Hi all, I'm trying to understand something about PCs and OSes and this seems the right place to solve misgivings I'm left with after browsing the web.
http://www.penguin.cz/~literakl/intel/i.html#IN
http://www.penguin.cz/~literakl/intel/o.html#OUT
The number of clocks it takes to do these instructions seems predefined/unchangeable/fixed.
But is the actual time different, depending on the device speed?
Or through the use of buffers and stuff the time for these operations never changes?
And what about the MOV instruction when reading/writing from/to RAM? To get the actual nanoseconds (or whatever) for that instruction should I add the fixed time for that instruction to the time needed to access ram, or is that time somehow included in the "clocks" value reported by these pages above because of cache and/or other tricks?
Sorry for the big confusion. ^ ^
Speed of IN, OUT, MOV; clock ticks, time
- Owen
- Member
- Posts: 1700
- Joined: Fri Jun 13, 2008 3:21 pm
- Location: Cambridge, United Kingdom
- Contact:
Re: Speed of IN, OUT, MOV; clock ticks, time
Those clocks only apply to the specific processors listed. Other processors of course have different cycle timings (and those timings are rarely documented)
In general, the only timing manufacturers will give you is the best case latency. Obviously they cant' tell you how long e.g. a DRAM access is going to take, or how long a device might assert the WAIT signal for.
In general, the only timing manufacturers will give you is the best case latency. Obviously they cant' tell you how long e.g. a DRAM access is going to take, or how long a device might assert the WAIT signal for.
Re: Speed of IN, OUT, MOV; clock ticks, time
Of course, that was not my doubt.Owen wrote:Other processors of course have different cycle timings
OK this is what I wanted to know, thanks.Owen wrote:Obviously they cant' tell you how long e.g. a DRAM access is going to take, or how long a device might assert the WAIT signal for.