MT/s vs. MHz: Accuracy in Measuring Memory Speed

[Peter_Brosdahl]

So doing a story on their latest SSD and saw this on their website and thought it'd be something interesting folks might want to check out. I think most of us have noticed how much of the industry has been switching from MHz to MT/s and this looks like a decent article on it.

MT/s vs. MHz: Accuracy in Measuring Memory Speed

When it comes to measuring memory speed, one of the most historically used measurements for memory performance is megahertz (commonly abbreviated as MHz). However, after years of debate between manufacturers, experts and gamers alike, many, including our team here at Patriot Memory, have started...

store.patriotmemory.com

 

[LazyGamer]

At a basic level, it's six of one and half a dozen of the other - but MT/s is the most technically correct terminology.

With just DDR signaling it's not really that big of a deal I think, since most of us are capable of just dividing by two. The days of QDR didn't really arrive, and I'm not aware of any standard used in desktop systems that run more than the two transfers per clock cycle of DDR signaling.

Where this makes more a difference is in situations where the clock signal is treated as a carrier signal, and the transfer signals are then sent dozens or hundreds of times per carrier signal pulse. This is a simplified view on how wireless transmissions send data, at which point the frequency of the carrier signal has as much to do with realized transfer speeds as it does effective signal range and power requirements.

 

[Peter_Brosdahl]

At a basic level, it's six of one and half a dozen of the other - but MT/s is the most technically correct terminology.

With just DDR signaling it's not really that big of a deal I think, since most of us are capable of just dividing by two. The days of QDR didn't really arrive, and I'm not aware of any standard used in desktop systems that run more than the two transfers per clock cycle of DDR signaling.

Where this makes more a difference is in situations where the clock signal is treated as a carrier signal, and the transfer signals are then sent dozens or hundreds of times per carrier signal pulse. This is a simplified view on how wireless transmissions send data, at which point the frequency of the carrier signal has as much to do with realized transfer speeds as it does effective signal range and power requirements.

I admit that I didn't know most of that, outside of halving of the DDR rate.