Redbook on the new z196 mainframes.
|
Prev: Effects of Memory Latency and Bandwidth on Supercomputer,Application Performance
Next: Changing the color of objects/primitives only ? (flat shading...) (massive parallel lookup hardware idea...)
From: Robert Myers on 3 Aug 2010 17:49 On Aug 3, 11:52 am, Bernd Paysan <bernd.pay...(a)gmx.de> wrote: > Andy Glew wrote: > > 256 *BYTE*? > > > 2048 bits? > > > Line sizes 4X the typical 64B line size of x86? > > > These aren't cache lines. They are disk blocks. > > What do you expect? Bandwidth and latency change differently. My rule of > thumb is that transfer time = access time, and that should work for disks as > well as for memories. That just repeats the old (and wrong) saw that you can always add bandwidth but latency is forever, but I'm getting tired of arguing about it. For some applications, latency is irrelevant and wasted bandwidth is intolerable. You are entitled to your prejudices and even your misconceptions, but I'm tired of seeing the same claims over and over without so much as an asterisk. Robert. The fact that disk sectors are still 512 bytes is > just legacy stuff, they should have increased to about a megabyte by now. > > -- > Bernd Paysan > "If you want it done right, you have to do it yourself!"http://www.jwdt.com/~paysan/
From: Noob on 4 Aug 2010 05:36
Bernd Paysan wrote: > The fact that disk sectors are still 512 bytes is just legacy stuff, > they should have increased to about a megabyte by now. Haven't most HDD manufacturers completed the transition to 4K sectors? http://en.wikipedia.org/wiki/Advanced_Format |