[CakeML-dev] sec_ok failure
Ramana.Kumar at cl.cam.ac.uk
Sat Mar 11 13:00:48 UTC 2017
In my attempt to debug this interactively the sec_ok tests succeeded.
So I don't know why I see EQT_ELIM exception non-interactively - it
could be an out-of-memory (Poly/ML) issue...
I've added something to print the instruction that fails if one ever
On 10 March 2017 at 20:44, Magnus Myreen <magnus.myreen at gmail.com> wrote:
> Would it be possible to make it print the asm instructions that failed sec_ok?
> On the topic of sec_ok, Anthony and I have discussed extending the
> base encoder configs with config-wrappers. The idea is to take the
> base encoding and use it to produce a new more capable encoding
> function by calling the base encoding on multiple asm instructions to
> implement a single asm instruction. I just wanted to mention this
> because it might not have been mentioned on the lists before.
> On 10 March 2017 at 07:41, Magnus Myreen <magnus.myreen at gmail.com> wrote:
>> I wonder if the range of the Loc instruction is too small at the moment on x64.
>> On 10 March 2017 at 00:03, Ramana Kumar <Ramana.Kumar at cl.cam.ac.uk> wrote:
>>> I'm seeing sec_ok failures in the compiler bootstrap from around at
>>> least 3072c3a8. Possibly these came out of the refs branch? I'll need
>>> to run it interactively to investigate..
>>> Developers mailing list
>>> Developers at cakeml.org
More information about the Developers