Intel Computer Hardware 80200 Manual de usuario Pagina 250
- Pagina / 289
- Tabla de contenidos
- MARCADORES
Valorado. / 5. Basado en revisión del cliente
B-22 March, 2003 Developer’s Manual
Intel
®
80200 Processor based on Intel
®
XScale
™
Microarchitecture
Optimization Guide
B.4.2.6. Data Alignment
Cache lines begin on 32-byte address boundaries. To maximize cache line use and minimize cache
pollution, data structures should be aligned on 32 byte boundaries and sized to multiple cache line
sizes. Aligning data structures on cache address boundaries simplifies later addition of prefetch
instructions to optimize performance.
Not aligning data on cache lines has the disadvantage of moving the prefetch address
correspondingly to the misalignment. Consider the following example:
struct {
long ia;
long ib;
long ic;
long id;
} tdata[IMAX];
for (i=0, i<IMAX; i++)
{
PREFETCH(tdata[i+1]);
tdata[i].ia = tdata[i].ib + tdata[i].ic _tdata[i].id];
....
tdata[i].id = 0;
}
In this case if tdata[] is not aligned to a cache line, then the prefetch using the address of
tdata[i+1].ia may not include element id. If the array was aligned on a cache line + 12 bytes, then
the prefetch would halve to be placed on &tdata[i+1].id.
If the structure is not sized to a multiple of the cache line size, then the prefetch address must be
advanced appropriately and requires extra prefetch instructions. Consider the following example:
struct {
long ia;
long ib;
long ic;
long id;
long ie;
} tdata[IMAX];
ADDRESS preadd = tdata
for (i=0, i<IMAX; i++)
{
PREFETCH(predata+=16);
tdata[I].ia = tdata[I].ib + tdata[I].ic _tdata[I].id] +
tdata[I].ie;
....
tdata[I].ie = 0;
}
In this case, the prefetch address was advanced by size of half a cache line and every other prefetch
instruction is ignored. Further, an additional register is required to track the next prefetch address.
Generally, not aligning and sizing data adds extra computational overhead.
Additional prefetch considerations are discussed in greater detail in following sections.
- 80200 Processor based on 1
- Microarchitecture 1
- Contents 3
- A Compatibility: Intel 9
- Introduction 17
- 1.1.2 Features 18
- 1.2.1 Number Representation 21
- 1.3 Other Relevant Documents 22
- Programming Model 23
- 2.2.5 Base Register Update 24
- 2.3.1 DSP Coprocessor 0 (CP0) 25
- 2.3.2 New Page Attributes 31
- 2.3.4 Event Architecture 34
- 2.3.4.3 Prefetch Aborts 35
- 2.3.4.4 Data Aborts 36
- Contents 38
- Memory Management 39
- 3.2 Architecture Model 40
- 3.2.3 Exceptions 42
- 3.4 Control 44
- 3.4.3 Locking Entries 45
- Instruction Cache 49
- 4.2 Operation 50
- 4.2.3 Fetch Policy 51
- 4.2.5 Parity Protection 52
- 4.3 Instruction Cache Control 54
- Branch Target Buffer 59
- 5.1.1 Reset 60
- 5.1.2 Update Policy 60
- 5.2 BTB Control 61
- Data Cache 63
- 6.2.3 Cache Policies 67
- 6.2.3.2 Read Miss Policy 68
- 6.2.3.3 Write Miss Policy 69
- 6.2.5 Parity Protection 70
- 6.2.6 Atomic Accesses 70
- 6.3.2 Enabling/Disabling 71
- Bits Description Notes 81
- 7.2 CP15 Registers 82
- 7.2.5 Register 4: Reserved 87
- • translation faults 89
- • domain faults 89
- • permission faults 89
- 7.3 CP14 Registers 98
- Configuration 100
- System Management 101
- 8.2 Processor Reset 103
- 8.2.2 Reset Effect on Outputs 104
- 8.3 Power Management 105
- Interrupts 107
- 9.3 Programmer Model 108
- 9.3.1 INTCTL 109
- 9.3.2 INTSRC 110
- 9.3.3 INTSTR 111
- External Bus 113
- 10.2 Signal Description 115
- 10.2.1 Request Bus 116
- 10.2.2 Data Bus 118
- 10.2.3 Critical Word First 119
- 10.2.4 Configuration Pins 120
- 10.2.5 Multimaster Support 121
- 10.2.6 Abort 123
- 10.2.7 ECC 124
- 10.3 Examples 126
- 10.3.4 Word Write 129
- 10.3.5.1 Write Burst 131
- 10.3.6 Write Burst, Coalesced 132
- 10.3.7 Pipelined Accesses 133
- 10.3.8 Locked Access 134
- 10.3.9 Aborted Access 135
- 10.3.10 Hold 136
- Bus Controller 137
- 11.3 Error Handling 138
- 11.3.2 ECC Errors 139
- 11.4 Programmer Model 141
- 11.4.2 ECC Error Registers 145
- Performance Monitoring 147
- 12.4.1 Managing PMNC 151
- 12.7 Examples 158
- Software Debug 159
- 13.3 Introduction 160
- 13.4.1 Global Enable Bit (GE) 162
- 13.4.2 Halt Mode Bit (H) 162
- 13.4.4 Sticky Abort Bit (SA) 163
- 13.5 Debug Exceptions 164
- 13.5.2 Monitor Mode 166
- 13.6 HW Breakpoint Resources 167
- 13.6.2 Data Breakpoints 168
- 13.7 Software Breakpoints 169
- 13.8.2 Overflow Flag (OV) 172
- 13.8.3 Download Flag (D) 172
- 13.9 Transmit Register (TX) 174
- 13.10 Receive Register (RX) 174
- 13.11 Debug JTAG Access 175
- 13.11.2 SELDCSR JTAG Register 176
- 13.11.2.1 DBG.HLD_RST 177
- 13.11.3 DBGTX JTAG Command 178
- 13.11.4 DBGTX JTAG Register 179
- 13.11.5 DBGRX JTAG Command 179
- 13.11.6 DBGRX JTAG Register 180
- 13.11.6.1 RX Write Logic 181
- 13.11.6.2 DBGRX Data Register 182
- 13.11.6.3 DBG.RR 182
- 13.11.6.4 DBG.V 183
- 13.11.6.5 DBG.RX 183
- 13.11.6.6 DBG.D 183
- 13.11.6.7 DBG.FLUSH 183
- 13.12 Trace Buffer 184
- 13.13 Trace Buffer Entries 186
- 13.13.1.3 Address Bytes 189
- 13.13.2 Trace Buffer Usage 190
- 13.14.1 LDIC JTAG Command 192
- 13.14.3 LDIC Cache Functions 194
- Debugger Actions 201
- Debug Handler Actions 201
- • a debug handler; 205
- • External memory 209
- 13.15.2.4 High-Speed Download 210
- • turn off all breakpoints; 211
- • invalidate the btb; 211
- Performance Considerations 213
- 14.2 Branch Prediction 214
- 14.3 Addressing Modes 214
- 14.4 Instruction Latencies 215
- • Minimum Resource Latency 216
- 14.4.8 Semaphore Instructions 221
- 14.4.11 Thumb* Instructions 221
- Compatibility: Intel 223
- 80200 Processor 223
- A.3 Architecture Deviations 225
- A.3.4 Write Buffer Behavior 226
- A.3.5 External Aborts 226
- A.3.6 Performance Differences 227
- Optimization Guide B 229
- B.2 Intel 230
- 80200 Processor Pipeline 230
- F1 F2 ID RF X1 X2 231
- M1 M2 Mx 231
- B.2.1.5. Use of Bypassing 232
- B.2.2.2. Pipeline Stalls 233
- B.2.3 Main Execution Pipeline 234
- Optimization Guide 235
- B.2.4 Memory Pipeline 236
- B.3 Basic Optimizations 237
- B.3.1.2. Optimizing Branches 238
- B.3.2 Bit Field Manipulation 241
- B.4.1 Instruction Cache 245
- • Interrupt handlers 246
- • Real time clock handlers 246
- • OS critical code 246
- B.4.2 Data and Mini Cache 247
- B.4.2.4. Creating On-chip RAM 248
- B.4.2.5. Mini-data Cache 249
- B.4.2.6. Data Alignment 250
- B.4.2.7. Literal Pools 251
- B.4.3 Cache Considerations 252
- B.4.4 Prefetch Considerations 253
- B.4.4.8. Cache Blocking 259
- B.4.4.9. Prefetch Unrolling 259
- B.4.4.10. Pointer Prefetch 260
- B.4.4.11. Loop Interchange 261
- B.4.4.12. Loop Fusion 261
- B.5 Instruction Scheduling 263
- B.6 Optimizing C Libraries 273
- B.7 Optimizations for Size 273
- Test Features C 275
- Test Features 276
- C.2.4.2. Bypass Register 280
- C.2.5 TAP Controller 281
- C.2.5.2. Run-Test/Idle State 282
- C.2.5.3. Select-DR-Scan State 282
- C.2.5.4. Capture-DR State 282
- C.2.5.5. Shift-DR State 283
- C.2.5.6. Exit1-DR State 283
- C.2.5.7. Pause-DR State 283
- C.2.5.8. Exit2-DR State 283
- C.2.5.9. Update-DR State 284
- C.2.5.11. Capture-IR State 284
- C.2.5.12. Shift-IR State 284
- C.2.5.13. Exit1-IR State 285
- C.2.5.14. Pause-IR State 285
- C.2.5.15. Exit2-IR State 285
- C.2.5.16. Update-IR State 285
Relacionado con productos y manuales para Tarjetas de desarrollo Intel Computer Hardware 80200
Tarjetas de desarrollo Intel IXC1100 Manual de usuario
(568 paginas)
(568 paginas)
© 2020, manymanuals.es. Todos los derechos reservados | 0.082 s |
Manymanuals.com
Manymanuals.de
Manymanuals.fr
Manymanuals.it
Manymanuals.pl
Manymanuals.cz
Manymanuals.es
Manymanuals-pt.com
Comentarios a estos manuales