update md files

Author: chuchunf

README.md

@@ -30,16 +30,12 @@ running all performance testing benchmarks
 ### Usage
 Please refer to [example](example) 
 
-## Versioning
-- 0.1
-    - Initial Release
-
-## Roadmap
-- [x] support more waiting strategies 
-- [ ] support publish N messages
-- [ ] support consume N messages
-- [ ] support proper DSL
-- [ ] more examples
+## Plan
+- [ ] Sequence 
+- [ ] Waiting Strategy
+- [ ] Sequence Barrier
+- [ ] Sequencer
+- [ ] RingBuffer
 
 ## License
 Distributed under the MIT License. See [LICENSE.txt](LICENSE.txt) for more information.

internal/sequence.md

@@ -1,8 +1,9 @@
 ## Sequence
-Sequence is the fundemtal data strcture for the entire package.
+Sequence is the fundamental data structure for the entire package.
 
-In short, sequence is just a thread-safe counter, which for the consumer and producer to wait, get then process
-a slot from the ring buffer. Refer to Sequencer for more details.
+In short, sequence is just a thread-safe counter, points to a slot in the ring buffer. 
+The consumer and producer uses a wait strategy to wait for the number. 
+Once get the slot, then the consumer or producer can continue its process. 
 
 Following Sequences are tested
 * struct with int64
@@ -13,22 +14,26 @@ All performance testing are done in Windows 10 with i5-11400F @ 2.60GHz in conso
 ```powershell
 #bechmark
 go test -benchmem -run=^$ -tags -race -bench ^BenchmarkSequenceGet$ goDisruptor/internal
+
 #profile
 go test -cpuprofile cpu.prof -memprofile mem.prof -benchmem -run=^$ -tags -race -bench ^BenchmarkSequenceGet$ goDisruptor/internal
+
 #get top function calls
 go tool pprof cpu.prof => top
+
 # generate call graph
 go tool pprof cpu.prof => png
+
 # get top memory allocation
 go tool pprof mem.prof => top
 ```
 
 ### Performance results 
-||Get without GC|Get with GC|Set without GC|Set with GC|Concurrent Get and Set without GC
-| -- | -- | -- | -- | -- | -- |
-|struct int64|10.18 ns/op|11.42 ns/op|4.624 ns/op|4.615 ns/op|21.70 ns/op|
-|struct [8]int64|23.10 ns/op|36.69 ns/op|4.714 ns/op|5.033 ns/op|49.09 ns/op|
-|int64|0.2708 ns/op|0.2416 ns/op|4.936 ns/op|5.150 ns/op|5.291 ns/op|
+|                 | Get without GC | Get with GC  | Set without GC | Set with GC | Concurrent Get and Set without GC |
+|-----------------|----------------|--------------|----------------|-------------|-----------------------------------|
+| struct int64    | 10.18 ns/op    | 11.42 ns/op  | 4.624 ns/op    | 4.615 ns/op | 21.70 ns/op                       |
+| struct [8]int64 | 23.10 ns/op    | 36.69 ns/op  | 4.714 ns/op    | 5.033 ns/op | 49.09 ns/op                       |
+| int64           | 0.2708 ns/op   | 0.2416 ns/op | 4.936 ns/op    | 5.150 ns/op | 5.291 ns/op                       |
 
 
 ### Get vs. Set
@@ -38,8 +43,8 @@ go tool pprof mem.prof => top
 // no significant difference in call graph, likely due to the I/O, lock, Timer, scheduling etc.
 
 ### Impact of malloc
-** mallocgc allocates 8 bytes cost around 8ns, compare to 20ns for 64 bytes
-** atomic.StoreInt64 doesn't call mallocgc, it is consistent 4ns
+** malloc allocates 8 bytes cost around 8ns, compare to 20ns for 64 bytes
+** atomic.StoreInt64 doesn't call malloc, it is consistent 4ns
 ** implementation using int64 directly, no allocation to heap, 0.2 ns only
 
 ### Impact of the struct