flagos-ai/FlagTree

View on GitHub

FlagTree is a unified compiler supporting multiple AI chip backends for custom Deep Learning operations, which is forked from triton-lang/triton.

Stars: 350
Forks: 149
Open beginner issues: 0
Indexed issues: 35
Avg merge: 2d 4h
Merged PRs (30d): 87

Dominant language: Python
License: MIT
Last GitHub push: Sep 15, 2026
Latest indexed: Sep 16, 2026
Contributing guide: Contributing guide
Code of conduct: Code of conduct
Beginner labels: No beginner labels indexed

Open All

32 open issues indexed

[Bug]MetaX在运行FlagGems时coredump，原因镜像中有环境vllm_metax 0.15.0+g24fb31.d20260310.maca3.5.3.20.torch2.8 Open

bug metax P0

flagos-ai/FlagTree#606 · 1 comment · 0 reactions · 1 assignee · Indexed Sep 16, 2026
[FR] [Bug] vllm-plugin-FL fails to load with FlagTree on MUSA S5000: triton_v3.2.x reports 0 active drivers, main/triton 3.6 also fails Open

bug mthreads

flagos-ai/FlagTree#773 · 8 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[BUILD] Packaging follow-ups: Ubuntu 22.04 compat, rpath audit, per-branch backends, vendor runtime deps Open

flagos-ai/FlagTree#798 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[Bug Report] HSACOError & FileNotFoundError when compiling Triton kernel on Hygon BW1000 HCU Open

hcu

flagos-ai/FlagTree#820 · 2 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[Kunlunxin] 昆仑芯XPU 使用flagtree 编译Triton Attention 后端报错了 Open

flagos-ai/FlagTree#922 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[Kunlunxin] 昆仑芯使用FlageTree Triton3.0 在测试vllm Triton attention时也会报错 Open

flagos-ai/FlagTree#935 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[Bug][ppu] not supported float8_e4m3fn on ppu backend Open

ppu

flagos-ai/FlagTree#975 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[BUG][mthreads] llc non-termination on giant IR: FFT N=250 naive DFT baseline compile hangs >15min Open

mthreads

flagos-ai/FlagTree#984 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[Bug][npu] not supported float8_e4m3fn on ascend npu Open

ascend

flagos-ai/FlagTree#987 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[BUG][ppu] fp8e4nv is gated by an NVIDIA compute-capability threshold, at both the frontend and in ConvertTritonGPUToLLVMPPU Open

ppu

flagos-ai/FlagTree#1005 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[BUG][mthreads]kernel crashes on device illegal memory access + core dump Open

mthreads

flagos-ai/FlagTree#1010 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[BUG][TLE.DSA][ascend] wrong result when replace tle.dsa.extract_slice with tle.dsa.extract_element Open

ascend

flagos-ai/FlagTree#1012 · 3 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
fix(nvidia): nvidia driver is_active() fires on Ascend NPU under torch_npu TORCH_TRANSFER_TO_NPU=1 Open

ascend

flagos-ai/FlagTree#1022 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[BUG][ppu] tl.atomic_add(..., sem="acquire") causing cross-block spin-barrier deadlock on ppu Open

ppu

flagos-ai/FlagTree#1029 · 2 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[BUG][hygon] tl.load(..., volatile=True) causing cross-block spin-barrier deadlock on hygon Open

hcu

flagos-ai/FlagTree#1030 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Ascend backend: decouple torch_npu so PrivateUse1 plugins can use it Open

ascend

flagos-ai/FlagTree#1046 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[BUG][TLE][PPU] PPU0010 async AIU load rejects i32 operands and lacks B32 lowering Open

flagos-ai/FlagTree#1050 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[BUG][hygon] Scalar atomic_add(..., 0) poll in a while-loop causing cross-block spin-barrier deadlock on hcu Open

flagos-ai/FlagTree#1051 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[XPU] tl.store 在向量计算之后写入标量时，实际写入的是过期的寄存器值 Open

flagos-ai/FlagTree#1056 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[KunlunXin] repeat_interleave_self_int 在总元素数超过 4M 时产生错误结果 Open

flagos-ai/FlagTree#1057 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[BUILD] Release tags never trigger build-deb/build-rpm: tags are pushed with GITHUB_TOKEN by the delivery workflows Open

flagos-ai/FlagTree#1062 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Help, please! Source code compile. Open

flagos-ai/FlagTree#1088 · 2 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[Bugs][Iluvatar] vllm 在安装了tree的环境中启动时 triton报错了 Open

iluvatar

flagos-ai/FlagTree#1093 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[DOC][TLE.DSA][ascend] tle.dsa.local_ptr is documented in tle.md but absent from every release Open

flagos-ai/FlagTree#1111 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 16, 2026
【FR】Bug: FlagTree Ascend backend fails to lower vLLM apply_top_k_top_p_triton (TritonIR→LinalgIR, () → tensor<1xf32> cannot be legalized) Open

flagos-ai/FlagTree#1117 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[BUG][ascend] uint16 and uint32 `>>` are lowered as arithmetic shifts, silently corrupting every value with the top bit set Open

flagos-ai/FlagTree#1121 · 3 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
XPU Compiler: OutOfResources on GDN fused_post_conv_kernel (Qwen3.6) Open

bug flagos2.2-rc0 P0 xpu

flagos-ai/FlagTree#1126 · 1 comment · 0 reactions · 2 assignees · Indexed Sep 16, 2026
[iluvatar] runtime path imports pytest: triton/spec/iluvatar/triton/testing.py::nvsmi reaches triton._internal_testing Open

flagos-ai/FlagTree#1142 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[Bug][metax][TLE] smem support is incomplete: scalar-indexed `local_ptr` asserts at LoadStoreOpToLLVM.cpp:427, smem atomics are unsupported, and smem+histogram silently corrupts data Open

metax

flagos-ai/FlagTree#1152 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[BUG] Incorrect matmul results with num_warps=8: missing cross-warp-group barrier after WGMMA wait Open

flagos-ai/FlagTree#1180 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Improve integer divisibility specialization to enable matmul `cp.async` pipelining Open

flagos-ai/FlagTree#1182 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[MThreads][TLE] Reusing a TME descriptor across K iterations crashes llc: cannot allocate physical SH register for TME operand Open

flagos-ai/FlagTree#1183 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026

flagos-ai/FlagTree

Get new issues in your inbox