OK, the CPU does all that crazy stuff but we're still assuming it doesn't cheat at CMOV?
-
-
Replying to @BRIAN_____
It is hard to imagine how it would cheat. The move itself is essentially free either way, and the condition has to be resolved first.
2 replies 0 retweets 0 likes -
Replying to @thecomp1ler @BRIAN_____
Super easy to cheat: is the branch predictor to predict the direction and try to speculate further. But we specifically asked about this and they confirm in their white paper that this works on today's processors.
1 reply 0 retweets 1 like -
Replying to @chandlerc1024 @BRIAN_____
Possible, but what if it mispredicts? That would be a poor optimization. Also would need to have several CMOV ops with same condition to be worthwhile, and that is beyond the current uop-fusion Intel can do.
1 reply 0 retweets 1 like -
Replying to @thecomp1ler @BRIAN_____
It mispredicting would be just as bad as the branch predictor mispredicting? And handled the same way? Predictors today have *extremely* high hit rates. It shouldn't matter how many cmov operations you have, it matters how much speculative execution predicting it unlocks.
2 replies 0 retweets 0 likes -
Replying to @chandlerc1024 @BRIAN_____
In cryptography it would in theory be unpredictable.
2 replies 0 retweets 0 likes
That's true on average, for random inputs. If somebody is explicitly trying to trigger the branch prediction so that they can learn a secret based on the number of conditional subtractions skipped (e.g. RSA p and q, some ECC inputs) then not necessarily.
-
-
Replying to @BRIAN_____ @chandlerc1024
No of course not, I mean as an optimization it wouldn’t be successful if it relied on the branch predictor.
0 replies 0 retweets 0 likesThanks. Twitter will use this to make your timeline better. UndoUndo
-
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.