8 ms·
Hmm. If this memory is faster than DRAM, wouldn't it be quick to test, say, ten trillion write/erase cycles? Why stop at 10M? Is the erase operation really slo
by runeks 3y ago
Hmm. If this memory is faster than DRAM, wouldn't it be quick to test, say, ten trillion write/erase cycles?
Why stop at 10M? Is the erase operation really slow?
- idiotsecant 3y agoProbably a case of don't ask questions you don't want answers to.
- dahfizz 3y agoThe paper says they tested the durability of the ram with a 5ms program-read-erase-read loop. Meaning each time they program-read-erase-read, it takes 5 milliseconds. Ten trillion cycles would take over 150 years. I'm guessing a silicon lab doesn't have "the rest of the computer" that would allow them to run this ram at full speed constantly. This UltraRAM isn't something they can just slot into their motherboard.
- DoctorOetker 3y agoThese labs have their custom structures synthesized, adding a small circuit specifically for endurance testing would be trivial compared to what they have achieved designing and implementing the structures they have.
- lawlessone 3y agoCould processors keep up with it?
- DoctorOetker 3y agoHalf in jest: if processors couldn't keep up with it, it would be used to replace flip-flops etc in processors...
- amelius 3y agoThat doesn't necessarily mean the processors would become any faster ...
- DougMerritt 3y agoIt kind of does, all else being equal. What else do you have in mind? Parasitic capacitance on signal lines?
- georgeburdell 3y agoThis is a common problem in memory. Oftentimes they use models to accelerate the wear and tear through temperature, voltage, etc. and extrapolate the lifetime
- hinkley 3y ago5ms is 200 cycles per second. 3600 seconds in an hour. 0.72 million writes per hour. Almost 40 million if I start it on Friday evening and stop it on Monday morning 10 million is 14 hours. It takes longer than that to prepare your documentation. Something is rotten in Denmark. A skeptic could very, very reasonably assume that cherry-picking is going on here, and that 10m to degradation isn't far off from the truth.
- antx 3y agoIndeed, the paper says: > Assuming ideal capacitive scaling[33] down to state-of-the-art feature sizes, the switching performance would be faster than DRAM, although testing on smaller feature size devices is required to confirm this. So, they have no idea of its performance. Yet.