You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Change SyncReadMem.read(addr, en) to not return the mport
[kevin] Just test what SiFive cores need here
[kevin] Remove CHIRRTL has been fundamentally broken for a long time and there are a lot of issues here related to enable inference, node vs. wire on read enable, etc.
[kevin] read enable in SRAM compiler: https://users.ece.cmu.edu/~koopman/ece548/hw/hw5/meml80.pdf
_“When CSN is high, the memory goes to stand-by (power down) mode and no access to the memory can”
**WARN: **_This datasheet might be from a very outdated 500nm technology from 1997. It would be nice if someone at SiFive could tell us if this is still relevant for 7nm.
[martin] The read enable in the data sheet is actually an output enable disabling the output driver to tri-state. This was needed in the old days for a bus based design. It might save some power, but most power saving in the datasheet is about chip select (CSN).
If we want to support OE then we should also call it output enable
The CSN also disable writes, not just reads
[kevin] the idea here was that the script invoking the SRAM compiler could take the conjunction of read and write enables to determine if the memory can be powered down
[jiuyang] How to cite SMT backend?
My project depends on SMT, and I'm wondering how to cite Kevin’s job.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
2021-08-10/11
Outreach
Discussion 1
_“When CSN is high, the memory goes to stand-by (power down) mode and no access to the memory can”
**WARN: **_This datasheet might be from a very outdated 500nm technology from 1997. It would be nice if someone at SiFive could tell us if this is still relevant for 7nm.
[schuyler] tl;dr: This is the plan to remove CircuitForm deprecations. This needs an owner.
sealed abstract class FirrtlEmitter(form: CircuitForm)
case class CircuitState(
Sorry, I was out-of-office last week, and failed to work on this.
I can spare time on this branch this week, do I need to work on this?
[schuyler] I will do something about this.
Discussion 2
instead ofin addition to -E experimental-smt2)Beta Was this translation helpful? Give feedback.
All reactions