2.5GHz = 20GB/s (but for addr and data)
- chip to chip: 3x of 8B in + 8B out, clock rate?
- node to node: 2x of 8B in + 8B out, clock rate?
). Below "B" means bytes the number then multiplied
by frequency to get a peak bandwidth. "wr" is write, "rd" is read.
Lots of problems with quoting peaks:
IBM loves "peaks" but they don't mean anything - please show what is delivered (since STREAM is so easy to run, why hide results? are the true numbers so bad?)
Here is the problem with marketing peaks see the "Beware the Ides..." posting: http://blogs.sun.com/bmseer/entry/beware_the_ides_of_may
|
|
|
|
