224 |
("BOOLTY", SizeOf.c_bool), |
("BOOLTY", SizeOf.c_bool), |
225 |
("CUDA_INNER_BLK_SIZE", "1024"), |
("CUDA_INNER_BLK_SIZE", "1024"), |
226 |
("CUDA_NO_BLOCKS", "1024"), |
("CUDA_NO_BLOCKS", "1024"), |
227 |
|
("CUDA_WORK_BLK_SIZES", "64"), |
228 |
("NEED_WORLD_COPY", if reqGlobalCopy then "#define GLOBAL_COPY_REQUIRED" else ""), |
("NEED_WORLD_COPY", if reqGlobalCopy then "#define GLOBAL_COPY_REQUIRED" else ""), |
229 |
("NEED_GLOBAL_COPY", if reqWorldCopy then "#define WORLD_COPY_REQUIRED" else ""), |
("NEED_GLOBAL_COPY", if reqWorldCopy then "#define WORLD_COPY_REQUIRED" else ""), |
230 |
(* START_{PARAMS,ARGS,ARGS_IN_WRLD} used for |
(* START_{PARAMS,ARGS,ARGS_IN_WRLD} used for |
374 |
val _ = if (#cudaPermute spec andalso #cudaGlobalQueue spec) |
val _ = if (#cudaPermute spec andalso #cudaGlobalQueue spec) |
375 |
then raise Fail "Cannot have permutation and GlobalQueueing enabled at the same time" |
then raise Fail "Cannot have permutation and GlobalQueueing enabled at the same time" |
376 |
else true |
else true |
377 |
|
val _ = if (#cudaBatch spec andalso #cudaGlobalQueue spec) |
378 |
|
then raise Fail "Cannot have batching and GlobalQueueing enabled at the same time" |
379 |
|
else true |
380 |
val cudaDefault = (not (#cudaPermute spec)) andalso (not (#cudaGlobalQueue spec)) |
val cudaDefault = (not (#cudaPermute spec)) andalso (not (#cudaGlobalQueue spec)) |
381 |
in |
in |
382 |
pp (#exec spec andalso #snapshot spec, "DIDEROT_EXEC_SNAPSHOT"); |
pp (#exec spec andalso #snapshot spec, "DIDEROT_EXEC_SNAPSHOT"); |