Download presentation
Presentation is loading. Please wait.
1
Modern processor design
2
Engineering design
3
The dynamic-static interface (DSI)
4
Conceptuall illustration of possible placement of DSI in ISA design
5
Performance simulation metods
Trace-driven simulation Execution-driven simulation
6
Scalar pipeline machine
7
Superpipelined machine of degree m=3
8
Superpipelined MIPS R4000 8-stage pipeline
9
Superscalar machine of degree n=3
10
VLIW machine of degree n=3
11
Pipeline example 4-stage 11-stage
12
Two commercial instruction pipelines
13
Activity of pipeline stages
14
6-stage instruction pipeline
15
I-cache and D-cache
16
Access to RF
17
RAW, WAR, and WAW data dependencies
18
WAW, WAR, and RAW hazards
19
Forwarding paths
20
Forwarding paths for ALU leading instructions
21
Forwarding paths for Load leading instructions
22
Impact on ALU, Load, and Branch penalties with increasing pipeline depth
23
Mitigating the Branch penalty impact of deep pipelines
24
Direct, associative, and set-associative caches
25
Strategy of cache design
26
Memory hierarchy
27
Main memory and I/O
28
DRAM accesses
29
Virtual to physical address translation
30
Processes switching
31
Paging tables
32
Direct cache
33
Associative cache
34
Set-associative cache
35
Virtual to physical address translation
36
Disk
37
I/O device communication
38
Stall cycle induced by backward propagation of stalling
39
Machine parallelism
40
Parallel pipeline of width s=3
41
Parallel pipeline - examples
5-stage Pentium parallel pipeline s=2 5-stage i486
42
Parallel pipeline with four execution pipes
43
The Motorola 88110 superscalar microprocessor
44
Interpipeline-stage buffers multi-entry buffer with reordering
single-entry buffer multi-entry buffer multi-entry buffer with reordering
45
Dynamic pipeline of width s=3
46
6-stage superscalar pipeline
47
Instruction dispatching in superscalar pipeline
48
Centralized reservation station
49
Distributed reservation stations
50
Dynamic pipeline with reservation stations and reorder buffer
51
Disruption of sequential control flow by Branch instructions
52
Branch target address generation penalties
53
Branch condition resolution penalties
54
Alpha pipeline stages
55
HP PA 7100 pipeline stages
56
IBM POWER (RIOS) pipeline stages
57
Intel i960CA pipeline stages
58
Intel Pentium pipeline stages
59
Cyrix 6x86 pipeline stages
60
Intel P6 pipeline stages
61
MIPS R10000 pipeline stages
62
Motorola MC68060 pipeline stages
63
IBM/Motorola PowerPC 604 pipeline stages
64
Sun UltraSPARC-I pipeline stages
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.