Geek Answers Handbook

Assessing the effect of processor frequency on I / O performance

I am researching dedicated I / O software that will run on consumer equipment. In essence, it comes down to saving huge data flows for further processing. Now I am looking for a model for evaluating x86 performance factors .

Take, for example, the new Macbook Pro :

Thunderbolt high-speed I / O technology (I / O) delivers amazing 10 gigabit per second transfer speeds

1.25 GB/ssounds good, but most of the day's processors are synchronized around 2 Ghz. Several cores do not matter much if only one can be assigned to each network channel.

Thus, even if the software acts as a miniature operating system and is limited to network / disk operations, the amount of data entering the repository cannot exceed [1] per second. Although this indicates an approximate performance limit, I feel that this is far from adequate. P / (2 * N)

What other considerations should I consider when evaluating I / O performance in terms of processor frequency and other hardware specifications? For simplicity, suppose that the repository runs instantly under any circumstances.

[1] P - ; N -

+5
2

, , /, PCIe, FSB, northbridge .

, , , , . , , , , , . , , parallelism .., - , .

, , , , , , .

+2

, - , - .. , (, ).

, , , , , () .

, , , - - . : , Apple .;)

Thunderbolt , , , , Thunderbolt, .

Thunderbolt - , - , Apple. , , , , .

- , . , - / , , . - / , ( , , , Finder) .

, CPU- , , , , , Thunderbolt .
+1

More articles:


All Articles