Various improvements to the docs

[giordano]

I had some...uhm...fun in the last couple of days trying to port some C++ CUDA code to CUDA.jl, and profile it. I dumped into this PR my experience, hoping to make lives of people after me a little bit easier 🙂

[github-actions]

Your PR requires formatting changes to meet the project's style guidelines.
Please consider running Runic (git runic master) to apply these changes.

Click here to view the suggested changes.

diff --git a/src/device/intrinsics/indexing.jl b/src/device/intrinsics/indexing.jl
index a42b003cd..36cde4ab9 100644
--- a/src/device/intrinsics/indexing.jl
+++ b/src/device/intrinsics/indexing.jl
@@ -66,32 +66,32 @@ end
 """
     gridDim()::NamedTuple
 
-Returns the dimensions of the grid as a `NamedTuple` with keys `x`, `y`, and `z`.
-These dimensions have the same starting index as the `gridDim` built-in variable in the C/C++ extension.
+    Returns the dimensions of the grid as a `NamedTuple` with keys `x`, `y`, and `z`.
+    These dimensions have the same starting index as the `gridDim` built-in variable in the C/C++ extension.
 """
 @inline gridDim() =   (x=gridDim_x(),   y=gridDim_y(),   z=gridDim_z())
 
 """
     blockIdx()::NamedTuple
 
-Returns the block index within the grid as a `NamedTuple` with keys `x`, `y`, and `z`.
-These indices are 1-based, unlike the `blockIdx` built-in variable in the C/C++ extension which is 0-based.
+    Returns the block index within the grid as a `NamedTuple` with keys `x`, `y`, and `z`.
+    These indices are 1-based, unlike the `blockIdx` built-in variable in the C/C++ extension which is 0-based.
 """
 @inline blockIdx() =  (x=blockIdx_x(),  y=blockIdx_y(),  z=blockIdx_z())
 
 """
     blockDim()::NamedTuple
 
-Returns the dimensions of the block as a `NamedTuple` with keys `x`, `y`, and `z`.
-These dimensions have the same starting index as the `blockDim` built-in variable in the C/C++ extension.
+    Returns the dimensions of the block as a `NamedTuple` with keys `x`, `y`, and `z`.
+    These dimensions have the same starting index as the `blockDim` built-in variable in the C/C++ extension.
 """
 @inline blockDim() =  (x=blockDim_x(),  y=blockDim_y(),  z=blockDim_z())
 
 """
     threadIdx()::NamedTuple
 
-Returns the thread index within the block as a `NamedTuple` with keys `x`, `y`, and `z`.
-These indices are 1-based, unlike the `threadIdx` built-in variable in the C/C++ extension which is 0-based.
+    Returns the thread index within the block as a `NamedTuple` with keys `x`, `y`, and `z`.
+    These indices are 1-based, unlike the `threadIdx` built-in variable in the C/C++ extension which is 0-based.
 """
 @inline threadIdx() = (x=threadIdx_x(), y=threadIdx_y(), z=threadIdx_z())
 
@@ -99,7 +99,7 @@ These indices are 1-based, unlike the `threadIdx` built-in variable in the C/C++
     warpsize()::Int32
 
 Returns the warp size (in threads).
-This corresponds to the `warpSize` built-in variable in the C/C++ extension.
+    This corresponds to the `warpSize` built-in variable in the C/C++ extension.
 """
 @inline warpsize() = ccall("llvm.nvvm.read.ptx.sreg.warpsize", llvmcall, Int32, ())
 
@@ -107,7 +107,7 @@ This corresponds to the `warpSize` built-in variable in the C/C++ extension.
     laneid()::Int32
 
 Returns the thread's lane within the warp.
-This ID is 1-based.
+    This ID is 1-based.
 """
 @inline laneid() = ccall("llvm.nvvm.read.ptx.sreg.laneid", llvmcall, Int32, ()) + 1i32
 

[github-actions]

CUDA.jl Benchmarks

Details

Benchmark suite	Current: 337b7a7	Previous: 7a27d77	Ratio
latency/precompile	44169430501.5 ns	44455759835 ns	0.99
latency/ttfp	13133988338 ns	13140153243 ns	1.00
latency/import	3764428041 ns	3755312424 ns	1.00
integration/volumerhs	9435111.5 ns	9442840 ns	1.00
integration/byval/slices=1	145882 ns	145598 ns	1.00
integration/byval/slices=3	423338 ns	422554 ns	1.00
integration/byval/reference	144101 ns	143811 ns	1.00
integration/byval/slices=2	284586 ns	284011 ns	1.00
integration/cudadevrt	102648 ns	102397 ns	1.00
kernel/indexing	13579 ns	13434 ns	1.01
kernel/indexing_checked	14218 ns	13908 ns	1.02
kernel/occupancy	647.9156626506024 ns	644.5636363636364 ns	1.01
kernel/launch	2059.4 ns	2090.3 ns	0.99
kernel/rand	14467 ns	14479 ns	1.00
array/reverse/1d	19050 ns	18661 ns	1.02
array/reverse/2dL_inplace	66409 ns	66252 ns	1.00
array/reverse/1dL	69247 ns	68893 ns	1.01
array/reverse/2d	21375 ns	21087 ns	1.01
array/reverse/1d_inplace	10869.333333333334 ns	10503.833333333332 ns	1.03
array/reverse/2d_inplace	11034 ns	11399.5 ns	0.97
array/reverse/2dL	73430.5 ns	73163 ns	1.00
array/reverse/1dL_inplace	66552 ns	66146 ns	1.01
array/copy	18566 ns	18502.5 ns	1.00
array/iteration/findall/int	146514 ns	146476.5 ns	1.00
array/iteration/findall/bool	130975 ns	130795 ns	1.00
array/iteration/findfirst/int	84094.5 ns	84133 ns	1.00
array/iteration/findfirst/bool	81344 ns	81624.5 ns	1.00
array/iteration/scalar	65699 ns	65804 ns	1.00
array/iteration/logical	199014 ns	198187.5 ns	1.00
array/iteration/findmin/1d	86127.5 ns	86504 ns	1.00
array/iteration/findmin/2d	117201 ns	117154 ns	1.00
array/reductions/reduce/Int64/1d	38955.5 ns	41088.5 ns	0.95
array/reductions/reduce/Int64/dims=1	51296 ns	52190.5 ns	0.98
array/reductions/reduce/Int64/dims=2	59074 ns	59179 ns	1.00
array/reductions/reduce/Int64/dims=1L	87341 ns	87126 ns	1.00
array/reductions/reduce/Int64/dims=2L	84657.5 ns	84418.5 ns	1.00
array/reductions/reduce/Float32/1d	34039.5 ns	34001 ns	1.00
array/reductions/reduce/Float32/dims=1	49262 ns	39890 ns	1.23
array/reductions/reduce/Float32/dims=2	56573 ns	55899 ns	1.01
array/reductions/reduce/Float32/dims=1L	51645 ns	51535 ns	1.00
array/reductions/reduce/Float32/dims=2L	69784.5 ns	69798 ns	1.00
array/reductions/mapreduce/Int64/1d	38808 ns	40980.5 ns	0.95
array/reductions/mapreduce/Int64/dims=1	51392.5 ns	41741 ns	1.23
array/reductions/mapreduce/Int64/dims=2	58970 ns	59036 ns	1.00
array/reductions/mapreduce/Int64/dims=1L	87347 ns	87134 ns	1.00
array/reductions/mapreduce/Int64/dims=2L	84631.5 ns	84427 ns	1.00
array/reductions/mapreduce/Float32/1d	33862 ns	33457 ns	1.01
array/reductions/mapreduce/Float32/dims=1	49049 ns	48711 ns	1.01
array/reductions/mapreduce/Float32/dims=2	56551 ns	55941 ns	1.01
array/reductions/mapreduce/Float32/dims=1L	51676 ns	51352 ns	1.01
array/reductions/mapreduce/Float32/dims=2L	69531 ns	68956 ns	1.01
array/broadcast	20650.5 ns	20251 ns	1.02
array/copyto!/gpu_to_gpu	10746.333333333334 ns	10684.333333333334 ns	1.01
array/copyto!/cpu_to_gpu	216903 ns	214898 ns	1.01
array/copyto!/gpu_to_cpu	283112 ns	281876 ns	1.00
array/accumulate/Int64/1d	118907 ns	118336 ns	1.00
array/accumulate/Int64/dims=1	79884 ns	79780 ns	1.00
array/accumulate/Int64/dims=2	157192 ns	155968.5 ns	1.01
array/accumulate/Int64/dims=1L	1707094 ns	1694089 ns	1.01
array/accumulate/Int64/dims=2L	961377 ns	960949 ns	1.00
array/accumulate/Float32/1d	101446.5 ns	100823 ns	1.01
array/accumulate/Float32/dims=1	76486 ns	76350 ns	1.00
array/accumulate/Float32/dims=2	144456.5 ns	144365 ns	1.00
array/accumulate/Float32/dims=1L	1585237 ns	1584729 ns	1.00
array/accumulate/Float32/dims=2L	657765 ns	656302 ns	1.00
array/construct	1277.9 ns	1283.1 ns	1.00
array/random/randn/Float32	43362 ns	36610 ns	1.18
array/random/randn!/Float32	30384 ns	30335 ns	1.00
array/random/rand!/Int64	29612 ns	26934 ns	1.10
array/random/rand!/Float32	8249.666666666666 ns	8186.666666666667 ns	1.01
array/random/rand/Int64	35367 ns	30201.5 ns	1.17
array/random/rand/Float32	12585 ns	12396 ns	1.02
array/permutedims/4d	51060.5 ns	52729 ns	0.97
array/permutedims/2d	52784.5 ns	52645 ns	1.00
array/permutedims/3d	53039 ns	53080 ns	1.00
array/sorting/1d	2735543 ns	2736443 ns	1.00
array/sorting/by	3305108.5 ns	3305811 ns	1.00
array/sorting/2d	1068212 ns	1071655.5 ns	1.00
cuda/synchronization/stream/auto	993.375 ns	1034.5263157894738 ns	0.96
cuda/synchronization/stream/nonblocking	7660.700000000001 ns	7705.9 ns	0.99
cuda/synchronization/stream/blocking	816.2019230769231 ns	784.4516129032259 ns	1.04
cuda/synchronization/context/auto	1146 ns	1133.5 ns	1.01
cuda/synchronization/context/nonblocking	7203.1 ns	7594.6 ns	0.95
cuda/synchronization/context/blocking	892.9795918367347 ns	885.6792452830189 ns	1.01

This comment was automatically generated by workflow using github-action-benchmark.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.48%. Comparing base (7a27d77) to head (337b7a7).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3030      +/-   ##
==========================================
+ Coverage   89.46%   89.48%   +0.01%     
==========================================
  Files         148      148              
  Lines       13047    13047              
==========================================
+ Hits        11673    11675       +2     
+ Misses       1374     1372       -2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.