What are the implications of sacrificing tighter latency timings for bus speed and could anyone give me an example when tighter memory timings would be preferable.
Iam asking the question because my Mobo allows certain tweaks to increase memory speed but it seems that it's a choice between FSB and tighter timings..
If anyone could expand on this further for me i would greatly appreciate it....
If you think of memory as a wherehouse, bus speeds represents the ammount of stuff a wherehouse employee can carry at a time. Whereas latency represents how fast he can find the stuff he needs.
In a situation where a lot of stuff is needed from the same place (a video file or picture for example), then a worker that can carry more (RAM with higher bandwidth) is better. But when lots of small things are needed from different places in the wherehouse then a worker that can find them quicker (lower latency RAM) is best.
This is why P4's (with thier high FSB's and dual channel support) are much better for video editing because large ammounts of data can be moved quickly by them.
Note: Sorry if this makes no sence I had 2 hours sleep this morning.
No need to apologise,Thats the kind of explanation that i find useful and understand.
I know there are probably other factors such as CPU cache and GPU onboard memory to take into account but for applications such as a graphically intensive game which one would yield the most benefit?
I have ran some tests using Aquamark 3 and tried to set it up in such a way as to represent both scenarios and my scores are more boosted by the tighter timings rather than FSB though this is probably only relevant in this particular example.
I suppose at the end of the day there are no rights or wrongs and both have their uses and your explanation of the P4 during video applications now makes sense to me.
Thanks a lot for your time aiding an interested learner.It's much appreciated....
Im not a programmer so I would just be guessing when I say which is better in games. But gun to my head, I would say that both are important for games. Large textures would benefit from high bandwidth, while physics engines would prob benefit from tighter timmings (lots of small modules of code) and large caches.
I admit both would be nice but it seems to come down to a choice between the two,on my Mobo anyway.
When you mention large textures benefiting from the higher bandwidth would this be handled by the graphics card primarily with the system memory taking up the slack when the graphics card ram is full?
I have often wondered whether it is the graphically intensive games that benefit from the tighter timings as oppose to a strategy game such as Medieval Total War for example which is less graphic intensive but still very processor intensive benefiting from the larger bandwidth?
I know there is probably no hard and fast rules when it comes to this kind of thing so my queries are probably unanswerable without inner knowlege of each game.
Thanks again for your input....