Is there any way to measure the number of cycles required for a
multiplication instruction without using the timestamp counter?
|
|
0
|
|
|
|
Reply
|
Vasanth
|
11/2/2004 1:38:24 AM |
|
See post (few lines lower): rdtsc and Pentium M- one more question (was
rdtsc ... ("john doe")
Zaharije Pasalic
|
|
0
|
|
|
|
Reply
|
hari4063
|
11/2/2004 2:37:19 AM
|
|
Vasanth Venkatachalam wrote:
> Is there any way to measure the number of cycles required for a
> multiplication instruction without using the timestamp counter?
Sure. Write a program that performs, say, 1e9 multiplications, run
it on an empty system, and time it with a plain old stopwatch.
--
Regards, Grumble
|
|
0
|
|
|
|
Reply
|
Grumble
|
11/2/2004 8:41:49 AM
|
|