Mojo module

info

Contains information about GPU architectures and their capabilities.

This module provides detailed specifications for various GPU models including NVIDIA and AMD GPUs. It includes information about compute capabilities, memory specifications, thread organization, and performance characteristics.

Aliases

A10

alias A10 = Info(__init__[__mlir_type.!kgen.string]("A10"), Vendor(__init__[__mlir_type.!pop.int_literal](2)), __init__[__mlir_type.!kgen.string]("cuda"), __init__[__mlir_type.!kgen.string]("ampere"), __init__[__mlir_type.!kgen.string]("nvptx-short-ptr=true"), __init__[__mlir_type.!pop.float_literal](8.5999999999999996), __init__[__mlir_type.!kgen.string]("sm_86"), 72, 32, 1536, 32, 64, 2048, 32, 102400, 65536, 256, __init__[__mlir_type.!kgen.string]("warp"), 255, 65536, 16, 128, 4, 1024)

A100

alias A100 = Info(__init__[__mlir_type.!kgen.string]("A100"), Vendor(__init__[__mlir_type.!pop.int_literal](2)), __init__[__mlir_type.!kgen.string]("cuda"), __init__[__mlir_type.!kgen.string]("ampere"), __init__[__mlir_type.!kgen.string]("nvptx-short-ptr=true"), __init__[__mlir_type.!pop.float_literal](8), __init__[__mlir_type.!kgen.string]("sm_80"), 108, 32, 2048, 32, 64, 2048, 32, 167936, 65536, 256, __init__[__mlir_type.!kgen.string]("warp"), 255, 65536, 32, 128, 4, 1024)

B100

alias B100 = Info(__init__[__mlir_type.!kgen.string]("B100"), Vendor(__init__[__mlir_type.!pop.int_literal](2)), __init__[__mlir_type.!kgen.string]("cuda"), __init__[__mlir_type.!kgen.string]("blackwell"), __init__[__mlir_type.!kgen.string]("nvptx-short-ptr=true"), __init__[__mlir_type.!pop.float_literal](10), __init__[__mlir_type.!kgen.string]("sm_100a"), 132, 32, -1, 32, 64, 1536, 32, 262144, 65536, 256, __init__[__mlir_type.!kgen.string]("warp"), 255, 65536, 32, 128, 4, 1024)

B200

alias B200 = Info(__init__[__mlir_type.!kgen.string]("B200"), Vendor(__init__[__mlir_type.!pop.int_literal](2)), __init__[__mlir_type.!kgen.string]("cuda"), __init__[__mlir_type.!kgen.string]("blackwell"), __init__[__mlir_type.!kgen.string]("nvptx-short-ptr=true"), __init__[__mlir_type.!pop.float_literal](10), __init__[__mlir_type.!kgen.string]("sm_100a"), 148, 32, -1, 32, 64, 2048, 32, 233472, 65536, 256, __init__[__mlir_type.!kgen.string]("warp"), 255, 65536, 32, 128, 4, 1024)

DEFAULT_GPU

alias DEFAULT_GPU = from_name[::StringSlice[::Bool()

DEFAULT_GPU_ARCH

alias DEFAULT_GPU_ARCH = _accelerator_arch()

DEFAULT_GPU_TARGET

alias DEFAULT_GPU_TARGET = from_name[::StringSlice[::Bool().target()

H100

alias H100 = Info(__init__[__mlir_type.!kgen.string]("H100"), Vendor(__init__[__mlir_type.!pop.int_literal](2)), __init__[__mlir_type.!kgen.string]("cuda"), __init__[__mlir_type.!kgen.string]("hopper"), __init__[__mlir_type.!kgen.string]("nvptx-short-ptr=true"), __init__[__mlir_type.!pop.float_literal](9), __init__[__mlir_type.!kgen.string]("sm_90a"), 132, 32, 2048, 32, 64, 2048, 32, 233472, 65536, 256, __init__[__mlir_type.!kgen.string]("warp"), 255, 65536, 32, 128, 4, 1024)

L4

alias L4 = Info(__init__[__mlir_type.!kgen.string]("L4"), Vendor(__init__[__mlir_type.!pop.int_literal](2)), __init__[__mlir_type.!kgen.string]("cuda"), __init__[__mlir_type.!kgen.string]("ada"), __init__[__mlir_type.!kgen.string]("nvptx-short-ptr=true"), __init__[__mlir_type.!pop.float_literal](8.9000000000000004), __init__[__mlir_type.!kgen.string]("sm_89"), 58, 32, 1536, 32, 64, 2048, 32, 102400, 65536, 256, __init__[__mlir_type.!kgen.string]("warp"), 255, 65536, 24, 128, 4, 1024)

MI300X

alias MI300X = Info(__init__[__mlir_type.!kgen.string]("MI300X"), Vendor(__init__[__mlir_type.!pop.int_literal](1)), __init__[__mlir_type.!kgen.string]("hip"), __init__[__mlir_type.!kgen.string]("gfx942"), __init__[__mlir_type.!kgen.string](""), __init__[__mlir_type.!pop.float_literal](9.4000000000000003), __init__[__mlir_type.!kgen.string]("CDNA3"), 304, 64, 2048, 64, 32, 2048, 2, 65536, 65536, 256, __init__[__mlir_type.!kgen.string]("warp"), 255, 65536, 2, 128, 4, 1024)

NoGPU

alias NoGPU = Info(__init__[__mlir_type.!kgen.string]("NoGPU"), Vendor(__init__[__mlir_type.!pop.int_literal](0)), __init__[__mlir_type.!kgen.string]("none"), __init__[__mlir_type.!kgen.string]("no_gpu"), __init__[__mlir_type.!kgen.string](""), __init__[__mlir_type.!pop.int_literal](0), __init__[__mlir_type.!kgen.string](""), 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, __init__[__mlir_type.!kgen.string]("none"), 0, 0, 0, 0, 0, 0)

OrinNano

alias OrinNano = Info(__init__[__mlir_type.!kgen.string]("Orin Nano"), Vendor(__init__[__mlir_type.!pop.int_literal](2)), __init__[__mlir_type.!kgen.string]("cuda"), __init__[__mlir_type.!kgen.string]("ampere"), __init__[__mlir_type.!kgen.string]("nvptx-short-ptr=true"), __init__[__mlir_type.!pop.float_literal](8.6999999999999993), __init__[__mlir_type.!kgen.string]("sm_87"), 8, 32, 1536, 32, 64, 2048, 32, 167936, 65536, 256, __init__[__mlir_type.!kgen.string]("warp"), 255, 65536, 16, 128, 4, 1024)

Radeon780m

alias Radeon780m = Info(__init__[__mlir_type.!kgen.string]("Radeon 780M"), Vendor(__init__[__mlir_type.!pop.int_literal](1)), __init__[__mlir_type.!kgen.string]("hip"), __init__[__mlir_type.!kgen.string]("gfx1103"), __init__[__mlir_type.!kgen.string](""), __init__[__mlir_type.!pop.float_literal](11), __init__[__mlir_type.!kgen.string]("RDNA3"), 12, 32, 1024, 32, 32, 1024, 2, 32768, 32768, 256, __init__[__mlir_type.!kgen.string]("warp"), 255, 32768, 2, 128, 4, 1024)

RTX2060

alias RTX2060 = Info(__init__[__mlir_type.!kgen.string]("RTX2060"), Vendor(__init__[__mlir_type.!pop.int_literal](2)), __init__[__mlir_type.!kgen.string]("cuda"), __init__[__mlir_type.!kgen.string]("turing"), __init__[__mlir_type.!kgen.string]("nvptx-short-ptr=true"), __init__[__mlir_type.!pop.float_literal](7.5), __init__[__mlir_type.!kgen.string]("sm_75"), 30, 32, 2048, 32, 64, 2048, 16, 65536, 65536, 256, __init__[__mlir_type.!kgen.string]("warp"), 255, 32768, 16, 32, 4, 1024)

RTX4090

alias RTX4090 = Info(__init__[__mlir_type.!kgen.string]("RTX4090"), Vendor(__init__[__mlir_type.!pop.int_literal](2)), __init__[__mlir_type.!kgen.string]("cuda"), __init__[__mlir_type.!kgen.string]("ada lovelace"), __init__[__mlir_type.!kgen.string]("nvptx-short-ptr=true"), __init__[__mlir_type.!pop.float_literal](8.9000000000000004), __init__[__mlir_type.!kgen.string]("sm_89"), 128, 32, -1, 32, 64, 1536, 24, 102400, 65536, 256, __init__[__mlir_type.!kgen.string]("warp"), 255, 65536, 24, 128, 4, 1024)

RTX4090m

alias RTX4090m = Info(__init__[__mlir_type.!kgen.string]("RTX4090m"), Vendor(__init__[__mlir_type.!pop.int_literal](2)), __init__[__mlir_type.!kgen.string]("cuda"), __init__[__mlir_type.!kgen.string]("ada lovelace"), __init__[__mlir_type.!kgen.string]("nvptx-short-ptr=true"), __init__[__mlir_type.!pop.float_literal](8.9000000000000004), __init__[__mlir_type.!kgen.string]("sm_89"), 76, 32, -1, 32, 64, 1536, 24, 102400, 65536, 256, __init__[__mlir_type.!kgen.string]("warp"), 255, 65536, 24, 128, 4, 1024)

RTX5090

alias RTX5090 = Info(__init__[__mlir_type.!kgen.string]("RTX5090"), Vendor(__init__[__mlir_type.!pop.int_literal](2)), __init__[__mlir_type.!kgen.string]("cuda"), __init__[__mlir_type.!kgen.string]("blackwell"), __init__[__mlir_type.!kgen.string]("nvptx-short-ptr=true"), __init__[__mlir_type.!pop.float_literal](12), __init__[__mlir_type.!kgen.string]("sm_120a"), 170, 32, -1, 32, 64, 1536, 32, 59392, 65536, 256, __init__[__mlir_type.!kgen.string]("warp"), 255, 65536, 32, 128, 4, 1024)

Structs

• ​Info: Comprehensive information about a GPU architecture.
• ​Vendor: Represents GPU vendors.

Functions

• ​is_cpu: Checks if the target is a CPU (compile-time version).
• ​is_gpu: Checks if the target is a GPU (compile-time version).
• ​is_valid_target: Checks if the target is valid (compile-time version).

Was this page helpful?

Thank you! We'll create more content like this.

Thank you for helping us improve!

😔 What went wrong?

Some code doesn’t work

It includes inaccurate information

It's missing information I need

It was difficult to understand

Other

Submit