style: format for 2024 edition

Signed-off-by: YdrMaster <[email protected]>

Author: YdrMaster

operators/src/add/args.rs

@@ -1,6 +1,6 @@
 use crate::{
-    rank_mismatch, shape_mismatch, shape_not_support, utils::type_distinct, ConstPtr, Hardware,
-    LaunchError, MutPtr, TensorLayout,
+    ConstPtr, Hardware, LaunchError, MutPtr, TensorLayout, rank_mismatch, shape_mismatch,
+    shape_not_support, utils::type_distinct,
 };
 use digit_layout::DigitLayout;
 use itertools::izip;

operators/src/add/common_cpu/mod.rs

@@ -1,5 +1,5 @@
-use super::{args::Scheme, Add, Args};
-use crate::{common_cpu::Cpu, ByteOf, LaunchError, QueueAlloc};
+use super::{Add, Args, args::Scheme};
+use crate::{ByteOf, LaunchError, QueueAlloc, common_cpu::Cpu};
 use digit_layout::types as ty;
 use half::f16;
 use rayon::iter::{IntoParallelIterator, ParallelIterator};

operators/src/add/cuda/mod.rs

@@ -1,9 +1,9 @@
-use super::{args::Scheme, Add, Args};
+use super::{Add, Args, args::Scheme};
 use crate::{
-    cuda::{dt_name, Gpu, Handle, ModuleBox},
+    ByteOf, LaunchError, QueueAlloc, SchemeDiversity,
+    cuda::{Gpu, Handle, ModuleBox, dt_name},
     shape_not_support, strides_not_support,
     utils::gcd,
-    ByteOf, LaunchError, QueueAlloc, SchemeDiversity,
 };
 use cuda::params;
 use digit_layout::DigitLayout;
@@ -110,8 +110,8 @@ mod test {
     use super::{Args, Gpu, Operator};
     use crate::{Hardware, Operator as _, TensorLayout};
     use digit_layout::{
-        types::{F16, F64},
         DigitLayout,
+        types::{F16, F64},
     };
 
     fn args<H: Hardware>(

operators/src/add_rows/args.rs

@@ -1,7 +1,6 @@
 use crate::{
-    type_not_support,
+    ConstPtr, Hardware, LaunchError, MutPtr, TensorLayout, type_not_support,
     utils::{dim_distinct, rank_error, type_distinct},
-    ConstPtr, Hardware, LaunchError, MutPtr, TensorLayout,
 };
 use digit_layout::{DigitLayout, LayoutContent::Unsigned};
 use std::ptr::{null, null_mut};

operators/src/add_rows/common_cpu/mod.rs

@@ -1,5 +1,5 @@
-use super::{args::Meta, AddRows, Args};
-use crate::{common_cpu::Cpu, ByteOf, LaunchError, QueueAlloc, Unsigned};
+use super::{AddRows, Args, args::Meta};
+use crate::{ByteOf, LaunchError, QueueAlloc, Unsigned, common_cpu::Cpu};
 use digit_layout::types as ty;
 use half::f16;
 use rayon::iter::{IntoParallelIterator, ParallelIterator};

operators/src/add_rows/cuda/mod.rs

@@ -1,10 +1,10 @@
 use super::{AddRows, Args};
 use crate::{
+    ByteOf, LaunchError, QueueAlloc, SchemeDiversity,
     add_rows::args::Meta,
-    cuda::{dt_name, Gpu, Handle, ModuleBox},
+    cuda::{Gpu, Handle, ModuleBox, dt_name},
     strides_not_support,
     utils::gcd,
-    ByteOf, LaunchError, QueueAlloc, SchemeDiversity,
 };
 use cuda::params;
 use digit_layout::DigitLayout;
@@ -144,11 +144,11 @@ extern "C" __global__ void {name}(
 #[cfg(test)]
 mod test {
     use super::{Args, Gpu, Operator};
-    use crate::{cuda::cast_load, Hardware, Operator as _, TensorLayout};
+    use crate::{Hardware, Operator as _, TensorLayout, cuda::cast_load};
     use cuda::memcpy_d2h;
     use digit_layout::{
-        types::{F16, F64, U32},
         DigitLayout,
+        types::{F16, F64, U32},
     };
     use half::f16;
 

operators/src/all_reduce/args.rs

@@ -1,6 +1,6 @@
-use super::ReduceOp;
+use super::ReduceOp;
 use crate::{
-    rearrange, shape_mismatch, strides_not_support, utils::type_distinct, Hardware, LaunchError,
+    Hardware, LaunchError, rearrange, shape_mismatch, strides_not_support, utils::type_distinct,
 };
 use digit_layout::DigitLayout;
 

operators/src/all_reduce/common_cpu.rs

@@ -1,8 +1,9 @@
-use super::{args::Meta, AllReduce, Args, ReduceOp};
+use super::{AllReduce, Args, ReduceOp, args::Meta};
 use crate::{
+    ByteOf, LaunchError, QueueAlloc, TopoNode,
     broadcast::{self, common_cpu::Operator as Broadcast},
     common_cpu::{Cpu, InprocNode},
-    rearrange, ByteOf, LaunchError, QueueAlloc, TopoNode,
+    rearrange,
 };
 use digit_layout::DigitLayout;
 use half::{bf16, f16};
@@ -138,7 +139,7 @@ fn sum<T: AddAssign + Clone>(len: usize, buf: *mut u8, src: *const u8) {
 
 #[test]
 fn test_comm() {
-    use crate::{common_cpu::ThisThread, Operator as _, TensorLayout};
+    use crate::{Operator as _, TensorLayout, common_cpu::ThisThread};
     use digit_layout::types::U32;
 
     InprocNode::new(4)

operators/src/all_reduce/infini.rs

@@ -1,8 +1,8 @@
-use super::{args::Meta, AllReduce, Args, ReduceOp};
+use super::{AllReduce, Args, ReduceOp, args::Meta};
 use crate::{
+    ByteOf, LaunchError, QueueAlloc,
     infini::{Device, InfiniNode},
     rearrange::{self, infini::Operator as Rearrange},
-    ByteOf, LaunchError, QueueAlloc,
 };
 use digit_layout::types as ty;
 use infini_ccl::bindings::InfiniDataType_t;

operators/src/all_reduce/nccl.rs

@@ -1,7 +1,8 @@
-use super::{args::Meta, AllReduce, Args, ReduceOp};
+use super::{AllReduce, Args, ReduceOp, args::Meta};
 use crate::{
+    ByteOf, LaunchError, LaunchError, QueueAlloc,
     cuda::{Gpu, NcclNode},
-    rearrange, ByteOf, LaunchError, LaunchError, QueueAlloc,
+    rearrange,
 };
 use std::{
     slice::{from_raw_parts, from_raw_parts_mut},