By: Chester (lamchester.delete@this.gmail.com), November 30, 2021 7:25 am
Room: Moderated Discussions
I've been loosely following the rowhammer discussion, and was wondering if we could just get DRAM cells resilient enough to not flip when the memory controller doesn't tell it to (aka, be not vulnerable to rowhammer in the simplest, most fundamental way). Or at least, even under the worst attack conditions, have an error rate that's not significantly different from normal operation.
What would the cost/performance penalty be for such DRAM? Does anyone have info? All the searches I've done turn up mitigation techniques.
What would the cost/performance penalty be for such DRAM? Does anyone have info? All the searches I've done turn up mitigation techniques.
Topic | Posted By | Date |
---|---|---|
Cost of rowhammer invulnerability? | Chester | 2021/11/30 07:25 AM |
Cost of rowhammer invulnerability? | Doug S | 2021/11/30 08:35 AM |
Cost of rowhammer invulnerability? | Dan Fay | 2021/11/30 11:12 AM |
Cost of rowhammer invulnerability? | Chester | 2021/12/02 07:29 AM |
Cost of rowhammer invulnerability? | Michael S | 2021/12/03 04:55 AM |
Cost of rowhammer invulnerability? | dmcq | 2021/11/30 09:29 AM |
Cost of rowhammer invulnerability? | Chester | 2021/12/02 11:42 AM |
Cost of rowhammer invulnerability? | Brendan | 2021/12/01 02:33 AM |
Cost of rowhammer invulnerability? | David Hess | 2021/12/04 01:55 AM |
Cost of rowhammer invulnerability? | peceed | 2021/12/29 01:20 PM |