There were a huge number of dynamic memory allocations that I eliminated and replaced by stack allocations.
Unfortunatly that means that some parameter sets use a loooooot of stack - often more than the 8 MiB default stack on Linux. I excluded the parameter sets that exceed the 4 MiB available on the mps2-an386:
aimer-l1-param4
aimer-l3-param3
aimer-l3-param4
aimer-l5-param3
aimer-l5-param4
On nucleo-l4r5zi (640KB RAM), we can run
aimer-l1-param1
aimer-l1-param2
aimer-l3-param1
I successfully ran test.py and testvectors.py on both qemu and the nucleo-l4r5zi
🪣 Resolves https://github.com/mupq/pqm4/issues/261
This adds the AIMer implementations from the NIST submission package.
There were a huge number of dynamic memory allocations that I eliminated and replaced by stack allocations. Unfortunatly that means that some parameter sets use a loooooot of stack - often more than the 8 MiB default stack on Linux. I excluded the parameter sets that exceed the 4 MiB available on the
mps2-an386
:aimer-l1-param4
aimer-l3-param3
aimer-l3-param4
aimer-l5-param3
aimer-l5-param4
On
nucleo-l4r5zi
(640KB RAM), we can runaimer-l1-param1
aimer-l1-param2
aimer-l3-param1
I successfully ran
test.py
andtestvectors.py
on both qemu and thenucleo-l4r5zi