Cray CX1000 Betriebsanweisung Seite 28
- Seite / 169
- Inhaltsverzeichnis
- LESEZEICHEN
Bewertet. / 5. Basierend auf Kundenbewertungen
Cray XMT™ Programming Environment User’s Guide
Subnormal numbers are less precise than normalized numbers. The smallest
subnormal number, min_denorm, has only one significant bit while the largest has
52 significant bits. However, whenever 0.5 <= x/y <= 2.0, the difference x-
y is exact, even though it may have less precision than x and y. This is not true for
machines that flush underflow to zero.
The Cray XMT floating-point hardware handles gradual underflow transparently.
Unlike many systems, the Cray XMT is not slowed by the presence (or possibility) of
subnormal numbers and gradual underflow in a computation.
3.4.1 Differences from IEEE Floating-point Arithmetic
The Cray XMT processors do not have 32-bit floating-point instructions. If you
are performing an operation on 32-bit floating-point numbers, you must first use
the MTA_FLOAT_REAL intrinsic function to convert each 32-bit number in the
operation to a 64-bit number. After the operation is complete, you can use the
MTA_REAL_FLOAT intrinsic function to round the results to 32-bit numbers. This
double rounding (first to 64 bits and then to 32 bits) is not the same as a single
rounding to 32 bits. For more information about how to use MTA_FLOAT_REAL and
MTA_REAL_FLOAT, see the mta_intrinsics(3) man page.
The Cray XMT does not provide you with control over rounding precision for
floating-point operations. The level of rounding precision is set on the processor
during the manufacturing process.
Traps on the Cray XMT are precise, but operands can be overwritten by the results of
an operation performed on the same or a different functional unit. This can make the
implementation of post-substitution difficult.
There is no exponent wrapping when an operation enables or takes an overflow or
underflow trap. The intent of wrapping is to provide for automatic rescaling when
products or quotients are used in subsequent operations. On the Cray XMT, you
must use care when rescaling.
The hardware supports fused multiply-add operations that only require a single issue
of an instruction. This operation facilitates certain computations by making it easy to
extract the lower half of the product of two 64-bit doubles. The problem is that the
compiler can evaluate statements such as the following in several different ways, each
of which may produce a different result:
x = a*b + c*d;
The previous statement can be evaluated as either:
temp = a*b;
x = temp + c*d; // For multiply-add operation
Or
temp = c*d;
x = a*b + temp; // For multiply-add operation
28 S–2479–20
- S–2479–20 1
- Changes to this Document 5
- Compiler Overview [7] 75 8
- S–2479–20 9 9
- Glossary 167 10
- Procedures 10
- Examples 10
- Contents 11
- S–2479–20 11 11
- Introduction [1] 13
- 14 S–2479–20 14
- 2.1 Setting Up a Secure Shell 15
- 2.1.2 Additional Information 16
- 2.2 Using Modules 17
- 2.2.3 Module Commands 18
- Developing an Application [3] 19
- 3.2.1 Generic Functions 20
- S–2479–20 21 21
- 22 S–2479–20 22
- S–2479–20 23 23
- 3.2.2 Intrinsic Functions 24
- 3.3.2 Avoiding Deadlock 25
- 30 S–2479–20 30
- 3.5.2 Anonymous futures 34
- 3.7 File I/O 36
- 3.7.2 System-level I/O 39
- S–2479–20 43 43
- 44 S–2479–20 44
- 3.9 Debugging the Program 45
- 46 S–2479–20 46
- 48 S–2479–20 48
- 4.2 Persisting Shared Memory 49
- 52 S–2479–20 52
- S–2479–20 53 53
- 54 S–2479–20 54
- S–2479–20 55 55
- 56 S–2479–20 56
- 5.5 LUC Client/Server Example 60
- 5.6 Fast I/O Memory Usage 65
- 66 S–2479–20 66
- Library [6] 67
- a file at which to 68
- a file from which to 68
- in the file, a pointer 69
- 70 S–2479–20 70
- 6.4 Examples 71
- 74 S–2479–20 74
- Compiler Overview [7] 75
- 7.1 The Compilation Process 76
- S–2479–20 77 77
- Whole-program Compilation 78
- Separate-module Compilation 78
- S–2479–20 79 79
- 7.2 Invoking the Compiler 80
- 7.3 Setting the Compiler Mode 80
- 7.3.1 Whole-program Mode 81
- 7.3.2 Separate-module Mode 82
- 7.3.3 Mixed Mode 83
- 7.4 Inlining Functions 84
- S–2479–20 85 85
- 86 S–2479–20 86
- 7.7 Creating New Libraries 87
- 7.8 Compiler Messages 88
- #define NUMSTREAMS 40 89
- S–2479–20 89 89
- 90 S–2479–20 90
- Running an Application [8] 91
- 8.3 Improving Performance 93
- 94 S–2479–20 94
- Optional Optimizations [9] 95
- 100 S–2479–20 100
- Error Messages [A] 101
- S–2479–20 101 101
- 102 S–2479–20 102
- User Runtime Functions [B] 103
- 104 S–2479–20 104
- S–2479–20 105 105
- 106 S–2479–20 106
- S–2479–20 107 107
- 108 S–2479–20 108
- C.1 Compilation Directives 109
- 110 S–2479–20 110
- 114 S–2479–20 114
- S–2479–20 115 115
- S–2479–20 119 119
- S–2479–20 121 121
- 124 S–2479–20 124
- C.3 Semantic Assertions 125
- 126 S–2479–20 126
- C.4 Implementation Hints 130
- S–2479–20 131 131
- 132 S–2479–20 132
- Condition Codes [D] 133
- 134 S–2479–20 134
- S–2479–20 135 135
- 136 S–2479–20 136
- Data Types [E] 137
- 138 S–2479–20 138
- Keywords [F] 139
- 140 S–2479–20 140
- S–2479–20 141 141
- 142 S–2479–20 142
- MTA_PARAMS [G] 143
- 144 S–2479–20 144
- S–2479–20 145 145
- 146 S–2479–20 146
- LUC API Reference [H] 147
- H.3 LUC Methods 149
- H.3.2 stopService Method 150
- H.3.3 getMyEndpointID Method 150
- H.3.4 remoteCall Method 151
- 152 S–2479–20 152
- H.3.5 remoteCallSync Method 153
- 154 S–2479–20 154
- H.3.7 setConfigValue Method 155
- 156 S–2479–20 156
- S–2479–20 157 157
- H.3.8 getConfigValue Method 158
- H.4 LUC Type Definitions 159
- H.5 LUC Callback Functions 160
- S–2479–20 161 161
- H.6 LUC Return Codes 162
- S–2479–20 163 163
- 164 S–2479–20 164
- S–2479–20 165 165
- 166 S–2479–20 166
- Glossary 167
- 168 S–2479–20 168
- S–2479–20 169 169
New Holland T8030 manuály
Uživatelské manuály a uživatelské příručky pro Zahradnické vybavení New Holland T8030.
Poskytujeme 1 manuály pdf New Holland T8030 ke stažení zdarma podle typů dokumentů: Uživatelský manuál
Kommentare zu diesen Handbüchern