Are there any cycle-accurate or -reporting modern x86(-64)+SIMD emulators (thus, for specific microarchs) beside https://software.intel.com/en-us/articles/intel-architecture-code-analyzer …?
-
-
Replying to @solardiz @matthew_d_green
Are modern x86 systems cycle accurate in any predictable sense?
2 replies 1 retweet 3 likes -
Given all the clock domains I'm not so sure...
1 reply 0 retweets 0 likes -
Are there multiple clocks inside Intel now?
1 reply 0 retweets 0 likes -
Don't they have separate clocks for uncore, cores, caches, memory controller, etc these days?
1 reply 0 retweets 0 likes -
Uncore has been spectate from core since Haswell I think. Not certain on other components.
1 reply 0 retweets 0 likes -
How separate? Literally independent crystals, asynchronously bound?
3 replies 0 retweets 0 likes -
Every Intel CPU since Pentium days uses 1 crystal which is multiplied up by a clock chip that Intel specifies, gives the bus & cpu clks
1 reply 0 retweets 0 likes -
integer multipliers of the same clock are, unfortunately, not what I'm looking for :\
1 reply 0 retweets 0 likes -
It's probably *possible* to extract entropy out of clock phasing, but *depending* on it would be a lot more difficult.
1 reply 0 retweets 0 likes
E.g. you can configure a DLL to be "loose" enough that it drifts quite a lot. Also, clocks are often spread spectrum.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.