1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
|
<!---======================= begin_copyright_notice ============================
Copyright (C) 2020-2021 Intel Corporation
SPDX-License-Identifier: MIT
============================= end_copyright_notice ==========================-->
## Opcode
SCATTER_SCALED = 0x79
## Format
| | | | | | | |
| --- | --- | --- | --- | --- | --- | --- |
| 0x79(SCATTER_SCALED) | Exec_size | Pred | Block_size | Num_blocks | Scale | Surface |
| | Offset | Element_offset | Src | | | |
## Semantics
for (i = 0; i < exec_size; ++i) {
if (ChEn[i]) {
surface[offset+element_offset[i]] = src[i]; //<num_blocks> bytes
}
}
## Description
Performs <exec_size> element scattered write into <surface> using the values from <src>. For each channel <num_blocks> bytes are written. The address to be written is determined by the offset and each element offset.
- **Exec_size(ub):** Execution size
- Bit[2..0]: size of the region for source and destination operands
- 0b000: 1 element (scalar)
- 0b001: 2 elements
- 0b010: 4 elements
- 0b011: 8 elements
- 0b100: 16 elements
- 0b101: 32 elements
- Bit[7..4]: execution mask (explicit control over the enabled channels)
- 0b0000: M1
- 0b0001: M2
- 0b0010: M3
- 0b0011: M4
- 0b0100: M5
- 0b0101: M6
- 0b0110: M7
- 0b0111: M8
- 0b1000: M1_NM
- 0b1001: M2_NM
- 0b1010: M3_NM
- 0b1011: M4_NM
- 0b1100: M5_NM
- 0b1101: M6_NM
- 0b1110: M7_NM
- 0b1111: M8_NM
- **Pred(uw):** Predication control
- **Block_size(ub):** This field is ignored and block size is always one byte
- **Num_blocks(ub):**
- Bit[1..0]: encodes the number of blocks to read at each address.
- 0b00: 1 block
- 0b01: 2 blocks
- 0b10: 4 blocks
- **Scale(uw):** This field is ignored and scale is always zero
- **Surface(ub):** Index of the surface variable. When scale is non-zero, only T0 (SLM) and T5 (stateless) are supported
- **Offset(scalar):** The global byte offset. Must have type UD
- **Element_offset(raw_operand):** The first Exec_size elements of the operand will be used as the byte offsets into the surface. Must have type UD
- **Src(raw_operand):** The variable providing the values to be written. The first exec_size elements will be used. For 1 and 2 byte accesses the upper bytes will be ignored. Must have type UD,D,F
#### Properties
- **Out-of-bound Access:** On write: data is dropped.
## Text
```
[(<P>)] SCATTER_SCALED.<num_blocks> (<exec_size>) <surface> <offset> <element_offset> <src>
```
## Notes
The behavior is undefined if more than one channel writes to the same address.
|
