ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Installer: enable ROCm torch on RDNA2 (gfx1030-1036) on Windows (#7277)

Browse source 
* Installer: enable ROCm torch on RDNA2 (gfx1030-1036) on Windows

repo.amd.com publishes a gfx103X-all wheel family with win_amd64 torch
2.9.1/2.10.0/2.11.0+rocm7.13.0 (cp310-313), but both Windows allowlists
omitted RDNA2, so RX 6000 cards (gfx1030/1032, etc.) fell back to CPU-only
torch. Map gfx1030-1036 to gfx103X-all in install.ps1 ($archFamilyMap) and
install_python_stack.py (_GFX_TO_AMD_INDEX_ARCH). No torch floor (mirrors
gfx110X-all: newest wheel, no _grouped_mm bug on RDNA2). NVIDIA/Mac/CPU and
Linux paths untouched; gfx906 stays CPU (no wheels published).

* Sync studio/setup.ps1 RDNA2 (gfx1030-1036) allowlists for PR #7277
This commit is contained in:
Daniel Han 2026-07-21 03:54:25 -07:00 • committed by GitHub
commit 35f887d795
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 17 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

4
install.ps1
View file

@ -2128,6 +2128,10 @@ exit 0
"gfx1151" = "gfx1151"; "gfx1150" = "gfx1150" # RDNA 3.5 (Strix Halo/Point)
"gfx1103" = "gfx110X-all"; "gfx1102" = "gfx110X-all" # RDNA 3
"gfx1101" = "gfx110X-all"; "gfx1100" = "gfx110X-all"
"gfx1036" = "gfx103X-all"; "gfx1035" = "gfx103X-all" # RDNA 2 (RX 6000)
"gfx1034" = "gfx103X-all"; "gfx1033" = "gfx103X-all"
"gfx1032" = "gfx103X-all"; "gfx1031" = "gfx103X-all"
"gfx1030" = "gfx103X-all"
"gfx90a" = "gfx90a"; "gfx908" = "gfx908" # MI200/MI100
}
# gfx120X (RDNA 4) and gfx1151/gfx1150 (Strix) have a null-pointer bug in

7
studio/install_python_stack.py
View file

@ -352,6 +352,13 @@ _GFX_TO_AMD_INDEX_ARCH: dict[str, str] = {
"gfx1102": "gfx110X-all", # RDNA 3
"gfx1101": "gfx110X-all",
"gfx1100": "gfx110X-all",
"gfx1036": "gfx103X-all",
"gfx1035": "gfx103X-all", # RDNA 2 (RX 6000)
"gfx1034": "gfx103X-all",
"gfx1033": "gfx103X-all",
"gfx1032": "gfx103X-all",
"gfx1031": "gfx103X-all",
"gfx1030": "gfx103X-all",
"gfx90a": "gfx90a",
"gfx908": "gfx908", # MI200/MI100
}

7
studio/setup.ps1
View file

@ -2774,6 +2774,7 @@ if ((Test-Path -LiteralPath $VenvDir -PathType Container) -and -not $NoTorchMode
"gfx1201", "gfx1200", # RDNA 4
"gfx1151", "gfx1150", # RDNA 3.5 (Strix Halo/Point)
"gfx1103", "gfx1102", "gfx1101", "gfx1100", # RDNA 3
"gfx1036", "gfx1035", "gfx1034", "gfx1033", "gfx1032", "gfx1031", "gfx1030", # RDNA 2 (RX 6000)
"gfx90a", "gfx908" # MI200 / MI100
)
if ($script:ROCmGfxArch -and ($_rocmWheelArches -contains $script:ROCmGfxArch)) {
@ -3064,6 +3065,10 @@ if (-not $TorchIndexPinned -and ($HasROCm -or $ROCmGfxArch) -and $CuTag -eq "cpu
"gfx1151" = "gfx1151"; "gfx1150" = "gfx1150" # RDNA 3.5 (Strix Halo/Point)
"gfx1103" = "gfx110X-all"; "gfx1102" = "gfx110X-all" # RDNA 3
"gfx1101" = "gfx110X-all"; "gfx1100" = "gfx110X-all"
"gfx1036" = "gfx103X-all"; "gfx1035" = "gfx103X-all" # RDNA 2 (RX 6000)
"gfx1034" = "gfx103X-all"; "gfx1033" = "gfx103X-all"
"gfx1032" = "gfx103X-all"; "gfx1031" = "gfx103X-all"
"gfx1030" = "gfx103X-all"
"gfx90a" = "gfx90a"; "gfx908" = "gfx908" # MI200/MI100
}
# gfx120X and Strix have a null _grouped_mm kernel on torch <2.11.0.
@ -3098,7 +3103,7 @@ if (-not $TorchIndexPinned -and ($HasROCm -or $ROCmGfxArch) -and $CuTag -eq "cpu
# GPU arch detected but not in the supported wheel map — warn explicitly
# so the user knows why they are getting CPU PyTorch instead of ROCm.
substep "[WARN] AMD GPU ($ROCmGfxArch) not in supported arch list -- falling back to CPU-only PyTorch" "Yellow"
substep " Supported: gfx1200/1201 (RDNA 4), gfx1150/1151 (RDNA 3.5), gfx1100-1103 (RDNA 3), gfx90a, gfx908" "Yellow"
substep " Supported: gfx1200/1201 (RDNA 4), gfx1150/1151 (RDNA 3.5), gfx1100-1103 (RDNA 3), gfx1030-1036 (RDNA 2), gfx90a, gfx908" "Yellow"
} else {
# HIP SDK present ($HasROCm=true via amd-smi) but gcnArchName was not
# readable — warn rather than silently falling back to CPU PyTorch.