Owner

beehive-lab

11 indexed repositories · View on GitHub

TornadoVM
Write Java. Run on GPUs. Fast.

Java · 1507 stars
mambo
A low-overhead dynamic binary instrumentation and modification tool for ARM (both AArch32 and AArch64 support) and RISC-V (RV64GC).

C · 398 stars
Maxine-VM
Maxine VM: A meta-circular research VM

Java · 341 stars
jllm
AI inference engine for Java in Java. Fast.

Java · 276 stars
FastPath_MP
FastPath_MP: An FPGA-based multi-path architecture for direct access from FPGA to NVMe SSD

C · 37 stars
docker-tornadovm
Docker build scripts for TornadoVM on GPUs: https://github.com/beehive-lab/TornadoVM

Shell · 32 stars
beehive-spirv-toolkit
Prototype for a SPIR-V assembler and dissasembler. It provides a composable Java interface for generating SPIR-V code at runtime.

Java · 16 stars
levelzero-jni
Intel LevelZero JNI library for TornadoVM

Java · 15 stars
docker-gpullama3.java
Docker build scripts for running GPULlama3.java on GPUs with TornadoVM

Shell · 3 stars
DFLOW
A framework for rapidly prototyping data DFLOW proof-of-concepts

C++ · 2 stars
pie
Instruction encoder / decoder generator

Ruby · 2 stars

Beginner friendly Everything

Open All

[CUDA] getHalfFloatValue() on a local HalfFloat[] element: ClassCastException HalfFloatStamp -> AbstractPointerStamp (silently bails to sequential) Open

Difficulty 4/5 3-5 days Newbie friendliness 48/100

beehive-lab/TornadoVM#1097 · Opened Sep 18, 2026
[Qwen3.8] Staged Transfers Open

beehive-lab/jllm#162 · 1 assignee · Opened Sep 16, 2026
[Libraries] [Qwen3-0.6B-fp16] Prefill Open

beehive-lab/jllm#158 · 1 comment · 1 assignee · Opened Sep 16, 2026
Gemma 4 Performance Open

Difficulty 5/5 Over a week Newbie friendliness 15/100

beehive-lab/jllm#157 · Opened Sep 16, 2026
[JIT] [Qwen3.8] Prefill Open

beehive-lab/jllm#154 · 1 assignee · Opened Sep 16, 2026
Qwen 3.8 27B Q4_0 performance Open

Difficulty 5/5 Over a week Newbie friendliness 30/100

beehive-lab/jllm#153 · Opened Sep 16, 2026
[cuda][profiler] Dispatch times are structurally 0 and copy-out time is never recorded, failing three non-whitelisted TestProfiler tests Open

beehive-lab/TornadoVM#1071 · 1 comment · 1 assignee · Opened Sep 4, 2026
[CUDA] Lowering fails with "Node implementing Lowerable not handled: NewInstance" when a ternary precedes an allocation (Math.max works) Open

bug compiler

beehive-lab/TornadoVM#1064 · 1 assignee · Opened Sep 2, 2026
[CUDA] CuDnn.sdpaForward launches no kernel: silently returns an all-zero result (shipped BenchmarkSdpa fails on 6.0.0) Open

bug CUDA

Difficulty 4/5 3-5 days Newbie friendliness 52/100

beehive-lab/TornadoVM#1063 · Opened Sep 2, 2026
TransformerKernelsTest#testReductionOneBlockWithLayer is non-deterministic: inter-workgroup race in the test kernel Open

Difficulty 4/5 3-5 days Newbie friendliness 55/100

beehive-lab/TornadoVM#1059 · Opened Aug 29, 2026
TornadoParallelScheduler: incorrect thread offset computation in scheduling for non-zero loop start with non-unit stride Open

Difficulty 3/5 1-2 days Newbie friendliness 78/100

beehive-lab/TornadoVM#1044 · 1 comment · Opened Aug 25, 2026
[bug] A batched plan with more than one output fails with a null device buffer at LAUNCH Open

bug

beehive-lab/TornadoVM#1034 · 1 assignee · Opened Aug 6, 2026
[perf] Host-side runtime overhead: where the time actually goes, the fixes, and what is left Open

beehive-lab/TornadoVM#1028 · 6 comments · 5 assignees · Opened Aug 6, 2026
Support the Qwen3.5 architecture Open

beehive-lab/jllm#143 · 1 comment · 1 assignee · Opened Aug 2, 2026
Reduce external implementation types exposed by the GPULlama3 integration API Open

1.0.1 good first issue refactoring

Difficulty 5/5 Over a week Newbie friendliness 35/100

beehive-lab/jllm#141 · Opened Jul 30, 2026
CUDA graph replay fails with CUresult=700 when a captured graph's buffers are re-pointed, and the failure is silent Open

Difficulty 4/5 3-5 days Newbie friendliness 45/100

beehive-lab/TornadoVM#1006 · Opened Jul 30, 2026
CUDA backend feature roadmap: KernelContext and ExecutionPlan coverage, gaps and priorities Open

Difficulty 5/5 Over a week Newbie friendliness 30/100

beehive-lab/TornadoVM#990 · Opened Jul 26, 2026
Auto-tuning of thread configuration via TornadoExecutionPlan.withAutoTune() Open

Difficulty 5/5 Over a week Newbie friendliness 35/100

beehive-lab/TornadoVM#979 · 1 comment · Opened Jul 24, 2026
TornadoVM 6.0.0 Roadmap: legacy backends & JVMCI removals, and NVIDIA broader library support Open

6.0.0

beehive-lab/TornadoVM#957 · 5 assignees · Opened Jul 20, 2026
[roadmap] NVIDIA ecosystem alignment based on findings from #953 Open

beehive-lab/TornadoVM#954 · 1 assignee · Opened Jul 20, 2026
Deprecate string-keyed consumeFromDevice/persistOnDevice in favour of residency-by-object-identity + async handles Open

beehive-lab/TornadoVM#942 · 3 comments · 3 assignees · Opened Jul 15, 2026
Async execution ergonomics: non-blocking executeAsync() + composable execution handles Open

CUDA

beehive-lab/TornadoVM#941 · 3 comments · 3 assignees · Opened Jul 15, 2026
CUDA/kernel backend: stronger compile-time verification of KernelContext kernels Open

CUDA

beehive-lab/TornadoVM#939 · 1 assignee · Opened Jul 15, 2026
CUDA backend: early returns in a helper inlined into a @Parallel loop mis-lower Open

Difficulty 4/5 3-5 days Newbie friendliness 48/100

beehive-lab/TornadoVM#931 · Opened Jul 14, 2026
CUDA backend: two loops in separate if/else branches merge incorrectly (phi lowering) Open

CUDA

beehive-lab/TornadoVM#930 · 1 assignee · Opened Jul 14, 2026
include RELEASE file in sdk distro Open

beehive-lab/TornadoVM#922 · 1 comment · 1 assignee · Opened Jul 13, 2026
Roadmap: vLLM-style serving features for TornadoVM batched decode Open

beehive-lab/jllm#130 · 1 comment · 4 assignees · Opened Jul 11, 2026
Metal backend reports grossly inflated GPU/dispatch timing metrics (TOTAL_KERNEL_TIME, TOTAL_DISPATCH_KERNEL_TIME, COPY_IN_TIME, COPY_OUT_TIME) despite comparable or better end-to-end execution time Open

beehive-lab/TornadoVM#905 · 2 comments · 1 assignee · Opened Jul 9, 2026
Support direct execution of K-quant GGUF weights on GPU (Q4_K, Q5_K, Q6_K) Open

enhancement quantization

Difficulty 4/5 3-5 days Newbie friendliness 48/100

beehive-lab/jllm#118 · Opened May 29, 2026
Invalid brackets in code generation for OpenCL and Metal backends Open

bug compiler metal OpenCL

beehive-lab/TornadoVM#837 · 2 assignees · Opened May 22, 2026
tornado --devices fails: libtornado-objc-metal.dylib” Not Opened Open

beehive-lab/TornadoVM#826 · 2 comments · 2 assignees · Opened Apr 10, 2026
TornadoVM hangs infinitely in TornadoSketcher.lookup when using Aviator Expression Engine Open

beehive-lab/TornadoVM#791 · 3 comments · 1 assignee · Opened Jan 29, 2026
Proposal: Same variable in different class Open

Difficulty 5/5 Over a week Newbie friendliness 35/100

beehive-lab/TornadoVM#784 · Opened Jan 20, 2026
[POC] Minimal Kernel Autotuning Support Open

enhancement

beehive-lab/jllm#96 · 1 assignee · Opened Dec 22, 2025
Proposal: Implement Triton-style Autotuning Support in TornadoVM Open

API performance runtime

beehive-lab/TornadoVM#778 · 1 assignee · Opened Dec 22, 2025
NullPointerException when using persistOnDevice + consumeFromDevice across multiple TaskGraphs in a single TornadoExecutionPlan Open

Difficulty 4/5 3-5 days Newbie friendliness 45/100

beehive-lab/TornadoVM#768 · 2 comments · Opened Dec 13, 2025
Sealed types cannot be instantiated. in Kotlin Open

beehive-lab/TornadoVM#766 · 1 comment · 1 assignee · Opened Dec 13, 2025
[CI] Add GA to enable automatic version changes on TornadoVM release Open

beehive-lab/docker-tornadovm#10 · 2 assignees · Opened Dec 9, 2025
RISC-V Vector Extension Support Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

beehive-lab/mambo#139 · 2 comments · Opened Nov 27, 2025
TornadoTaskRuntimeException when using Phi-3-mini-4k-instruct-fp16.gguf during TornadoVM initialization Open

bug

beehive-lab/jllm#72 · 5 comments · 1 assignee · Opened Nov 25, 2025
[bug] `WithThreadInfo` does not work with `@Parallel` tasks Open

bug

beehive-lab/TornadoVM#735 · 1 assignee · Opened Nov 13, 2025
[proposal] Multithreaded executions with different inputs Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

beehive-lab/TornadoVM#729 · Opened Oct 24, 2025
Different result for CPU and GPU Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

beehive-lab/TornadoVM#728 · 9 comments · Opened Oct 22, 2025
[proposal] One kernel for multiple Tasks Open

question

beehive-lab/TornadoVM#724 · 5 comments · 1 assignee · Opened Oct 3, 2025
[ERROR] OpenCL JNI Library not found - Windows 11 Open

Difficulty 4/5 3-5 days Newbie friendliness 28/100

beehive-lab/TornadoVM#707 · 5 comments · 1 reaction · Opened Aug 18, 2025
support for llama3.2 vision Open

question

Difficulty 5/5 Over a week Newbie friendliness 25/100

beehive-lab/jllm#43 · 1 comment · Opened Aug 5, 2025
OutOfMemory / Misleading Hint Open

documentation enhancement

beehive-lab/jllm#30 · 1 comment · 1 reaction · 1 assignee · Opened Jun 19, 2025
Add pre-commit hook Open

good first issue Tooling

Difficulty 3/5 1-2 days Newbie friendliness 38/100

beehive-lab/jllm#29 · Opened Jun 19, 2025
Support Images for GPULlama3.java Open

beehive-lab/docker-gpullama3.java#1 · 1 assignee · Opened Jun 18, 2025
Build GUI Chatbox for GPULlama3.java Inference with embedded GPU resource monitoring Open

good first issue Tooling

Difficulty 5/5 Over a week Newbie friendliness 28/100

beehive-lab/jllm#24 · 2 comments · Opened Jun 13, 2025

Showing the newest 100

This page lists what was indexed most recently. Advanced filter has the whole inventory, narrowed by language, difficulty and how long a task takes.

Open advanced filter

beehive-lab

Showing the newest 100

Get new issues in your inbox