Prepare CUTEst.jl for single and quadruple precision #353
Conversation
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #353 +/- ##
===========================================
- Coverage 89.11% 60.66% -28.46%
===========================================
Files 5 8 +3
Lines 790 1388 +598
===========================================
+ Hits 704 842 +138
- Misses 86 546 +460 ☔ View full report in Codecov by Sentry. |
amontoison
changed the title
amontoison
force-pushed
the
multiprecision
branch
from
4df0046 to
2c95975
Compare
| global cutest_instances_single = 0 | ||
| global cutest_instances_double = 0 | ||
| global cutest_instances_quadruple = 0 | ||
| global cutest_lib_single = C_NULL | ||
| global cutest_lib_double = C_NULL | ||
| global cutest_lib_quadruple = C_NULL |
There was a problem hiding this comment.
Maybe we should at some point create a structure for this? That could be an issue with the tag good first issue.
There was a problem hiding this comment.
I added a SIF demuxer in the latest version of CUTEst such that we don't need anymore to update the global variables.
The issue is that package was not updated since long time we need many modifications if we want to easily upgrade it.
| @@ -226,7 +226,7 @@ function CUTEstModel( | |||
| nnzj = nnzj[] |> Int | |||
| nnzh = nnzh[] |> Int | |||
|
|
|||
| work = Vector{Int32}(undef, ncon) | |||
| work = Vector{Float64}(undef, ncon) | |||
There was a problem hiding this comment.
Could you add a one-line comment for this vector?
There was a problem hiding this comment.
I think we should add two vectors in the structure (one of size nvar and another of size ncon). We need these kind of buffer / workspace for a lot of CUTEst routines.
amontoison
force-pushed
the
multiprecision
branch
2 times, most recently
from
b6ddce9 to
e8ec19f
Compare
amontoison
force-pushed
the
multiprecision
branch
from
e8ec19f to
22d8487
Compare
Should be merged after #344, #349, #350, #351 and #352