youngkingdom/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix several broken links (#1168)

Browse Source 
Co-authored-by: isaacw <isaacw@nvidia.com>
This commit is contained in:
wang-y-z
2023-11-03 12:01:25 +08:00
committed by GitHub
parent 557be3ab0e
commit 1d7f2a207e
5 changed files with 11 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
media/docs/build/building_in_windows_with_visual_studio.md vendored
View File

@ -1,4 +1,4 @@
[README](../README.md#documentation) > **CUTLASS 3.0: Building on Windows with Visual Studio**
[README](/README.md#documentation) > **CUTLASS 3.0: Building on Windows with Visual Studio**
# Building on Windows with Visual Studio

2
media/docs/build/building_with_clang_as_host_compiler.md vendored
View File

@ -1,4 +1,4 @@
[README](../README.md#documentation) > **CUTLASS 3: Building with Clang as host compiler**
[README](/README.md#documentation) > **CUTLASS 3: Building with Clang as host compiler**
# Building with Clang as host compiler

2
media/docs/efficient_gemm.md
View File

@ -241,7 +241,7 @@ The third kernel design is the [*Warp-Specialized Persistent Ping-Pong*](/includ
Like the Warp-Specialized Persistent Cooperative, kernel the concepts of warp groups, barrier synchronization between warp groups, and the shape of the grid launch remain the same in the persistent ping-pong design.
The distinctive feature of the Warp-Specialized Persistent Ping-Pong kernel is the following :
* The two *consumer* warp groups are assigned a different output tile using the Tile Scheduler. This allows for *epilogue* of one *consumer* warp group to be overlapped with the math operations of the other *consumer* warp group - thus maximizing tensor core utilization.
* The *producer* warp group synchronizes using the [Ordered Sequence Barrier](/include/cutlass/pipeline.hpp) to fill buffers of the two *consumer* warp groups one after the other in order.
* The *producer* warp group synchronizes using the [Ordered Sequence Barrier](/include/cutlass/pipeline/pipeline.hpp) to fill buffers of the two *consumer* warp groups one after the other in order.
# Resources

2
media/docs/pipeline.md
View File

@ -174,7 +174,7 @@ Please note that this is a basic example.
There are different versions possible,
depending on what the producer and consumer threads are doing.
Please refer to our [unit tests](/test/unit/pipeline)
and the other [pipeline classes](/include/cutlass/pipeline.hpp)
and the other [pipeline classes](/include/cutlass/pipeline/pipeline.hpp)
for more details.
# Copyright