Finding Memory System Latency: Interleaving Effects

AI Thread Summary
To determine memory system latency, the architecture significantly influences the results, with serially architected memory exhibiting longer latency but higher bandwidth compared to random access memory. The discussion revolves around calculating latency given specific parameters like bandwidth, precharge time, and data transfer size, particularly in scenarios with and without memory interleaving. A general latency equation used for hard drives is not applicable to system memory, prompting the need for a more relevant formula. The user seeks clarification on how to apply these concepts to their homework question involving specific bandwidth and interleaving conditions. The conversation highlights the complexity of memory latency calculations and the importance of understanding the underlying architecture.
Knowledge92
Messages
3
Reaction score
0
I've been searching for some type of equation or explanation of how to find the latency (access time) of a memory system when given the bandwidth, the precharge (refreshing) time between accesses, and how many bytes of data are being transferred...

Also I am not sure as to how memory interleaving would change this latency...

Any help is greatly appreciated!
 
Technology news on Phys.org


It depends on the architecture of the memory, so I don't think there is a general relation. A serially architected memory can have a very long latency, but then a very high bandwidth once you've arrived at the correct location, while a random access memory will have a much smaller latency.
 


Thanks for that information, but maybe it'd help if I just posted my homework question:

Assume a memory system has a bandwidth of 120 MBytes (M = 2^20) per second and has aprecharge (refreshing) of 5 ns between accesses. What is the memory systems latency (access time) for the following two cases?

a. Two bytes of data are transferred and memory interleaving is not used.

b. 4-way memory interleaving is used and each module is able to transfer two bytes per access.

I am unsure how to start this. I found a computer engineering equation that latency = seek time + rotational delay + transfer time + controller delay, but I have no idea how to apply this general form of a latency equation (or if I am supposed to use this equation at all) to this type of problem.

Thank you very much!
 


Knowledge92 said:
latency = seek time + rotational delay + transfer time + controller delay
This is a formula used for hard drives, not system memory.
 


rcgldr said:
This is a formula used for hard drives, not system memory.

OK.. is there a formula for the latency of system memory/what is it?
 
Thread 'Star maps using Blender'
Blender just recently dropped a new version, 4.5(with 5.0 on the horizon), and within it was a new feature for which I immediately thought of a use for. The new feature was a .csv importer for Geometry nodes. Geometry nodes are a method of modelling that uses a node tree to create 3D models which offers more flexibility than straight modeling does. The .csv importer node allows you to bring in a .csv file and use the data in it to control aspects of your model. So for example, if you...
I tried a web search "the loss of programming ", and found an article saying that all aspects of writing, developing, and testing software programs will one day all be handled through artificial intelligence. One must wonder then, who is responsible. WHO is responsible for any problems, bugs, deficiencies, or whatever malfunctions which the programs make their users endure? Things may work wrong however the "wrong" happens. AI needs to fix the problems for the users. Any way to...
Back
Top