The AMDGPU backend provides ISA code generation for AMD GPUs, starting with the R600 family up until the current GCN families. It lives in the lib/Target/AMDGPU directory.
Use the clang -target <Architecture>-<Vendor>-<OS>-<Environment> option to specify the target triple:
AMDGPU Target Triples Architecture Vendor OS Environment r600 amd <empty> <empty> amdgcn amd <empty> <empty> amdgcn amd amdhsa <empty> amdgcn amd amdhsa opencl amdgcn amd amdhsa amdgizcl amdgcn amd amdhsa amdgiz amdgcn amd amdhsa hcc
Use the clang -mcpu <Processor> option to specify the AMD GPU processor. The names from both the Processor and Alternative Processor can be used.
AMDGPU Processors Processor Alternative Processor Target Triple Architecture dGPU/ APU Runtime Support Example Products R600 [AMD-R6xx] r600 r600 dGPU r630 r600 dGPU rs880 r600 dGPU rv670 r600 dGPU R700 [AMD-R7xx] rv710 r600 dGPU rv730 r600 dGPU rv770 r600 dGPU Evergreen [AMD-Evergreen] cedar r600 dGPU redwood r600 dGPU sumo r600 dGPU juniper r600 dGPU cypress r600 dGPU Northern Islands [AMD-Cayman-Trinity] barts r600 dGPU turks r600 dGPU caicos r600 dGPU cayman r600 dGPU GCN GFX6 (Southern Islands (SI)) [AMD-Souther-Islands] gfx600
- SI
- tahiti
amdgcn dGPU gfx601
- pitcairn
- verde
- oland
- hainan
amdgcn dGPU GCN GFX7 (Sea Islands (CI)) [AMD-Sea-Islands] gfx700
- bonaire
amdgcn dGPU
- Radeon HD 7790
- Radeon HD 8770
- R7 260
- R7 260X
- kaveri
amdgcn APU
- A6-7000
- A6 Pro-7050B
- A8-7100
- A8 Pro-7150B
- A10-7300
- A10 Pro-7350B
- FX-7500
- A8-7200P
- A10-7400P
- FX-7600P
gfx701
- hawaii
amdgcn dGPU ROCm
- FirePro W8100
- FirePro W9100
- FirePro S9150
- FirePro S9170
gfx702 dGPU ROCm
- Radeon R9 290
- Radeon R9 290x
- Radeon R390
- Radeon R390x
gfx703
- kabini
- mullins
amdgcn APU
- E1-2100
- E1-2200
- E1-2500
- E2-3000
- E2-3800
- A4-5000
- A4-5100
- A6-5200
- A4 Pro-3340B
GCN GFX8 (Volcanic Islands (VI)) [AMD-Volcanic-Islands] gfx800
- iceland
amdgcn dGPU
- FirePro S7150
- FirePro S7100
- FirePro W7100
- Radeon R285
- Radeon R9 380
- Radeon R9 385
- Mobile FirePro M7170
gfx801
- carrizo
amdgcn APU
- A6-8500P
- Pro A6-8500B
- A8-8600P
- Pro A8-8600B
- FX-8800P
- Pro A12-8800B
amdgcn APU ROCm
- A10-8700P
- Pro A10-8700B
- A10-8780P
amdgcn APU
- A10-9600P
- A10-9630P
- A12-9700P
- A12-9730P
- FX-9800P
- FX-9830P
amdgcn APU
- E2-9010
- A6-9210
- A9-9410
gfx802
- tonga
amdgcn dGPU ROCm Same as gfx800 gfx803
- fiji
amdgcn dGPU ROCm
- Radeon R9 Nano
- Radeon R9 Fury
- Radeon R9 FuryX
- Radeon Pro Duo
- FirePro S9300x2
- polaris10
amdgcn dGPU ROCm
- Radeon RX 470
- Radeon RX 480
- polaris11
amdgcn dGPU ROCm
- Radeon RX 460
gfx804 amdgcn dGPU Same as gfx803 gfx810
- stoney
amdgcn APU GCN GFX9 gfx900 amdgcn dGPU
- Radeon Vega Frontier Edition
gfx901 amdgcn dGPU ROCm Same as gfx900 except XNACK is enabled gfx902 amdgcn APU TBA gfx903 amdgcn APU Same as gfx902 except XNACK is enabled
The AMDGPU backend uses the following address space mappings.
The memory space names used in the table, aside from the region memory space, is from the OpenCL standard.
LLVM Address Space number is used throughout LLVM (for example, in LLVM IR).
Address Space Mapping LLVM Address Space Memory Space Current Default amdgiz/amdgizcl hcc Future Default 0 Private (Scratch) Generic (Flat) Generic (Flat) Generic (Flat) 1 Global Global Global Global 2 Constant Constant Constant Region (GDS) 3 Local (group/LDS) Local (group/LDS) Local (group/LDS) Local (group/LDS) 4 Generic (Flat) Region (GDS) Region (GDS) Constant 5 Region (GDS) Private (Scratch) Private (Scratch) Private (Scratch)
This section provides LLVM memory synchronization scopes supported by the AMDGPU backend memory model when the target triple OS is amdhsa (see Memory Model and Target Triples).
The memory model supported is based on the HSA memory model [HSA] which is based in turn on HRF-indirect with scope inclusion [HRF]. The happens-before relation is transitive over the synchonizes-with relation independent of scope, and synchonizes-with allows the memory scope instances to be inclusive (see table AMDHSA LLVM Sync Scopes for AMDHSA).
This is different to the OpenCL [OpenCL] memory model which does not have scope inclusion and requires the memory scopes to exactly match. However, this is conservatively correct for OpenCL.
AMDHSA LLVM Sync Scopes for AMDHSA LLVM Sync Scope Description none The default: system.
Synchronizes with, and participates in modification and seq_cst total orderings with, other operations (except image operations) for all address spaces (except private, or generic that accesses private) provided the other operation’s sync scope is:
- system.
- agent and executed by a thread on the same agent.
- workgroup and executed by a thread in the same workgroup.
- wavefront and executed by a thread in the same wavefront.
agent Synchronizes with, and participates in modification and seq_cst total orderings with, other operations (except image operations) for all address spaces (except private, or generic that accesses private) provided the other operation’s sync scope is:
- system or agent and executed by a thread on the same agent.
- workgroup and executed by a thread in the same workgroup.
- wavefront and executed by a thread in the same wavefront.
workgroup Synchronizes with, and participates in modification and seq_cst total orderings with, other operations (except image operations) for all address spaces (except private, or generic that accesses private) provided the other operation’s sync scope is:
- system, agent or workgroup and executed by a thread in the same workgroup.
- wavefront and executed by a thread in the same wavefront.
wavefront Synchronizes with, and participates in modification and seq_cst total orderings with, other operations (except image operations) for all address spaces (except private, or generic that accesses private) provided the other operation’s sync scope is:
- system, agent, workgroup or wavefront and executed by a thread in the same wavefront.
singlethread Only synchronizes with, and participates in modification and seq_cst total orderings with, other operations (except image operations) running in the same thread for all address spaces (for example, in signal handlers).
The AMDGPU backend generates a standard ELF [ELF] relocatable code object that can be linked by lld to produce a standard ELF shared code object which can be loaded and executed on an AMDGPU target.
The AMDGPU backend uses the following ELF header:
AMDGPU ELF Header Field Value e_ident[EI_CLASS] ELFCLASS64 e_ident[EI_DATA] ELFDATA2LSB e_ident[EI_OSABI] ELFOSABI_AMDGPU_HSA e_ident[EI_ABIVERSION] ELFABIVERSION_AMDGPU_HSA e_type ET_REL or ET_DYN e_machine EM_AMDGPU e_entry 0 e_flags 0
AMDGPU ELF Header Enumeration Values Name Value EM_AMDGPU 224 ELFOSABI_AMDGPU_HSA 64 ELFABIVERSION_AMDGPU_HSA 1
Can be one of the following values:
The AMD HSA runtime loader requires a ET_DYN code object.
An AMDGPU target ELF code object has the standard ELF sections which include:
AMDGPU ELF Sections Name Type Attributes .bss SHT_NOBITS SHF_ALLOC + SHF_WRITE .data SHT_PROGBITS SHF_ALLOC + SHF_WRITE .debug_* SHT_PROGBITS none .dynamic SHT_DYNAMIC SHF_ALLOC .dynstr SHT_PROGBITS SHF_ALLOC .dynsym SHT_PROGBITS SHF_ALLOC .got SHT_PROGBITS SHF_ALLOC + SHF_WRITE .hash SHT_HASH SHF_ALLOC .note SHT_NOTE none .relaname SHT_RELA none .rela.dyn SHT_RELA none .rodata SHT_PROGBITS SHF_ALLOC .shstrtab SHT_STRTAB none .strtab SHT_STRTAB none .symtab SHT_SYMTAB none .text SHT_PROGBITS SHF_ALLOC + SHF_EXECINSTR
These sections have their standard meanings (see [ELF]) and are only generated if needed.
For relocatable code objects, name is the name of the section that the relocation records apply. For example, .rela.text is the section name for relocation records associated with the .text section.
For linked shared code objects, .rela.dyn contains all the relocation records from each of the relocatable code object’s .relaname sections.
See Relocation Records for the relocation records supported by the AMDGPU backend.
As required by ELFCLASS64, minimal zero byte padding must be generated after the name field to ensure the desc field is 4 byte aligned. In addition, minimal zero byte padding must be generated to ensure the desc field size is a multiple of 4 bytes. The sh_addralign field of the .note section must be at least 4 to indicate at least 8 byte alignment.
The AMDGPU backend code object uses the following ELF note records in the .note section. The Description column specifies the layout of the note record’s desc field. All fields are consecutive bytes. Note records with variable size strings have a corresponding *_size field that specifies the number of bytes, including the terminating null character, in the string. The string(s) come immediately after the preceding fields.
Additional note records can be present.
AMDGPU ELF Note Records Name Type Description “AMD” NT_AMD_AMDGPU_METADATA <metadata null terminated string> “AMD” NT_AMD_AMDGPU_ISA <isa name null terminated string>
AMDGPU ELF Note Record Enumeration Values Name Value reserved 0-9 NT_AMD_AMDGPU_METADATA 10 NT_AMD_AMDGPU_ISA 11
Specifies the instruction set architecture used by the machine code contained in the code object.
This note record is required for code objects containing machine code for processors matching the amdgcn architecture in table Processors.
The null terminated string has the following syntax:
architecture-vendor-os-environment-processor
where:
- architecture
The architecture from table AMDGPU Target Triples.
This is always amdgcn when the target triple OS is amdhsa (see Target Triples).
- vendor
The vendor from table AMDGPU Target Triples.
For the AMDGPU backend this is always amd.
- os
- The OS from table AMDGPU Target Triples.
- environment
An environment from table AMDGPU Target Triples, or blank if the environment has no affect on the execution of the code object.
For the AMDGPU backend this is currently always blank.
- processor
- The processor from table AMDGPU Processors.
For example:
amdgcn-amd-amdhsa--gfx901
Specifies extensible metadata associated with the code object. See Code Object Metadata for the syntax of the code object metadata string.
This note record is required and must contain the minimum information necessary to support the ROCM kernel queries. For example, the segment sizes needed in a dispatch packet. In addition, a high level language runtime may require other information to be included. For example, the AMD OpenCL runtime records kernel argument information.
The code object metadata is specified by the NT_AMD_AMDHSA_METADATA note record (see Note Records).
The metadata is specified as a YAML formatted string (see [YAML] and YAML I/O).
The metadata is represented as a single YAML document comprised of the mapping defined in table AMDHSA Code Object Metadata Mapping and referenced tables.
For boolean values, the string values of false and true are used for false and true respectively.
Additional information can be added to the mappings. To avoid conflicts, any non-AMD key names should be prefixed by “vendor-name.”.
AMDHSA Code Object Metadata Mapping String Key Value Type Required? Description “Version” sequence of 2 integers Required
- The first integer is the major version. Currently 1.
- The second integer is the minor version. Currently 0.
“Printf” sequence of strings Each string is encoded information about a printf function call. The encoded information is organized as fields separated by colon (‘:’):
ID:N:S[0]:S[1]:...:S[N-1]:FormatString
where:
- ID
- A 32 bit integer as a unique id for each printf function call
- N
- A 32 bit integer equal to the number of arguments of printf function call minus 1
- S[i] (where i = 0, 1, ... , N-1)
- 32 bit integers for the size in bytes of the i-th FormatString argument of the printf function call
- FormatString
- The format string passed to the printf function call.
“Kernels” sequence of mapping Required Sequence of the mappings for each kernel in the code object. See AMDHSA Code Object Kernel Metadata Mapping for the definition of the mapping.
AMDHSA Code Object Kernel Metadata Mapping String Key Value Type Required? Description “Name” string Required Source name of the kernel. “SymbolName” string Required Name of the kernel descriptor ELF symbol. “Language” string Source language of the kernel. Values include:
- “OpenCL C”
- “OpenCL C++”
- “HCC”
- “OpenMP”
“LanguageVersion” sequence of 2 integers
- The first integer is the major version.
- The second integer is the minor version.
“Attrs” mapping Mapping of kernel attributes. See AMDHSA Code Object Kernel Attribute Metadata Mapping for the mapping definition. “Arguments” sequence of mapping Sequence of mappings of the kernel arguments. See AMDHSA Code Object Kernel Argument Metadata Mapping for the definition of the mapping. “CodeProps” mapping Mapping of properties related to the kernel code. See AMDHSA Code Object Kernel Code Properties Metadata Mapping for the mapping definition. “DebugProps” mapping Mapping of properties related to the kernel debugging. See AMDHSA Code Object Kernel Debug Properties Metadata Mapping for the mapping definition.
AMDHSA Code Object Kernel Attribute Metadata Mapping String Key Value Type Required? Description “ReqdWorkGroupSize” sequence of 3 integers The dispatch work-group size X, Y, Z must correspond to the specified values.
Corresponds to the OpenCL reqd_work_group_size attribute.
“WorkGroupSizeHint” sequence of 3 integers The dispatch work-group size X, Y, Z is likely to be the specified values.
Corresponds to the OpenCL work_group_size_hint attribute.
“VecTypeHint” string The name of a scalar or vector type.
Corresponds to the OpenCL vec_type_hint attribute.
AMDHSA Code Object Kernel Argument Metadata Mapping String Key Value Type Required? Description “Name” string Kernel argument name. “TypeName” string Kernel argument type name. “Size” integer Required Kernel argument size in bytes. “Align” integer Required Kernel argument alignment in bytes. Must be a power of two. “ValueKind” string Required Kernel argument kind that specifies how to set up the corresponding argument. Values include:
- “ByValue”
- The argument is copied directly into the kernarg.
- “GlobalBuffer”
- A global address space pointer to the buffer data is passed in the kernarg.
- “DynamicSharedPointer”
- A group address space pointer to dynamically allocated LDS is passed in the kernarg.
- “Sampler”
- A global address space pointer to a S# is passed in the kernarg.
- “Image”
- A global address space pointer to a T# is passed in the kernarg.
- “Pipe”
- A global address space pointer to an OpenCL pipe is passed in the kernarg.
- “Queue”
- A global address space pointer to an OpenCL device enqueue queue is passed in the kernarg.
- “HiddenGlobalOffsetX”
- The OpenCL grid dispatch global offset for the X dimension is passed in the kernarg.
- “HiddenGlobalOffsetY”
- The OpenCL grid dispatch global offset for the Y dimension is passed in the kernarg.
- “HiddenGlobalOffsetZ”
- The OpenCL grid dispatch global offset for the Z dimension is passed in the kernarg.
- “HiddenNone”
- An argument that is not used by the kernel. Space needs to be left for it, but it does not need to be set up.
- “HiddenPrintfBuffer”
- A global address space pointer to the runtime printf buffer is passed in kernarg.
- “HiddenDefaultQueue”
- A global address space pointer to the OpenCL device enqueue queue that should be used by the kernel by default is passed in the kernarg.
- “HiddenCompletionAction”
- TBD
“ValueType” string Required Kernel argument value type. Only present if “ValueKind” is “ByValue”. For vector data types, the value is for the element type. Values include:
- “Struct”
- “I8”
- “U8”
- “I16”
- “U16”
- “F16”
- “I32”
- “U32”
- “F32”
- “I64”
- “U64”
- “F64”
“PointeeAlign” integer Alignment in bytes of pointee type for pointer type kernel argument. Must be a power of 2. Only present if “ValueKind” is “DynamicSharedPointer”. “AddrSpaceQual” string Kernel argument address space qualifier. Only present if “ValueKind” is “GlobalBuffer” or “DynamicSharedPointer”. Values are:
- “Private”
- “Global”
- “Constant”
- “Local”
- “Generic”
- “Region”
“AccQual” string Kernel argument access qualifier. Only present if “ValueKind” is “Image” or “Pipe”. Values are:
- “ReadOnly”
- “WriteOnly”
- “ReadWrite”
“ActualAcc” string The actual memory accesses performed by the kernel on the kernel argument. Only present if “ValueKind” is “GlobalBuffer”, “Image”, or “Pipe”. This may be more restrictive than indicated by “AccQual” to reflect what the kernel actual does. If not present then the runtime must assume what is implied by “AccQual” and “IsConst”. Values are:
- “ReadOnly”
- “WriteOnly”
- “ReadWrite”
“IsConst” boolean Indicates if the kernel argument is const qualified. Only present if “ValueKind” is “GlobalBuffer”. “IsRestrict” boolean Indicates if the kernel argument is restrict qualified. Only present if “ValueKind” is “GlobalBuffer”. “IsVolatile” boolean Indicates if the kernel argument is volatile qualified. Only present if “ValueKind” is “GlobalBuffer”. “IsPipe” boolean Indicates if the kernel argument is pipe qualified. Only present if “ValueKind” is “Pipe”.
AMDHSA Code Object Kernel Code Properties Metadata Mapping String Key Value Type Required? Description “KernargSegmentSize” integer Required The size in bytes of the kernarg segment that holds the values of the arguments to the kernel. “GroupSegmentFixedSize” integer Required The amount of group segment memory required by a work-group in bytes. This does not include any dynamically allocated group segment memory that may be added when the kernel is dispatched. “PrivateSegmentFixedSize” integer Required The amount of fixed private address space memory required for a work-item in bytes. If IsDynamicCallstack is 1 then additional space must be added to this value for the call stack. “KernargSegmentAlign” integer Required The maximum byte alignment of arguments in the kernarg segment. Must be a power of 2. “WavefrontSize” integer Required Wavefront size. Must be a power of 2. “NumSGPRs” integer Number of scalar registers used by a wavefront for GFX6-GFX9. This includes the special SGPRs for VCC, Flat Scratch (GFX7-GFX9) and XNACK (for GFX8-GFX9). It does not include the 16 SGPR added if a trap handler is enabled. It is not rounded up to the allocation granularity. “NumVGPRs” integer Number of vector registers used by each work-item for GFX6-GFX9 “MaxFlatWorkgroupSize” integer Maximum flat work-group size supported by the kernel in work-items. “IsDynamicCallStack” boolean Indicates if the generated machine code is using a dynamically sized call stack. “IsXNACKEnabled” boolean Indicates if the generated machine code is capable of supporting XNACK.
AMDHSA Code Object Kernel Debug Properties Metadata Mapping String Key Value Type Required? Description “DebuggerABIVersion” string “ReservedNumVGPRs” integer “ReservedFirstVGPR” integer “PrivateSegmentBufferSGPR” integer “WavefrontPrivateSegmentOffsetSGPR” integer
Symbols include the following:
AMDGPU ELF Symbols Name Type Section Description link-name STT_OBJECT
- .data
- .rodata
- .bss
Global variable link-name@kd STT_OBJECT
- .rodata
Kernel descriptor link-name STT_FUNC
- .text
Kernel entry point
Global variables both used and defined by the compilation unit.
If the symbol is defined in the compilation unit then it is allocated in the appropriate section according to if it has initialized data or is readonly.
If the symbol is external then its section is STN_UNDEF and the loader will resolve relocations using the definition provided by another code object or explicitly defined by the runtime.
All global symbols, whether defined in the compilation unit or external, are accessed by the machine code indirectly through a GOT table entry. This allows them to be preemptable. The GOT table is only supported when the target triple OS is amdhsa (see Target Triples).
AMDGPU backend generates Elf64_Rela relocation records. Supported relocatable fields are:
Following notations are used for specifying relocation calculations:
The following relocation types are supported:
AMDGPU ELF Relocation Records Relocation Type Value Field Calculation R_AMDGPU_NONE 0 none none R_AMDGPU_ABS32_LO 1 word32 (S + A) & 0xFFFFFFFF R_AMDGPU_ABS32_HI 2 word32 (S + A) >> 32 R_AMDGPU_ABS64 3 word64 S + A R_AMDGPU_REL32 4 word32 S + A - P R_AMDGPU_REL64 5 word64 S + A - P R_AMDGPU_ABS32 6 word32 S + A R_AMDGPU_GOTPCREL 7 word32 G + GOT + A - P R_AMDGPU_GOTPCREL32_LO 8 word32 (G + GOT + A - P) & 0xFFFFFFFF R_AMDGPU_GOTPCREL32_HI 9 word32 (G + GOT + A - P) >> 32 R_AMDGPU_REL32_LO 10 word32 (S + A - P) & 0xFFFFFFFF R_AMDGPU_REL32_HI 11 word32 (S + A - P) >> 32
Standard DWARF [DWARF] Version 2 sections can be generated. These contain information that maps the code object executable code and data to the source language constructs. It can be used by tools such as debuggers and profilers.
The following address space mapping is used:
AMDGPU DWARF Address Space Mapping DWARF Address Space Memory Space 1 Private (Scratch) 2 Local (group/LDS) omitted Global omitted Constant omitted Generic (Flat) not supported Region (GDS)
See Address Spaces for infomration on the memory space terminology used in the table.
An address_class attribute is generated on pointer type DIEs to specify the DWARF address space of the value of the pointer when it is in the private or local address space. Otherwise the attribute is omitted.
An XDEREF operation is generated in location list expressions for variables that are allocated in the private and local address space. Otherwise no XDREF is omitted.
This section is WIP.
This section is WIP.
This section provides code conventions used when the target triple OS is amdhsa (see Target Triples).
The HSA architected queuing language (AQL) defines a user space memory interface that can be used to control the dispatch of kernels, in an agent independent way. An agent can have zero or more AQL queues created for it using the ROCm runtime, in which AQL packets (all of which are 64 bytes) can be placed. See the HSA Platform System Architecture Specification [HSA] for the AQL queue mechanics and packet layouts.
The packet processor of a kernel agent is responsible for detecting and dispatching HSA kernels from the AQL queues associated with it. For AMD GPUs the packet processor is implemented by the hardware command processor (CP), asynchronous dispatch controller (ADC) and shader processor input controller (SPI).
The ROCm runtime can be used to allocate an AQL queue object. It uses the kernel mode driver to initialize and register the AQL queue with CP.
To dispatch a kernel the following actions are performed. This can occur in the CPU host program, or from an HSA kernel executing on a GPU.
The memory space properties are:
AMDHSA Memory Spaces Memory Space Name HSA Segment Name Hardware Name Address Size NULL Value Private private scratch 32 0x00000000 Local group LDS 32 0xFFFFFFFF Global global global 64 0x0000000000000000 Constant constant same as global 64 0x0000000000000000 Generic flat flat 64 0x0000000000000000 Region N/A GDS 32 not implemented for AMDHSA
The global and constant memory spaces both use global virtual addresses, which are the same virtual address space used by the CPU. However, some virtual addresses may only be accessible to the CPU, some only accessible by the GPU, and some by both.
Using the constant memory space indicates that the data will not change during the execution of the kernel. This allows scalar read instructions to be used. The vector and scalar L1 caches are invalidated of volatile data before each kernel dispatch execution to allow constant memory to change values between kernel dispatches.
The local memory space uses the hardware Local Data Store (LDS) which is automatically allocated when the hardware creates work-groups of wavefronts, and freed when all the wavefronts of a work-group have terminated. The data store (DS) instructions can be used to access it.
The private memory space uses the hardware scratch memory support. If the kernel uses scratch, then the hardware allocates memory that is accessed using wavefront lane dword (4 byte) interleaving. The mapping used from private address to physical address is:
wavefront-scratch-base + (private-address * wavefront-size * 4) + (wavefront-lane-id * 4)
There are different ways that the wavefront scratch base address is determined by a wavefront (see Initial Kernel Execution State). This memory can be accessed in an interleaved manner using buffer instruction with the scratch buffer descriptor and per wave scratch offset, by the scratch instructions, or by flat instructions. If each lane of a wavefront accesses the same private address, the interleaving results in adjacent dwords being accessed and hence requires fewer cache lines to be fetched. Multi-dword access is not supported except by flat and scratch instructions in GFX9.
The generic address space uses the hardware flat address support available in GFX7-GFX9. This uses two fixed ranges of virtual addresses (the private and local appertures), that are outside the range of addressible global memory, to map from a flat address to a private or local address.
FLAT instructions can take a flat address and access global, private (scratch) and group (LDS) memory depending in if the address is within one of the apperture ranges. Flat access to scratch requires hardware aperture setup and setup in the kernel prologue (see Flat Scratch). Flat access to LDS requires hardware aperture setup and M0 (GFX7-GFX8) register setup (see M0).
To convert between a segment address and a flat address the base address of the appertures address can be used. For GFX7-GFX8 these are available in the HSA AQL Queue the address of which can be obtained with Queue Ptr SGPR (see Initial Kernel Execution State). For GFX9 the appature base addresses are directly available as inline constant registers SRC_SHARED_BASE/LIMIT and SRC_PRIVATE_BASE/LIMIT. In 64 bit address mode the apperture sizes are 2^32 bytes and the base is aligned to 2^32 which makes it easier to convert from flat to segment or segment to flat.
Image and sample handles created by the ROCm runtime are 64 bit addresses of a hardware 32 byte V# and 48 byte S# object respectively. In order to support the HSA query_sampler operations two extra dwords are used to store the HSA BRIG enumeration values for the queries that are not trivially deducible from the S# representation.
Signal handles created by the ROCm runtime are 64 bit addresses of a structure allocated in memory accessible from both the CPU and GPU. The structure is defined by the ROCm runtime and subject to change between releases (see [AMD-ROCm-github]).
The AQL queue structure is defined by the ROCm runtime and subject to change between releases (see [AMD-ROCm-github]). For some processors it contains fields needed to implement certain language features such as the flat address aperture bases. It also contains fields used by CP such as managing the allocation of scratch memory.
A kernel descriptor consists of the information needed by CP to initiate the execution of a kernel, including the entry point address of the machine code that implements the kernel.
CP microcode requires the Kernel descritor to be allocated on 64 byte alignment.
Kernel Descriptor for GFX6-GFX9 Bits Size Field Name Description 31:0 4 bytes group_segment_fixed_size The amount of fixed local address space memory required for a work-group in bytes. This does not include any dynamically allocated local address space memory that may be added when the kernel is dispatched. 63:32 4 bytes private_segment_fixed_size The amount of fixed private address space memory required for a work-item in bytes. If is_dynamic_callstack is 1 then additional space must be added to this value for the call stack. 95:64 4 bytes max_flat_workgroup_size Maximum flat work-group size supported by the kernel in work-items. 96 1 bit is_dynamic_call_stack Indicates if the generated machine code is using a dynamically sized call stack. 97 1 bit is_xnack_enabled Indicates if the generated machine code is capable of suppoting XNACK. 127:98 30 bits Reserved. Must be 0. 191:128 8 bytes kernel_code_entry_byte_offset Byte offset (possibly negative) from base address of kernel descriptor to kernel’s entry point instruction which must be 256 byte aligned. 383:192 24 bytes Reserved. Must be 0. 415:384 4 bytes compute_pgm_rsrc1 Compute Shader (CS) program settings used by CP to set up COMPUTE_PGM_RSRC1 configuration register. See compute_pgm_rsrc1 for GFX6-GFX9. 447:416 4 bytes compute_pgm_rsrc2 Compute Shader (CS) program settings used by CP to set up COMPUTE_PGM_RSRC2 configuration register. See compute_pgm_rsrc2 for GFX6-GFX9. 448 1 bit enable_sgpr_private_segment _buffer Enable the setup of the SGPR user data registers (see Initial Kernel Execution State).
The total number of SGPR user data registers requested must not exceed 16 and match value in compute_pgm_rsrc2.user_sgpr.user_sgpr_count. Any requests beyond 16 will be ignored.
449 1 bit enable_sgpr_dispatch_ptr see above 450 1 bit enable_sgpr_queue_ptr see above 451 1 bit enable_sgpr_kernarg_segment_ptr see above 452 1 bit enable_sgpr_dispatch_id see above 453 1 bit enable_sgpr_flat_scratch_init see above 454 1 bit enable_sgpr_private_segment _size see above 455 1 bit enable_sgpr_grid_workgroup _count_X Not implemented in CP and should always be 0. 456 1 bit enable_sgpr_grid_workgroup _count_Y Not implemented in CP and should always be 0. 457 1 bit enable_sgpr_grid_workgroup _count_Z Not implemented in CP and should always be 0. 463:458 6 bits Reserved. Must be 0. 511:464 4 bytes Reserved. Must be 0. 512 Total size 64 bytes.
compute_pgm_rsrc1 for GFX6-GFX9 Bits Size Field Name Description 5:0 6 bits granulated_workitem_vgpr_count Number of vector registers used by each work-item, granularity is device specific:
- GFX6-9
- roundup((max-vgpg + 1) / 4) - 1
Used by CP to set up COMPUTE_PGM_RSRC1.VGPRS.
9:6 4 bits granulated_wavefront_sgpr_count Number of scalar registers used by a wavefront, granularity is device specific:
- GFX6-8
- roundup((max-sgpg + 1) / 8) - 1
- GFX9
- roundup((max-sgpg + 1) / 16) - 1
Includes the special SGPRs for VCC, Flat Scratch (for GFX7 onwards) and XNACK (for GFX8 onwards). It does not include the 16 SGPR added if a trap handler is enabled.
Used by CP to set up COMPUTE_PGM_RSRC1.SGPRS.
11:10 2 bits priority Must be 0.
Start executing wavefront at the specified priority.
CP is responsible for filling in COMPUTE_PGM_RSRC1.PRIORITY.
13:12 2 bits float_mode_round_32 Wavefront starts execution with specified rounding mode for single (32 bit) floating point precision floating point operations.
Floating point rounding mode values are defined in Floating Point Rounding Mode Enumeration Values.
Used by CP to set up COMPUTE_PGM_RSRC1.FLOAT_MODE.
15:14 2 bits float_mode_round_16_64 Wavefront starts execution with specified rounding denorm mode for half/double (16 and 64 bit) floating point precision floating point operations.
Floating point rounding mode values are defined in Floating Point Rounding Mode Enumeration Values.
Used by CP to set up COMPUTE_PGM_RSRC1.FLOAT_MODE.
17:16 2 bits float_mode_denorm_32 Wavefront starts execution with specified denorm mode for single (32 bit) floating point precision floating point operations.
Floating point denorm mode values are defined in Floating Point Denorm Mode Enumeration Values.
Used by CP to set up COMPUTE_PGM_RSRC1.FLOAT_MODE.
19:18 2 bits float_mode_denorm_16_64 Wavefront starts execution with specified denorm mode for half/double (16 and 64 bit) floating point precision floating point operations.
Floating point denorm mode values are defined in Floating Point Denorm Mode Enumeration Values.
Used by CP to set up COMPUTE_PGM_RSRC1.FLOAT_MODE.
20 1 bit priv Must be 0.
Start executing wavefront in privilege trap handler mode.
CP is responsible for filling in COMPUTE_PGM_RSRC1.PRIV.
21 1 bit enable_dx10_clamp Wavefront starts execution with DX10 clamp mode enabled. Used by the vector ALU to force DX-10 style treatment of NaN’s (when set, clamp NaN to zero, otherwise pass NaN through).
Used by CP to set up COMPUTE_PGM_RSRC1.DX10_CLAMP.
22 1 bit debug_mode Must be 0.
Start executing wavefront in single step mode.
CP is responsible for filling in COMPUTE_PGM_RSRC1.DEBUG_MODE.
23 1 bit enable_ieee_mode Wavefront starts execution with IEEE mode enabled. Floating point opcodes that support exception flag gathering will quiet and propagate signaling-NaN inputs per IEEE 754-2008. Min_dx10 and max_dx10 become IEEE 754-2008 compliant due to signaling-NaN propagation and quieting.
Used by CP to set up COMPUTE_PGM_RSRC1.IEEE_MODE.
24 1 bit bulky Must be 0.
Only one work-group allowed to execute on a compute unit.
CP is responsible for filling in COMPUTE_PGM_RSRC1.BULKY.
25 1 bit cdbg_user Must be 0.
Flag that can be used to control debugging code.
CP is responsible for filling in COMPUTE_PGM_RSRC1.CDBG_USER.
31:26 6 bits Reserved. Must be 0. 32 Total size 4 bytes
compute_pgm_rsrc2 for GFX6-GFX9 Bits Size Field Name Description 0 1 bit enable_sgpr_private_segment _wave_offset Enable the setup of the SGPR wave scratch offset system register (see Initial Kernel Execution State).
Used by CP to set up COMPUTE_PGM_RSRC2.SCRATCH_EN.
5:1 5 bits user_sgpr_count The total number of SGPR user data registers requested. This number must match the number of user data registers enabled.
Used by CP to set up COMPUTE_PGM_RSRC2.USER_SGPR.
6 1 bit enable_trap_handler Set to 1 if code contains a TRAP instruction which requires a trap handler to be enabled.
CP sets COMPUTE_PGM_RSRC2.TRAP_PRESENT if the runtime has installed a trap handler regardless of the setting of this field.
7 1 bit enable_sgpr_workgroup_id_x Enable the setup of the system SGPR register for the work-group id in the X dimension (see Initial Kernel Execution State).
Used by CP to set up COMPUTE_PGM_RSRC2.TGID_X_EN.
8 1 bit enable_sgpr_workgroup_id_y Enable the setup of the system SGPR register for the work-group id in the Y dimension (see Initial Kernel Execution State).
Used by CP to set up COMPUTE_PGM_RSRC2.TGID_Y_EN.
9 1 bit enable_sgpr_workgroup_id_z Enable the setup of the system SGPR register for the work-group id in the Z dimension (see Initial Kernel Execution State).
Used by CP to set up COMPUTE_PGM_RSRC2.TGID_Z_EN.
10 1 bit enable_sgpr_workgroup_info Enable the setup of the system SGPR register for work-group information (see Initial Kernel Execution State).
Used by CP to set up COMPUTE_PGM_RSRC2.TGID_SIZE_EN.
12:11 2 bits enable_vgpr_workitem_id Enable the setup of the VGPR system registers used for the work-item ID. System VGPR Work-Item ID Enumeration Values defines the values.
Used by CP to set up COMPUTE_PGM_RSRC2.TIDIG_CMP_CNT.
13 1 bit enable_exception_address_watch Must be 0.
Wavefront starts execution with address watch exceptions enabled which are generated when L1 has witnessed a thread access an address of interest.
CP is responsible for filling in the address watch bit in COMPUTE_PGM_RSRC2.EXCP_EN_MSB according to what the runtime requests.
14 1 bit enable_exception_memory Must be 0.
Wavefront starts execution with memory violation exceptions exceptions enabled which are generated when a memory violation has occurred for this wave from L1 or LDS (write-to-read-only-memory, mis-aligned atomic, LDS address out of range, illegal address, etc.).
CP sets the memory violation bit in COMPUTE_PGM_RSRC2.EXCP_EN_MSB according to what the runtime requests.
23:15 9 bits granulated_lds_size Must be 0.
CP uses the rounded value from the dispatch packet, not this value, as the dispatch may contain dynamically allocated group segment memory. CP writes directly to COMPUTE_PGM_RSRC2.LDS_SIZE.
Amount of group segment (LDS) to allocate for each work-group. Granularity is device specific:
- GFX6:
- roundup(lds-size / (64 * 4))
- GFX7-GFX9:
- roundup(lds-size / (128 * 4))
24 1 bit enable_exception_ieee_754_fp _invalid_operation Wavefront starts execution with specified exceptions enabled.
Used by CP to set up COMPUTE_PGM_RSRC2.EXCP_EN (set from bits 0..6).
IEEE 754 FP Invalid Operation
25 1 bit enable_exception_fp_denormal _source FP Denormal one or more input operands is a denormal number 26 1 bit enable_exception_ieee_754_fp _division_by_zero IEEE 754 FP Division by Zero 27 1 bit enable_exception_ieee_754_fp _overflow IEEE 754 FP FP Overflow 28 1 bit enable_exception_ieee_754_fp _underflow IEEE 754 FP Underflow 29 1 bit enable_exception_ieee_754_fp _inexact IEEE 754 FP Inexact 30 1 bit enable_exception_int_divide_by _zero Integer Division by Zero (rcp_iflag_f32 instruction only) 31 1 bit Reserved. Must be 0. 32 Total size 4 bytes.
Floating Point Rounding Mode Enumeration Values Enumeration Name Value Description AMD_FLOAT_ROUND_MODE_NEAR_EVEN 0 Round Ties To Even AMD_FLOAT_ROUND_MODE_PLUS_INFINITY 1 Round Toward +infinity AMD_FLOAT_ROUND_MODE_MINUS_INFINITY 2 Round Toward -infinity AMD_FLOAT_ROUND_MODE_ZERO 3 Round Toward 0
Floating Point Denorm Mode Enumeration Values Enumeration Name Value Description AMD_FLOAT_DENORM_MODE_FLUSH_SRC_DST 0 Flush Source and Destination Denorms AMD_FLOAT_DENORM_MODE_FLUSH_DST 1 Flush Output Denorms AMD_FLOAT_DENORM_MODE_FLUSH_SRC 2 Flush Source Denorms AMD_FLOAT_DENORM_MODE_FLUSH_NONE 3 No Flush
System VGPR Work-Item ID Enumeration Values Enumeration Name Value Description AMD_SYSTEM_VGPR_WORKITEM_ID_X 0 Set work-item X dimension ID. AMD_SYSTEM_VGPR_WORKITEM_ID_X_Y 1 Set work-item X and Y dimensions ID. AMD_SYSTEM_VGPR_WORKITEM_ID_X_Y_Z 2 Set work-item X, Y and Z dimensions ID. AMD_SYSTEM_VGPR_WORKITEM_ID_UNDEFINED 3 Undefined.
This section defines the register state that will be set up by the packet processor prior to the start of execution of every wavefront. This is limited by the constraints of the hardware controllers of CP/ADC/SPI.
The order of the SGPR registers is defined, but the compiler can specify which ones are actually setup in the kernel descriptor using the enable_sgpr_* bit fields (see Kernel Descriptor). The register numbers used for enabled registers are dense starting at SGPR0: the first enabled register is SGPR0, the next enabled register is SGPR1 etc.; disabled registers do not have an SGPR number.
The initial SGPRs comprise up to 16 User SRGPs that are set by CP and apply to all waves of the grid. It is possible to specify more than 16 User SGPRs using the enable_sgpr_* bit fields, in which case only the first 16 are actually initialized. These are then immediately followed by the System SGPRs that are set up by ADC/SPI and can have different values for each wave of the grid dispatch.
SGPR register initial state is defined in SGPR Register Set Up Order.
SGPR Register Set Up Order SGPR Order Name (kernel descriptor enable field) Number of SGPRs Description First Private Segment Buffer (enable_sgpr_private _segment_buffer) 4 V# that can be used, together with Scratch Wave Offset as an offset, to access the private memory space using a segment address.
CP uses the value provided by the runtime.
then Dispatch Ptr (enable_sgpr_dispatch_ptr) 2 64 bit address of AQL dispatch packet for kernel dispatch actually executing. then Queue Ptr (enable_sgpr_queue_ptr) 2 64 bit address of amd_queue_t object for AQL queue on which the dispatch packet was queued. then Kernarg Segment Ptr (enable_sgpr_kernarg _segment_ptr) 2 64 bit address of Kernarg segment. This is directly copied from the kernarg_address in the kernel dispatch packet.
Having CP load it once avoids loading it at the beginning of every wavefront.
then Dispatch Id (enable_sgpr_dispatch_id) 2 64 bit Dispatch ID of the dispatch packet being executed. then Flat Scratch Init (enable_sgpr_flat_scratch _init) 2 This is 2 SGPRs:
- GFX6
- Not supported.
- GFX7-GFX8
The first SGPR is a 32 bit byte offset from SH_HIDDEN_PRIVATE_BASE_VIMID to per SPI base of memory for scratch for the queue executing the kernel dispatch. CP obtains this from the runtime.
This is the same offset used in computing the Scratch Segment Buffer base address. The value of Scratch Wave Offset must be added by the kernel machine code and moved to SGPRn-4 for use as the FLAT SCRATCH BASE in flat memory instructions.
The second SGPR is 32 bit byte size of a single work-item’s scratch memory usage. This is directly loaded from the kernel dispatch packet Private Segment Byte Size and rounded up to a multiple of DWORD.
The kernel code must move to SGPRn-3 for use as the FLAT SCRATCH SIZE in flat memory instructions. Having CP load it once avoids loading it at the beginning of every wavefront.
- GFX9
- This is the 64 bit base address of the per SPI scratch backing memory managed by SPI for the queue executing the kernel dispatch. CP obtains this from the runtime (and divides it if there are multiple Shader Arrays each with its own SPI). The value of Scratch Wave Offset must be added by the kernel machine code and moved to SGPRn-4 and SGPRn-3 for use as the FLAT SCRATCH BASE in flat memory instructions.
then Private Segment Size (enable_sgpr_private _segment_size) 1 The 32 bit byte size of a single work-item’s scratch memory allocation. This is the value from the kernel dispatch packet Private Segment Byte Size rounded up by CP to a multiple of DWORD.
Having CP load it once avoids loading it at the beginning of every wavefront.
This is not used for GFX7-GFX8 since it is the same value as the second SGPR of Flat Scratch Init. However, it may be needed for GFX9 which changes the meaning of the Flat Scratch Init value.
then Grid Work-Group Count X (enable_sgpr_grid _workgroup_count_X) 1 32 bit count of the number of work-groups in the X dimension for the grid being executed. Computed from the fields in the kernel dispatch packet as ((grid_size.x + workgroup_size.x - 1) / workgroup_size.x). then Grid Work-Group Count Y (enable_sgpr_grid _workgroup_count_Y && less than 16 previous SGPRs) 1 32 bit count of the number of work-groups in the Y dimension for the grid being executed. Computed from the fields in the kernel dispatch packet as ((grid_size.y + workgroup_size.y - 1) / workgroupSize.y).
Only initialized if <16 previous SGPRs initialized.
then Grid Work-Group Count Z (enable_sgpr_grid _workgroup_count_Z && less than 16 previous SGPRs) 1 32 bit count of the number of work-groups in the Z dimension for the grid being executed. Computed from the fields in the kernel dispatch packet as ((grid_size.z + workgroup_size.z - 1) / workgroupSize.z).
Only initialized if <16 previous SGPRs initialized.
then Work-Group Id X (enable_sgpr_workgroup_id _X) 1 32 bit work-group id in X dimension of grid for wavefront. then Work-Group Id Y (enable_sgpr_workgroup_id _Y) 1 32 bit work-group id in Y dimension of grid for wavefront. then Work-Group Id Z (enable_sgpr_workgroup_id _Z) 1 32 bit work-group id in Z dimension of grid for wavefront. then Work-Group Info (enable_sgpr_workgroup _info) 1 {first_wave, 14’b0000, ordered_append_term[10:0], threadgroup_size_in_waves[5:0]} then Scratch Wave Offset (enable_sgpr_private _segment_wave_offset) 1 32 bit byte offset from base of scratch base of queue executing the kernel dispatch. Must be used as an offset with Private segment address when using Scratch Segment Buffer. It must be used to set up FLAT SCRATCH for flat addressing (see Flat Scratch).
The order of the VGPR registers is defined, but the compiler can specify which ones are actually setup in the kernel descriptor using the enable_vgpr* bit fields (see Kernel Descriptor). The register numbers used for enabled registers are dense starting at VGPR0: the first enabled register is VGPR0, the next enabled register is VGPR1 etc.; disabled registers do not have a VGPR number.
VGPR register initial state is defined in VGPR Register Set Up Order.
VGPR Register Set Up Order VGPR Order Name (kernel descriptor enable field) Number of VGPRs Description First Work-Item Id X (Always initialized) 1 32 bit work item id in X dimension of work-group for wavefront lane. then Work-Item Id Y (enable_vgpr_workitem_id > 0) 1 32 bit work item id in Y dimension of work-group for wavefront lane. then Work-Item Id Z (enable_vgpr_workitem_id > 1) 1 32 bit work item id in Z dimension of work-group for wavefront lane.
The setting of registers is is done by GPU CP/ADC/SPI hardware as follows:
Flat Scratch register pair are adjacent SGRRs so they can be moved as a 64 bit value to the hardware required SGPRn-3 and SGPRn-4 respectively.
The global segment can be accessed either using buffer instructions (GFX6 which has V# 64 bit address support), flat instructions (GFX7-9), or global instructions (GFX9).
If buffer operations are used then the compiler can generate a V# with the following properties:
If the kernel may use flat operations to access scratch memory, the prolog code must set up FLAT_SCRATCH register pair (FLAT_SCRATCH_LO/FLAT_SCRATCH_HI which are in SGPRn-4/SGPRn-3). Initialization uses Flat Scratch Init and Scratch Wave Offset SGPR registers (see Initial Kernel Execution State):
This section describes the mapping of LLVM memory model onto AMDGPU machine code (see Memory Model for Concurrent Operations). The implementation is WIP.
The AMDGPU backend supports the memory synchronization scopes specified in Memory Scopes.
The code sequences used to implement the memory model are defined in table AMDHSA Memory Model Code Sequences GFX6-GFX9.
The sequences specify the order of instructions that a single thread must execute. The s_waitcnt and buffer_wbinvl1_vol are defined with respect to other memory instructions executed by the same thread. This allows them to be moved earlier or later which can allow them to be combined with other instances of the same instruction, or hoisted/sunk out of loops to improve performance. Only the instructions related to the memory model are given; additional s_waitcnt instructions are required to ensure registers are defined before being used. These may be able to be combined with the memory model s_waitcnt instructions as described above.
The AMDGPU memory model supports both the HSA [HSA] memory model, and the OpenCL [OpenCL] memory model. The HSA memory model uses a single happens-before relation for all address spaces (see Address Spaces). The OpenCL memory model which has separate happens-before relations for the global and local address spaces, and only a fence specifying both global and local address space joins the relationships. Since the LLVM memfence instruction does not allow an address space to be specified the OpenCL fence has to convervatively assume both local and global address space was specified. However, optimizations can often be done to eliminate the additional s_waitcnt``instructions when there are no intervening corresponding ``ds/flat_load/store/atomic memory instructions. The code sequences in the table indicate what can be omitted for the OpenCL memory. The target triple environment is used to determine if the source language is OpenCL (see OpenCL).
ds/flat_load/store/atomic instructions to local memory are termed LDS operations.
buffer/global/flat_load/store/atomic instructions to global memory are termed vector memory operations.
For GFX6-GFX9:
Private address space uses buffer_load/store using the scratch V# (GFX6-8), or scratch_load/store (GFX9). Since only a single thread is accessing the memory, atomic memory orderings are not meaningful and all accesses are treated as non-atomic.
Constant address space uses buffer/global_load instructions (or equivalent scalar memory instructions). Since the constant address space contents do not change during the execution of a kernel dispatch it is not legal to perform stores, and atomic memory orderings are not meaningful and all access are treated as non-atomic.
A memory synchronization scope wider than work-group is not meaningful for the group (LDS) address space and is treated as work-group.
The memory model does not support the region address space which is treated as non-atomic.
Acquire memory ordering is not meaningful on store atomic instructions and is treated as non-atomic.
Release memory ordering is not meaningful on load atomic instructions and is treated a non-atomic.
Acquire-release memory ordering is not meaningful on load or store atomic instructions and is treated as acquire and release respectively.
AMDGPU backend only uses scalar memory operations to access memory that is proven to not change during the execution of the kernel dispatch. This includes constant address space and global address space for program scope const variables. Therefore the kernel machine code does not have to maintain the scalar L1 cache to ensure it is coherent with the vector L1 cache. The scalar and vector L1 caches are invalidated between kernel dispatches by CP since constant address space data may change between kernel dispatch executions. See Memory Spaces.
The one execption is if scalar writes are used to spill SGPR registers. In this case the AMDGPU backend ensures the memory location used to spill is never accessed by vector memory operations at the same time. If scalar writes are used then a s_dcache_wb is inserted before the s_endpgm and before a function return since the locations may be used for vector memory instructions by a future wave that uses the same scratch area, or a function call that creates a frame at the same address, respectively. There is no need for a s_dcache_inv as all scalar writes are write-before-read in the same thread.
Scratch backing memory (which is used for the private address space) is accessed with MTYPE NC_NV (non-coherenent non-volatile). Since the private address space is only accessed by a single thread, and is always write-before-read, there is never a need to invalidate these entries from the L1 cache. Hence all cache invalidates are done as *_vol to only invalidate the volatile cache lines.
On dGPU the kernarg backing memory is accessed as UC (uncached) to avoid needing to invalidate the L2 cache. This also causes it to be treated as non-volatile and so is not invalidated by *_vol. On APU it is accessed as CC (cache coherent) and so the L2 cache will coherent with the CPU and other agents.
AMDHSA Memory Model Code Sequences GFX6-GFX9 LLVM Instr LLVM Memory Ordering LLVM Memory Sync Scope AMDGPU Address Space AMDGPU Machine Code Non-Atomic load none none
- global
- generic
- non-volatile
- buffer/global/flat_load
- volatile
- buffer/global/flat_load glc=1
load none none
- local
- ds_load
store none none
- global
- generic
- buffer/global/flat_store
store none none
- local
- ds_store
Unordered Atomic load atomic unordered any any Same as non-atomic. store atomic unordered any any Same as non-atomic. atomicrmw unordered any any Same as monotonic atomic. Monotonic Atomic load atomic monotonic
- singlethread
- wavefront
- workgroup
- global
- generic
- buffer/global/flat_load
load atomic monotonic
- singlethread
- wavefront
- workgroup
- local
- ds_load
load atomic monotonic
- agent
- system
- global
- generic
- buffer/global/flat_load glc=1
store atomic monotonic
- singlethread
- wavefront
- workgroup
- agent
- system
- global
- generic
- buffer/global/flat_store
store atomic monotonic
- singlethread
- wavefront
- workgroup
- local
- ds_store
atomicrmw monotonic
- singlethread
- wavefront
- workgroup
- agent
- system
- global
- generic
- buffer/global/flat_atomic
atomicrmw monotonic
- singlethread
- wavefront
- workgroup
- local
- ds_atomic
Acquire Atomic load atomic acquire
- singlethread
- wavefront
- global
- local
- generic
- buffer/global/ds/flat_load
load atomic acquire
- workgroup
- global
- buffer/global_load
load atomic acquire
- workgroup
- local
- generic
- ds/flat_load
- s_waitcnt lgkmcnt(0)
- If OpenCL, omit waitcnt.
- Must happen before any following global/generic load/load atomic/store/store atomic/atomicrmw.
- Ensures any following global data read is no older than the load atomic value being acquired.
load atomic acquire
- agent
- system
- global
- buffer/global_load glc=1
- s_waitcnt vmcnt(0)
- Must happen before following buffer_wbinvl1_vol.
- Ensures the load has completed before invalidating the cache.
- buffer_wbinvl1_vol
- Must happen before any following global/generic load/load atomic/atomicrmw.
- Ensures that following loads will not see stale global data.
load atomic acquire
- agent
- system
- generic
- flat_load glc=1
- s_waitcnt vmcnt(0) & lgkmcnt(0)
- If OpenCL omit lgkmcnt(0).
- Must happen before following buffer_wbinvl1_vol.
- Ensures the flat_load has completed before invalidating the cache.
- buffer_wbinvl1_vol
- Must happen before any following global/generic load/load atomic/atomicrmw.
- Ensures that following loads will not see stale global data.
atomicrmw acquire
- singlethread
- wavefront
- global
- local
- generic
- buffer/global/ds/flat_atomic
atomicrmw acquire
- workgroup
- global
- buffer/global_atomic
atomicrmw acquire
- workgroup
- local
- generic
- ds/flat_atomic
- waitcnt lgkmcnt(0)
- If OpenCL, omit waitcnt.
- Must happen before any following global/generic load/load atomic/store/store atomic/atomicrmw.
- Ensures any following global data read is no older than the atomicrmw value being acquired.
atomicrmw acquire
- agent
- system
- global
- buffer/global_atomic
- s_waitcnt vmcnt(0)
- Must happen before following buffer_wbinvl1_vol.
- Ensures the atomicrmw has completed before invalidating the cache.
- buffer_wbinvl1_vol
- Must happen before any following global/generic load/load atomic/atomicrmw.
- Ensures that following loads will not see stale global data.
atomicrmw acquire
- agent
- system
- generic
- flat_atomic
- s_waitcnt vmcnt(0) & lgkmcnt(0)
- If OpenCL, omit lgkmcnt(0).
- Must happen before following buffer_wbinvl1_vol.
- Ensures the atomicrmw has completed before invalidating the cache.
- buffer_wbinvl1_vol
- Must happen before any following global/generic load/load atomic/atomicrmw.
- Ensures that following loads will not see stale global data.
fence acquire
- singlethread
- wavefront
none none fence acquire
- workgroup
none
- s_waitcnt lgkmcnt(0)
- If OpenCL and address space is not generic, omit waitcnt. However, since LLVM currently has no address space on the fence need to conservatively always generate. If fence had an address space then set to address space of OpenCL fence flag, or to generic if both local and global flags are specified.
- Must happen after any preceding local/generic load atomic/atomicrmw with an equal or wider sync scope and memory ordering stronger than unordered (this is termed the fence-paired-atomic).
- Must happen before any following global/generic load/load atomic/store/store atomic/atomicrmw.
- Ensures any following global data read is no older than the value read by the fence-paired-atomic.
fence acquire
- agent
- system
none
- s_waitcnt vmcnt(0) & lgkmcnt(0)
- If OpenCL and address space is not generic, omit lgkmcnt(0). However, since LLVM currently has no address space on the fence need to conservatively always generate (see comment for previous fence).
- Could be split into separate s_waitcnt vmcnt(0) and s_waitcnt lgkmcnt(0) to allow them to be independently moved according to the following rules.
- s_waitcnt vmcnt(0) must happen after any preceding global/generic load atomic/atomicrmw with an equal or wider sync scope and memory ordering stronger than unordered (this is termed the fence-paired-atomic).
- s_waitcnt lgkmcnt(0) must happen after any preceding group/generic load atomic/atomicrmw with an equal or wider sync scope and memory ordering stronger than unordered (this is termed the fence-paired-atomic).
- Must happen before the following buffer_wbinvl1_vol.
- Ensures that the fence-paired atomic has completed before invalidating the cache. Therefore any following locations read must be no older than the value read by the fence-paired-atomic.
- buffer_wbinvl1_vol
- Must happen before any following global/generic load/load atomic/store/store atomic/atomicrmw.
- Ensures that following loads will not see stale global data.
Release Atomic store atomic release
- singlethread
- wavefront
- global
- local
- generic
- buffer/global/ds/flat_store
store atomic release
- workgroup
- global
- generic
- s_waitcnt lgkmcnt(0)
- If OpenCL, omit waitcnt.
- Must happen after any preceding local/generic load/store/load atomic/store atomic/atomicrmw.
- Must happen before the following store.
- Ensures that all memory operations to local have completed before performing the store that is being released.
- buffer/global/flat_store
store atomic release
- workgroup
- local
- ds_store
store atomic release
- agent
- system
- global
- generic
- s_waitcnt vmcnt(0) & lgkmcnt(0)
- If OpenCL, omit lgkmcnt(0).
- Could be split into separate s_waitcnt vmcnt(0) and s_waitcnt lgkmcnt(0) to allow them to be independently moved according to the following rules.
- s_waitcnt vmcnt(0) must happen after any preceding global/generic load/store/load atomic/store atomic/atomicrmw.
- s_waitcnt lgkmcnt(0) must happen after any preceding local/generic load/store/load atomic/store atomic/atomicrmw.
- Must happen before the following store.
- Ensures that all memory operations to global have completed before performing the store that is being released.
- buffer/global/ds/flat_store
atomicrmw release
- singlethread
- wavefront
- global
- local
- generic
- buffer/global/ds/flat_atomic
atomicrmw release
- workgroup
- global
- generic
- s_waitcnt lgkmcnt(0)
- If OpenCL, omit waitcnt.
- Must happen after any preceding local/generic load/store/load atomic/store atomic/atomicrmw.
- Must happen before the following atomicrmw.
- Ensures that all memory operations to local have completed before performing the atomicrmw that is being released.
- buffer/global/flat_atomic
atomicrmw release
- workgroup
- local
- ds_atomic
atomicrmw release
- agent
- system
- global
- generic
- s_waitcnt vmcnt(0) & lgkmcnt(0)
- If OpenCL, omit lgkmcnt(0).
- Could be split into separate s_waitcnt vmcnt(0) and s_waitcnt lgkmcnt(0) to allow them to be independently moved according to the following rules.
- s_waitcnt vmcnt(0) must happen after any preceding global/generic load/store/load atomic/store atomic/atomicrmw.
- s_waitcnt lgkmcnt(0) must happen after any preceding local/generic load/store/load atomic/store atomic/atomicrmw.
- Must happen before the following atomicrmw.
- Ensures that all memory operations to global and local have completed before performing the atomicrmw that is being released.
- buffer/global/ds/flat_atomic*
fence release
- singlethread
- wavefront
none none fence release
- workgroup
none
- s_waitcnt lgkmcnt(0)
- If OpenCL and address space is not generic, omit waitcnt. However, since LLVM currently has no address space on the fence need to conservatively always generate (see comment for previous fence).
- Must happen after any preceding local/generic load/load atomic/store/store atomic/atomicrmw.
- Must happen before any following store atomic/atomicrmw with an equal or wider sync scope and memory ordering stronger than unordered (this is termed the fence-paired-atomic).
- Ensures that all memory operations to local have completed before performing the following fence-paired-atomic.
fence release
- agent
- system
none
- s_waitcnt vmcnt(0) & lgkmcnt(0)
- If OpenCL and address space is not generic, omit lgkmcnt(0). However, since LLVM currently has no address space on the fence need to conservatively always generate (see comment for previous fence).
- Could be split into separate s_waitcnt vmcnt(0) and s_waitcnt lgkmcnt(0) to allow them to be independently moved according to the following rules.
- s_waitcnt vmcnt(0) must happen after any preceding global/generic load/store/load atomic/store atomic/atomicrmw.
- s_waitcnt lgkmcnt(0) must happen after any preceding local/generic load/store/load atomic/store atomic/atomicrmw.
- Must happen before any following store atomic/atomicrmw with an equal or wider sync scope and memory ordering stronger than unordered (this is termed the fence-paired-atomic).
- Ensures that all memory operations to global have completed before performing the following fence-paired-atomic.
Acquire-Release Atomic atomicrmw acq_rel
- singlethread
- wavefront
- global
- local
- generic
- buffer/global/ds/flat_atomic
atomicrmw acq_rel
- workgroup
- global
- s_waitcnt lgkmcnt(0)
- If OpenCL, omit waitcnt.
- Must happen after any preceding local/generic load/store/load atomic/store atomic/atomicrmw.
- Must happen before the following atomicrmw.
- Ensures that all memory operations to local have completed before performing the atomicrmw that is being released.
- buffer/global_atomic
atomicrmw acq_rel
- workgroup
- local
- ds_atomic
- s_waitcnt lgkmcnt(0)
- If OpenCL, omit waitcnt.
- Must happen before any following global/generic load/load atomic/store/store atomic/atomicrmw.
- Ensures any following global data read is no older than the load atomic value being acquired.
atomicrmw acq_rel
- workgroup
- generic
- s_waitcnt lgkmcnt(0)
- If OpenCL, omit waitcnt.
- Must happen after any preceding local/generic load/store/load atomic/store atomic/atomicrmw.
- Must happen before the following atomicrmw.
- Ensures that all memory operations to local have completed before performing the atomicrmw that is being released.
- flat_atomic
- s_waitcnt lgkmcnt(0)
- If OpenCL, omit waitcnt.
- Must happen before any following global/generic load/load atomic/store/store atomic/atomicrmw.
- Ensures any following global data read is no older than the load atomic value being acquired.
atomicrmw acq_rel
- agent
- system
- global
- s_waitcnt vmcnt(0) & lgkmcnt(0)
- If OpenCL, omit lgkmcnt(0).
- Could be split into separate s_waitcnt vmcnt(0) and s_waitcnt lgkmcnt(0) to allow them to be independently moved according to the following rules.
- s_waitcnt vmcnt(0) must happen after any preceding global/generic load/store/load atomic/store atomic/atomicrmw.
- s_waitcnt lgkmcnt(0) must happen after any preceding local/generic load/store/load atomic/store atomic/atomicrmw.
- Must happen before the following atomicrmw.
- Ensures that all memory operations to global have completed before performing the atomicrmw that is being released.
- buffer/global_atomic
- s_waitcnt vmcnt(0)
- Must happen before following buffer_wbinvl1_vol.
- Ensures the atomicrmw has completed before invalidating the cache.
- buffer_wbinvl1_vol
- Must happen before any following global/generic load/load atomic/atomicrmw.
- Ensures that following loads will not see stale global data.
atomicrmw acq_rel
- agent
- system
- generic
- s_waitcnt vmcnt(0) & lgkmcnt(0)
- If OpenCL, omit lgkmcnt(0).
- Could be split into separate s_waitcnt vmcnt(0) and s_waitcnt lgkmcnt(0) to allow them to be independently moved according to the following rules.
- s_waitcnt vmcnt(0) must happen after any preceding global/generic load/store/load atomic/store atomic/atomicrmw.
- s_waitcnt lgkmcnt(0) must happen after any preceding local/generic load/store/load atomic/store atomic/atomicrmw.
- Must happen before the following atomicrmw.
- Ensures that all memory operations to global have completed before performing the atomicrmw that is being released.
- flat_atomic
- s_waitcnt vmcnt(0) & lgkmcnt(0)
- If OpenCL, omit lgkmcnt(0).
- Must happen before following buffer_wbinvl1_vol.
- Ensures the atomicrmw has completed before invalidating the cache.
- buffer_wbinvl1_vol
- Must happen before any following global/generic load/load atomic/atomicrmw.
- Ensures that following loads will not see stale global data.
fence acq_rel
- singlethread
- wavefront
none none fence acq_rel
- workgroup
none
- s_waitcnt lgkmcnt(0)
- If OpenCL and address space is not generic, omit waitcnt. However, since LLVM currently has no address space on the fence need to conservatively always generate (see comment for previous fence).
- Must happen after any preceding local/generic load/load atomic/store/store atomic/atomicrmw.
- Must happen before any following global/generic load/load atomic/store/store atomic/atomicrmw.
- Ensures that all memory operations to local have completed before performing any following global memory operations.
- Ensures that the preceding local/generic load atomic/atomicrmw with an equal or wider sync scope and memory ordering stronger than unordered (this is termed the fence-paired-atomic) has completed before following global memory operations. This satisfies the requirements of acquire.
- Ensures that all previous memory operations have completed before a following local/generic store atomic/atomicrmw with an equal or wider sync scope and memory ordering stronger than unordered (this is termed the fence-paired-atomic). This satisfies the requirements of release.
fence acq_rel
- agent
- system
none
- s_waitcnt vmcnt(0) & lgkmcnt(0)
- If OpenCL and address space is not generic, omit lgkmcnt(0). However, since LLVM currently has no address space on the fence need to conservatively always generate (see comment for previous fence).
- Could be split into separate s_waitcnt vmcnt(0) and s_waitcnt lgkmcnt(0) to allow them to be independently moved according to the following rules.
- s_waitcnt vmcnt(0) must happen after any preceding global/generic load/store/load atomic/store atomic/atomicrmw.
- s_waitcnt lgkmcnt(0) must happen after any preceding local/generic load/store/load atomic/store atomic/atomicrmw.
- Must happen before the following buffer_wbinvl1_vol.
- Ensures that the preceding global/local/generic load atomic/atomicrmw with an equal or wider sync scope and memory ordering stronger than unordered (this is termed the fence-paired-atomic) has completed before invalidating the cache. This satisfies the requirements of acquire.
- Ensures that all previous memory operations have completed before a following global/local/generic store atomic/atomicrmw with an equal or wider sync scope and memory ordering stronger than unordered (this is termed the fence-paired-atomic). This satisfies the requirements of release.
- buffer_wbinvl1_vol
- Must happen before any following global/generic load/load atomic/store/store atomic/atomicrmw.
- Ensures that following loads will not see stale global data. This satisfies the requirements of acquire.
Sequential Consistent Atomic load atomic seq_cst
- singlethread
- wavefront
- workgroup
- global
- local
- generic
Same as corresponding load atomic acquire. load atomic seq_cst
- agent
- system
- global
- local
- generic
- s_waitcnt vmcnt(0)
- Must happen after preceding global/generic load atomic/store atomic/atomicrmw with memory ordering of seq_cst and with equal or wider sync scope. (Note that seq_cst fences have their own s_waitcnt vmcnt(0) and so do not need to be considered.)
- Ensures any preceding sequential consistent global memory instructions have completed before executing this sequentially consistent instruction. This prevents reordering a seq_cst store followed by a seq_cst load (Note that seq_cst is stronger than acquire/release as the reordering of load acquire followed by a store release is prevented by the waitcnt vmcnt(0) of the release, but there is nothing preventing a store release followed by load acquire from competing out of order.)
- Following instructions same as corresponding load atomic acquire.
store atomic seq_cst
- singlethread
- wavefront
- workgroup
- global
- local
- generic
Same as corresponding store atomic release. store atomic seq_cst
- agent
- system
- global
- generic
Same as corresponding store atomic release. atomicrmw seq_cst
- singlethread
- wavefront
- workgroup
- global
- local
- generic
Same as corresponding atomicrmw acq_rel. atomicrmw seq_cst
- agent
- system
- global
- generic
Same as corresponding atomicrmw acq_rel. fence seq_cst
- singlethread
- wavefront
- workgroup
- agent
- system
none Same as corresponding fence acq_rel.
The memory order also adds the single thread optimization constrains defined in table AMDHSA Memory Model Single Thread Optimization Constraints GFX6-GFX9.
AMDHSA Memory Model Single Thread Optimization Constraints GFX6-GFX9 LLVM Memory Optimization Constraints Ordering unordered none monotonic none acquire
- If a load atomic/atomicrmw then no following load/load atomic/store/ store atomic/atomicrmw/fence instruction can be moved before the acquire.
- If a fence then same as load atomic, plus no preceding associated fence-paired-atomic can be moved after the fence.
release
- If a store atomic/atomicrmw then no preceding load/load atomic/store/ store atomic/atomicrmw/fence instruction can be moved after the release.
- If a fence then same as store atomic, plus no following associated fence-paired-atomic can be moved before the fence.
acq_rel Same constraints as both acquire and release. seq_cst
- If a load atomic then same constraints as acquire, plus no preceding sequentially consistent load atomic/store atomic/atomicrmw/fence instruction can be moved after the seq_cst.
- If a store atomic then the same constraints as release, plus no following sequentially consistent load atomic/store atomic/atomicrmw/fence instruction can be moved before the seq_cst.
- If an atomicrmw/fence then same constraints as acq_rel.
For code objects generated by AMDGPU backend for HSA [HSA] compatible runtimes (such as ROCm [AMD-ROCm]), the runtime installs a trap handler that supports the s_trap instruction with the following usage:
AMDGPU Trap Handler for AMDHSA OS Usage Code Sequence Trap Handler Inputs Description reserved s_trap 0x00 Reserved by hardware. debugtrap(arg) s_trap 0x01
- SGPR0-1:
- queue_ptr
- VGPR0:
- arg
Reserved for HSA debugtrap intrinsic (not implemented). llvm.trap s_trap 0x02
- SGPR0-1:
- queue_ptr
Causes dispatch to be terminated and its associated queue put into the error state. llvm.debugtrap s_trap 0x03
- SGPR0-1:
- queue_ptr
If debugger not installed handled same as llvm.trap. debugger breakpoint s_trap 0x07 Reserved for debugger breakpoints. debugger s_trap 0x08 Reserved for debugger. debugger s_trap 0xfe Reserved for debugger. debugger s_trap 0xff Reserved for debugger.
For code objects generated by AMDGPU backend for non-amdhsa OS, the runtime does not install a trap handler. The llvm.trap and llvm.debugtrap instructions are handled as follows:
AMDGPU Trap Handler for Non-AMDHSA OS Usage Code Sequence Description llvm.trap s_endpgm Causes wavefront to be terminated. llvm.debugtrap none Compiler warning given that there is no trap handler installed.
When generating code for the OpenCL language the target triple environment should be opencl or amdgizcl (see Target Triples).
When the language is OpenCL the following differences occur:
When generating code for the OpenCL language the target triple environment should be hcc (see Target Triples).
When the language is OpenCL the following differences occur:
AMDGPU backend has LLVM-MC based assembler which is currently in development. It supports AMDGCN GFX6-GFX8.
This section describes general syntax for instructions and operands. For more information about instructions, their semantics and supported combinations of operands, refer to one of instruction set architecture manuals [AMD-Souther-Islands] [AMD-Sea-Islands] [AMD-Volcanic-Islands].
An instruction has the following syntax (register operands are normally comma-separated while extra operands are space-separated):
<opcode> <register_operand0>, ... <extra_operand0> ...
The following syntax for register operands is supported:
The following extra operands are supported:
ds_add_u32 v2, v4 offset:16
ds_write_src2_b64 v2 offset0:4 offset1:8
ds_cmpst_f32 v2, v4, v6
ds_min_rtn_f64 v[8:9], v2, v[4:5]
For full list of supported instructions, refer to “LDS/GDS instructions” in ISA Manual.
flat_load_dword v1, v[3:4]
flat_store_dwordx3 v[3:4], v[5:7]
flat_atomic_swap v1, v[3:4], v5 glc
flat_atomic_cmpswap v1, v[3:4], v[5:6] glc slc
flat_atomic_fmax_x2 v[1:2], v[3:4], v[5:6] glc
For full list of supported instructions, refer to “FLAT instructions” in ISA Manual.
buffer_load_dword v1, off, s[4:7], s1
buffer_store_dwordx4 v[1:4], v2, ttmp[4:7], s1 offen offset:4 glc tfe
buffer_store_format_xy v[1:2], off, s[4:7], s1
buffer_wbinvl1
buffer_atomic_inc v1, v2, s[8:11], s4 idxen offset:4 slc
For full list of supported instructions, refer to “MUBUF Instructions” in ISA Manual.
s_load_dword s1, s[2:3], 0xfc
s_load_dwordx8 s[8:15], s[2:3], s4
s_load_dwordx16 s[88:103], s[2:3], s4
s_dcache_inv_vol
s_memtime s[4:5]
For full list of supported instructions, refer to “Scalar Memory Operations” in ISA Manual.
s_mov_b32 s1, s2
s_mov_b64 s[0:1], 0x80000000
s_cmov_b32 s1, 200
s_wqm_b64 s[2:3], s[4:5]
s_bcnt0_i32_b64 s1, s[2:3]
s_swappc_b64 s[2:3], s[4:5]
s_cbranch_join s[4:5]
For full list of supported instructions, refer to “SOP1 Instructions” in ISA Manual.
s_add_u32 s1, s2, s3
s_and_b64 s[2:3], s[4:5], s[6:7]
s_cselect_b32 s1, s2, s3
s_andn2_b32 s2, s4, s6
s_lshr_b64 s[2:3], s[4:5], s6
s_ashr_i32 s2, s4, s6
s_bfm_b64 s[2:3], s4, s6
s_bfe_i64 s[2:3], s[4:5], s6
s_cbranch_g_fork s[4:5], s[6:7]
For full list of supported instructions, refer to “SOP2 Instructions” in ISA Manual.
s_cmp_eq_i32 s1, s2
s_bitcmp1_b32 s1, s2
s_bitcmp0_b64 s[2:3], s4
s_setvskip s3, s5
For full list of supported instructions, refer to “SOPC Instructions” in ISA Manual.
s_barrier
s_nop 2
s_endpgm
s_waitcnt 0 ; Wait for all counters to be 0
s_waitcnt vmcnt(0) & expcnt(0) & lgkmcnt(0) ; Equivalent to above
s_waitcnt vmcnt(1) ; Wait for vmcnt counter to be 1.
s_sethalt 9
s_sleep 10
s_sendmsg 0x1
s_sendmsg sendmsg(MSG_INTERRUPT)
s_trap 1
For full list of supported instructions, refer to “SOPP Instructions” in ISA Manual.
Unless otherwise mentioned, little verification is performed on the operands of SOPP Instructions, so it is up to the programmer to be familiar with the range or acceptable values.
For vector ALU instruction opcodes (VOP1, VOP2, VOP3, VOPC, VOP_DPP, VOP_SDWA), the assembler will automatically use optimal encoding based on its operands. To force specific encoding, one can add a suffix to the opcode of the instruction:
VOP1/VOP2/VOP3/VOPC examples:
v_mov_b32 v1, v2
v_mov_b32_e32 v1, v2
v_nop
v_cvt_f64_i32_e32 v[1:2], v2
v_floor_f32_e32 v1, v2
v_bfrev_b32_e32 v1, v2
v_add_f32_e32 v1, v2, v3
v_mul_i32_i24_e64 v1, v2, 3
v_mul_i32_i24_e32 v1, -3, v3
v_mul_i32_i24_e32 v1, -100, v3
v_addc_u32 v1, s[0:1], v2, v3, s[2:3]
v_max_f16_e32 v1, v2, v3
VOP_DPP examples:
v_mov_b32 v0, v0 quad_perm:[0,2,1,1]
v_sin_f32 v0, v0 row_shl:1 row_mask:0xa bank_mask:0x1 bound_ctrl:0
v_mov_b32 v0, v0 wave_shl:1
v_mov_b32 v0, v0 row_mirror
v_mov_b32 v0, v0 row_bcast:31
v_mov_b32 v0, v0 quad_perm:[1,3,0,1] row_mask:0xa bank_mask:0x1 bound_ctrl:0
v_add_f32 v0, v0, |v0| row_shl:1 row_mask:0xa bank_mask:0x1 bound_ctrl:0
v_max_f16 v1, v2, v3 row_shl:1 row_mask:0xa bank_mask:0x1 bound_ctrl:0
VOP_SDWA examples:
v_mov_b32 v1, v2 dst_sel:BYTE_0 dst_unused:UNUSED_PRESERVE src0_sel:DWORD
v_min_u32 v200, v200, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:BYTE_1 src1_sel:DWORD
v_sin_f32 v0, v0 dst_unused:UNUSED_PAD src0_sel:WORD_1
v_fract_f32 v0, |v0| dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1
v_cmpx_le_u32 vcc, v1, v2 src0_sel:BYTE_2 src1_sel:WORD_0
For full list of supported instructions, refer to “Vector ALU instructions”.
AMDGPU ABI defines auxiliary data in output code object. In assembly source, one can specify them with assembler directives.
major and minor are integers that specify the version of the HSA code object that will be generated by the assembler.
major, minor, and stepping are all integers that describe the instruction set architecture (ISA) version of the assembly program.
vendor and arch are quoted strings. vendor should always be equal to “AMD” and arch should always be equal to “AMDGPU”.
By default, the assembler will derive the ISA version, vendor, and arch from the value of the -mcpu option that is passed to the assembler.
This directives specifies that the symbol with given name is a kernel entry point (label) and the object should contain corresponding symbol of type STT_AMDGPU_HSA_KERNEL.
This directive marks the beginning of a list of key / value pairs that are used to specify the amd_kernel_code_t object that will be emitted by the assembler. The list must be terminated by the .end_amd_kernel_code_t directive. For any amd_kernel_code_t values that are unspecified a default value will be used. The default value for all keys is 0, with the following exceptions:
The .amd_kernel_code_t directive must be placed immediately after the function label and before any instructions.
For a full list of amd_kernel_code_t keys, refer to AMDGPU ABI document, comments in lib/Target/AMDGPU/AmdKernelCodeT.h and test/CodeGen/AMDGPU/hsa.s.
Here is an example of a minimal amd_kernel_code_t specification:
.hsa_code_object_version 1,0
.hsa_code_object_isa
.hsatext
.globl hello_world
.p2align 8
.amdgpu_hsa_kernel hello_world
hello_world:
.amd_kernel_code_t
enable_sgpr_kernarg_segment_ptr = 1
is_ptr64 = 1
compute_pgm_rsrc1_vgprs = 0
compute_pgm_rsrc1_sgprs = 0
compute_pgm_rsrc2_user_sgpr = 2
kernarg_segment_byte_size = 8
wavefront_sgpr_count = 2
workitem_vgpr_count = 3
.end_amd_kernel_code_t
s_load_dwordx2 s[0:1], s[0:1] 0x0
v_mov_b32 v0, 3.14159
s_waitcnt lgkmcnt(0)
v_mov_b32 v1, s0
v_mov_b32 v2, s1
flat_store_dword v[1:2], v0
s_endpgm
.Lfunc_end0:
.size hello_world, .Lfunc_end0-hello_world
| [AMD-R6xx] | AMD R6xx shader ISA |
| [AMD-R7xx] | AMD R7xx shader ISA |
| [AMD-Evergreen] | AMD Evergreen shader ISA |
| [AMD-Cayman-Trinity] | AMD Cayman/Trinity shader ISA |
| [AMD-Souther-Islands] | (1, 2) AMD Southern Islands Series ISA |
| [AMD-Sea-Islands] | (1, 2) AMD Sea Islands Series ISA |
| [AMD-Volcanic-Islands] | (1, 2) AMD GCN3 Instruction Set Architecture |
| [AMD-OpenCL_Programming-Guide] | AMD Accelerated Parallel Processing OpenCL Programming Guide |
| [AMD-APP-SDK] | AMD Accelerated Parallel Processing APP SDK Documentation |
| [AMD-ROCm] | (1, 2, 3, 4) ROCm: Open Platform for Development, Discovery and Education Around GPU Computing |
| [AMD-ROCm-github] | (1, 2) ROCm github |
| [HSA] | (1, 2, 3, 4, 5, 6, 7, 8, 9, 10) Heterogeneous System Architecture (HSA) Foundation |
| [ELF] | (1, 2) Executable and Linkable Format (ELF) |
| [DWARF] | DWARF Debugging Information Format |
| [YAML] | YAML Ain’t Markup Language (YAML™) Version 1.2 |
| [OpenCL] | (1, 2) The OpenCL Specification Version 2.0 |
| [HRF] | Heterogeneous-race-free Memory Models |
| [AMD-AMDGPU-Compute-Application-Binary-Interface] | AMDGPU Compute Application Binary Interface |