W11D1 & W14D2 : Jiang Ruohong

W11D1/w11d1note_jiangruohong.md

@@ -0,0 +1,29 @@
+## Out-of-order Execution
+### History
+1960s: Vector(or Array) vs Superscalar vs Pipeline+OoO  
+Scoreboarding(Thorton 1964) vs Tomasulo(IBM 1965)
+### How to WAR/WAW
+~~Scoreboarding~~  
+RAW: distance: forwarding/code movement  
+
+WAW/WAR(pseudo): renaming(tomasulo)  
+**Reservation Station**  
+- Op: operation in the unit
+- Vj, Vk: Value of source operands 
+- Qj,Qk: id of reservation stations producing source registers (0 means value is ready)
+
+### Challenges
+- branch
+- exception
+- process context switching(multiprocess OS)
+
+### Precise Interrupt
+Tomasulo: In-order issue, out-of-order execution, and out-of-order completion  
+"fix" the out-of-order completion  
+
+Reorder Buffer(ROB): all instructions commit in order  
+- Use ROB number instead of RS when execution completes
+- supplies operands between execution complete & commit
+- inst commit, when inst commits, put result into register
+
+easy to undo speculated inst/mispredicted branches/exceptions

W14D2/w14d2note_jiangruohong.md

@@ -0,0 +1,24 @@
+RAID 5
+
+Small writes (write to one disk): 
+- Option 1: read other data disks, create new sum and write to Parity 
+Disk
+- Option 2: since P has old sum, compare old data to new data, add 
+the difference to P
+• Small writes are limited by Parity Disk: Write to D0, D5 both also write to P disk
+
+![](w14d2note_jiangruohong_pic1.png)
+
+### Hamming code
+use $\log_2n$ extra parity bits  
+possible to find out wrong bit if one of the data bits flipped
+
+(7,4) Hamming code  
+![](w14d2note_jiangruohong_pic2.png)
+
+```
+bit 001 010 011 100 101 110 111
+    p1  p2  d1  p3  d2  d3  d4
+```
+
+XOR sum in every circle should be 0.