Thanks! I will take a look at this and see if my branch changes things at all. I’d also be curious about benchmarks across a variety of data sizes - these 4x3 matrices are probably not indicative of the types of data where we have performance issues [edit: oops, I see there are some with 20000 x 1000 as well]
seantalts
15
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Use of operands_and_partials slows down things? | 11 | 954 | February 4, 2018 | |
| Math library input checks | 53 | 3303 | March 7, 2019 | |
| A proposal for sparse matrices and GPs in Stan | 35 | 5212 | January 18, 2018 | |
| Operands and Partials: partials_ vs partials_vec_ | 10 | 1158 | February 3, 2018 | |
| Stan SIMD & Performance | 23 | 5008 | January 23, 2020 |