portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-04-09 21:50:45 +01:00
Code Releases Activity

Update todo list

Browse Source
This commit is contained in:
Peter Boyle 2019-01-02 22:07:51 +00:00
parent e73b909a48
commit 535a6aaf05
1 changed files with 22 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

36
TODO
View File

@ -1,10 +1,11 @@
TODO:
---------------
GPU branch code item work list GPU branch code item work list
----------------------------- -----------------------------
- GPU offload reductions; thrust initial ; inclusive_scan vs reduce? TODO:
---------------
- Make GPU offload reductions optionally deterministic
- Accelerate the cshift - Accelerate the cshift
- Accelerate non-dslash elements of Mobius; check accelerator_loop uniformly used in fermion operators - Accelerate non-dslash elements of Mobius; check accelerator_loop uniformly used in fermion operators
- Gamma tables on GPU - Gamma tables on GPU
@ -24,17 +25,12 @@ GPU branch code item work list
- AVX512 still broken, lebesgue order missing ? - AVX512 still broken, lebesgue order missing ?
- Neon ?? - Neon ??
DONE:
----------------------------- -----------------------------
Physics item work list: - GPU offload reductions: using thrust::reduce?
- Deprecate JSON.
2)- Consistent linear solver flop count/rate -- PARTIAL, time but no flop/s yet - pugixml difficult.
4)- Multigrid Wilson and DWF, compare to other Multigrid implementations - Eigen problematic.
5)- HDCR resume
-----------------------------
DONE
- Audit HMC timestep / traj length size - Audit HMC timestep / traj length size
- GPU offload reductions; thrust initial ; inclusive_scan vs reduce? - GPU offload reductions; thrust initial ; inclusive_scan vs reduce?
- Pragmas.h - prune and remove strong_inline (?) - Pragmas.h - prune and remove strong_inline (?)
@ -45,7 +41,7 @@ DONE
- - (3) omp for collapse(n) - - (3) omp for collapse(n)
- - (4) omp parallel for collapse(n) - - (4) omp parallel for collapse(n)
- - Only (1) has a natural mirror in accelerator_loop - - Only (1) has a natural mirror in accelerator_loop
- - Nested loop macros get cumbersome - - Nested loop macros get cumbersome made a generic interface for N deep
- - Don't like thread_region and thread_loop_in_region - - Don't like thread_region and thread_loop_in_region
- - Could replace with - - Could replace with
@ -64,6 +60,18 @@ DONE
and same "in_region". and same "in_region".
-----------------------------
Physics item work list:
-----------------------------
2)- Consistent linear solver flop count/rate -- PARTIAL, time but no flop/s yet
4)- Multigrid Wilson and DWF, compare to other Multigrid implementations
5)- HDCR resume
-----------------------------
Nov 2018 Nov 2018
1)- BG/Q port and check ; Andrew says ok. 1)- BG/Q port and check ; Andrew says ok.