the cool M1/2 chips himself are top but look like apple did not eat the spectre still --> i will be very surprise if still a meltdown side channel is possible is less difficult to mitigate --> but so far there is no research inside public
such design flaws like rowhammer + meltdown--> but logic flaws like spectre are bastards that will sit for long --> cool for us)
meltdown is easy fix --> this is not a secret but with rowhammer --> if you can change physics --> you can fix him
--> some extra mitigations sit on top of ECC is good --> there are still top experimental papers --> but they are still beer talk --> not many practical results
everything for M1 is from 5nm from TSMC)) --> sad cry from amd+intel
everything is thermal cool because of less power consumption compare to amd 7mm from TSMC +intel latest chips (10mm?)
there is speculative execution --> so all the rich cool members with the cool M1/2 chips himself --> can anyone throw some spectre gadgets --> forget about meltdown
--> simple script kid cannot clone such git repo simple side channel inside x86 arch --> change intrinsics --> port spectre to arm
from my knowledge --> there is no rollback for modify to L1 or L2 in case of branch misprediction --> is that correct?
i immediately remember reading such ARM killed array bounds spectre with a new instruction for speculative barrier --> insert between array bounds test + array access himself
arstechnica.com/gadgets/2018/01/meltdown-and-spectre-heres-what-intel-apple-microsoft-others-are-doing-about-it/
my final thoughts -->
meltdown --> 90% killed
rowhammer --> if you study physics --> you can say he is still alive lurking inside DRAM to wait for the first bit flips))
--> also yellow raise such good point --> memory cell size is decrease himself
but you can ask ECC will protect my DRAM from attackers like drpalpatine? --> funny but only 60% true --> ECC rely on accuracy inside signal detection --> with decrease cell size --> SNR decreases + there is too much error rate fuck ECC inside ass for example high frequency/clusters
spectre --> 90% alive --> but i am hope that remaining 10% is true --> birth to new variants of spectre/ even new side channels himself
there is top interest but not hardware --> please who knows what? --> did anyone study speculative execution inside the apple silicon?
such design flaws like rowhammer + meltdown--> but logic flaws like spectre are bastards that will sit for long --> cool for us)
meltdown is easy fix --> this is not a secret but with rowhammer --> if you can change physics --> you can fix him
--> some extra mitigations sit on top of ECC is good --> there are still top experimental papers --> but they are still beer talk --> not many practical results
everything for M1 is from 5nm from TSMC)) --> sad cry from amd+intel
everything is thermal cool because of less power consumption compare to amd 7mm from TSMC +intel latest chips (10mm?)
there is speculative execution --> so all the rich cool members with the cool M1/2 chips himself --> can anyone throw some spectre gadgets --> forget about meltdown
--> simple script kid cannot clone such git repo simple side channel inside x86 arch --> change intrinsics --> port spectre to arm
from my knowledge --> there is no rollback for modify to L1 or L2 in case of branch misprediction --> is that correct?
i immediately remember reading such ARM killed array bounds spectre with a new instruction for speculative barrier --> insert between array bounds test + array access himself
arstechnica.com/gadgets/2018/01/meltdown-and-spectre-heres-what-intel-apple-microsoft-others-are-doing-about-it/
my final thoughts -->
meltdown --> 90% killed
rowhammer --> if you study physics --> you can say he is still alive lurking inside DRAM to wait for the first bit flips))
--> also yellow raise such good point --> memory cell size is decrease himself
but you can ask ECC will protect my DRAM from attackers like drpalpatine? --> funny but only 60% true --> ECC rely on accuracy inside signal detection --> with decrease cell size --> SNR decreases + there is too much error rate fuck ECC inside ass for example high frequency/clusters
spectre --> 90% alive --> but i am hope that remaining 10% is true --> birth to new variants of spectre/ even new side channels himself
there is top interest but not hardware --> please who knows what? --> did anyone study speculative execution inside the apple silicon?
Последнее редактирование:
