You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: CHANGELOG
+9-7Lines changed: 9 additions & 7 deletions
Original file line number
Diff line number
Diff line change
@@ -11,6 +11,11 @@ Summary
11
11
objects, which can be waited upon to synchronize with the completion of the
12
12
parallel operation.
13
13
14
+
Breaking API Changes
15
+
Synchronous Thrust algorithms now block until all of their operations have
16
+
completed. Use the new asynchronous Thrust algorithms for non-blocking
17
+
behavior.
18
+
14
19
New Features
15
20
`thrust::event` and `thrust::future<T>`, uniquely-owned asynchronous handles consisting of a state (ready or not ready), content (some value; for `thrust::future` only), and an optional set of objects that should be destroyed only when the future's value is ready and has been consumed.
16
21
The design is loosely based on C++11's `std::future`.
@@ -196,7 +201,7 @@ Bug Fixes
196
201
1777043 `complex` does not work with `sequence`.
197
202
198
203
#######################################
199
-
# Thrust v1.9.1-2 (CUDA 9.1) #
204
+
# Thrust v1.9.1 (CUDA 9.1) #
200
205
#######################################
201
206
202
207
Summary
@@ -209,7 +214,7 @@ Bug Fixes
209
214
1904217 Allow callables that take non-const refs to be used with reduce and scan.
210
215
211
216
#######################################
212
-
# Thrust v1.9.0-4 (CUDA 9.0) #
217
+
# Thrust v1.9.0 (CUDA 9.0) #
213
218
#######################################
214
219
215
220
Summary
@@ -242,14 +247,14 @@ Bug Fixes
242
247
Acknowledgments
243
248
Thanks to Manuel Schiller for contributing a C++11 based enhancement
244
249
regarding the deduction of functor return types, improving the performance
245
-
of thrust::unique and implementing transform_output_iterator.
250
+
of `thrust::unique` and implementing `thrust::transform_output_iterator`.
246
251
Thanks to Thibault Notargiacomo for the implementation of move semantics for
247
252
the vector_base based class.
248
253
Thanks to Duane Merrill for developing CUB and helping to integrate it into
249
254
Thrust's backend.
250
255
251
256
#######################################
252
-
# Thrust v1.8.3-2 (CUDA 8.0) #
257
+
# Thrust v1.8.3 (CUDA 8.0) #
253
258
#######################################
254
259
255
260
Summary
@@ -309,9 +314,6 @@ Summary
309
314
request a CUDA stream for kernels launched during algorithm execution. Finally, new CUDA algorithm
310
315
implementations provide substantial performance improvements.
311
316
312
-
Breaking API Changes
313
-
None.
314
-
315
317
New Features
316
318
Algorithms in CUDA __device__ code
317
319
Thrust algorithms may now be invoked from CUDA __device__ and __host__ __device__ functions.
0 commit comments