deepseek-ai / DeepEP - Conditional Complexity

Intro

Conditional complexity (also called cyclomatic complexity) is a term used to measure the complexity of software. The term refers to the number of possible paths through a program function. A higher value ofter means higher maintenance and testing costs (infosecinstitute.com).
Conditional complexity is calculated by counting all conditions in the program that can affect the execution path (e.g. if statement, loops, switches, and/or operators, try and catch blocks...).
Conditional complexity is measured at the unit level (methods, functions...).
Units are classified in four categories based on the measured McCabe index: 1-5 (simple units), 6-10 (medium complex units), 11-25 (complex units), 26+ (very complex units).

Learn more...

Conditional Complexity Overall

There are 45 units with 399 lines of code in units (7.4% of code).

0 very complex units (0 lines of code)
0 complex units (0 lines of code)
1 medium complex units (32 lines of code)
3 simple units (114 lines of code)
41 very simple units (253 lines of code)

Legend:

51+

26-50

11-25

6-10

1-5

Alternative Visuals

3D view of all units

Conditional Complexity per Extension

51+

26-50

11-25

6-10

1-5

Conditional Complexity per Logical Component

primary logical decomposition

51+

26-50

11-25

6-10

1-5

Most Complex Units

Top 45 most complex units

Unit	# lines	McCabe index	# params
def dispatch() in deep_ep/buffer.py	32	11	16
def internode_dispatch() in deep_ep/buffer.py	43	9	16
def __init__() in deep_ep/buffer.py	32	8	6
void Buffer::sync() in csrc/deep_ep.cpp	39	7	3
def combine() in deep_ep/buffer.py	15	4	8
def low_latency_dispatch() in deep_ep/buffer.py	14	3	8
torch::Tensor Buffer::get_local_buffer_tensor() in csrc/deep_ep.cpp	7	3	3
def get_local_buffer_tensor() in deep_ep/buffer.py	7	2	5
def low_latency_combine() in deep_ep/buffer.py	10	2	9
def __exit__() in deep_ep/utils.py	3	2	4
size_t get_rdma_buffer_size_hint() in csrc/config.hpp	20	2	2
int Buffer::get_root_rdma_rank() in csrc/deep_ep.cpp	3	2	1
def set_num_sms() in deep_ep/buffer.py	3	1	1
def capture() in deep_ep/buffer.py	2	1	0
def get_low_latency_rdma_size_hint() in deep_ep/buffer.py	2	1	4
def get_dispatch_config() in deep_ep/buffer.py	15	1	1
def get_combine_config() in deep_ep/buffer.py	15	1	1
def get_dispatch_layout() in deep_ep/buffer.py	8	1	6
def internode_combine() in deep_ep/buffer.py	18	1	9
def clean_low_latency_buffer() in deep_ep/buffer.py	2	1	4
def get_next_low_latency_combine_buffer() in deep_ep/buffer.py	3	1	2
def __init__() in deep_ep/utils.py	4	1	3
def current_stream_wait() in deep_ep/utils.py	3	1	1
def __enter__() in deep_ep/utils.py	2	1	1
explicit EventHandle() in csrc/event.hpp	4	1	1
void current_stream_wait() in csrc/event.hpp	3	1	0
torch::Event create_event() in csrc/event.hpp	5	1	1
void stream_wait() in csrc/event.hpp	4	1	2
void stream_wait() in csrc/event.hpp	3	1	2
dtype_t cell_div() in csrc/config.hpp	3	1	2
dtype_t align() in csrc/config.hpp	3	1	2
size_t get_nvl_buffer_size_hint() in csrc/config.hpp	18	1	2
std::pair clean_meta() in csrc/config.hpp	4	1	0
out_ptr_t advance() in csrc/config.hpp	3	1	2
size_t get_low_latency_rdma_size_hint() in csrc/config.hpp	4	1	4
void Buffer::move_fifo_slots() in csrc/deep_ep.cpp	3	1	1
bool Buffer::is_available() in csrc/deep_ep.cpp	3	1	0
bool Buffer::is_internode_available() in csrc/deep_ep.cpp	3	1	0
int Buffer::get_num_rdma_ranks() in csrc/deep_ep.cpp	3	1	0
int Buffer::get_rdma_rank() in csrc/deep_ep.cpp	3	1	0
int Buffer::get_local_device_id() in csrc/deep_ep.cpp	3	1	0
pybind11::bytearray Buffer::get_local_ipc_handle() in csrc/deep_ep.cpp	3	1	0
pybind11::bytearray Buffer::get_local_nvshmem_unique_id() in csrc/deep_ep.cpp	5	1	0
void Buffer::clean_low_latency_buffer() in csrc/deep_ep.cpp	15	1	3
return torch::from_blob() in csrc/deep_ep.cpp	2	1	0