heggenhougen

5 ms·

1.
by heggenhougen 7mo ago
For anyone who wants to stress-test the claims — the verifier at rolv.ai runs on any hardware in about 10 minutes and generates your own SHA-256 hash. That's the most direct way to engage with the results. I'm here all day to answ
2.
Show HN: 1k desktop beats vendor sparse library 474× on Mistral-7B(rolv.ai)
1 points by heggenhougen 7mo ago|1 comments
3.
by heggenhougen 7mo ago
Methodology and reproducibility details: All benchmarks were run on the same machine: HP All‑in‑One, Intel i7‑1165G7 (4 cores), 64 GB RAM. All tests use identical inputs, identical weights, identical precision, and identical batch size. Den
4.
LLM FFN benchmarks on a 4‑core HP All‑in‑One(rolv.ai)
1 points by heggenhougen 7mo ago|2 comments
5.
by heggenhougen 7mo ago
I ran four feed‑forward network (FFN) layers from real LLMs on a consumer HP All‑in‑One PC (Intel i7‑1165G7, 4 cores, 64 GB RAM). Each test compares: vendor dense baseline vendor sparse baseline a custom sparse operator identical inputs ide
6.
Benchmarking rolvsparse on DeepSeek-R1 and Llama 4 – up to 82x vs. cuBLAS(rolv.ai)
2 points by heggenhougen 8mo ago|1 comments
7.
by heggenhougen 8mo ago
We have been running a sparse matrix library called rolvsparse on real model weights downloaded directly from HuggingFace and measuring throughput and energy against cuBLAS on an NVIDIA B200. Here are the results across five models so far.
8.
by heggenhougen 8mo ago
Happy to answer questions. Quick note on methodology: the TFLOPS figure is effective (computed as if doing the full dense multiply) — ROLV doesn't violate hardware peak, it avoids work entirely via structured sparsity. Weights are pull
9.
Show HN: ROLV – 20x faster MoE FFN inference on Llama 4 Maverick vs. cuBLAS(rolv.ai)
1 points by heggenhougen 8mo ago|1 comments
10.
ROLV: 50.5× speedup and 91.4% energy savings Qwen2.5-72B MoE expert FFN (1xB200)(rolv.ai)
1 points by heggenhougen 8mo ago|1 comments
11.
by heggenhougen 8mo ago
We just ran the exact expert FFN slice from Qwen2.5-72B-Instruct (8192×28672, batch 512) on a single NVIDIA B200. Results (ROLV vs vendor-best cuBLAS): - Speedup : 50.5× (4953% faster) - Energy Savings : 91.4% - Tokens/s :
12.
by heggenhougen 8mo ago
Almost unbelievable: === FINITE ELEMENT SOLVER FOR MOBILE PHONE DESIGN @ 80% SPARSITY === Stiffness matrix: 8192 × 8192 Sparsity : 80% Solver calls : 100000 Mode : Multi-CPU (optimized for Intel Intel(R) Xeon(R) CPU @ 2.20
13.
100% secure way to test proprietary sparse AI kernel (ROLV) on any hardware(img1.wsimg.com)
1 points by heggenhougen 8mo ago|1 comments
14.
by heggenhougen 8mo ago
100% secure way to test proprietary sparse AI kernel (ROLV) on any hardware We just released a public Verification Kit that lets anyone test ROLV on *their own hardware* — laptop, on-prem cluster, H100, MI300X, custom ASIC, phone, whatever
15.
by heggenhougen 8mo ago
I’m on the Rolv.ai team — below is a detailed expansion on our benchmarks and what to inspect in the PDF. This focuses on artifact parity, methodology, standout results, and the exact verification checks reviewers can run to validate our c
16.
ROLV – Beats Vendor Kernels, Cross‑Platform(img1.wsimg.com)
1 points by heggenhougen 8mo ago|4 comments
17.
by heggenhougen 8mo ago
rolv.ai publishing reproducible benchmarks showing that ROLV outperforms vendor dense kernels and vendor sparse libraries across GPUs, CPUs, and TPUs while producing the same hash/artifact on every platform. Highlights: commodity Intel
18.
by heggenhougen 8mo ago
ROLV is not an optimization, a kernel or a library. It is a new compute primitive—a universal sparse operator that works across GPUs, TPUs, CPUs, mobile SoCs, and next-generation accelerators. ROLV.ai produces identical normalized outputs a
19.
Rolv.ai: Universal sparse compute primitive, back end-agnostic reproducibility(rolv.ai)
1 points by heggenhougen 8mo ago|0 comments
20.
The ROLV CPU Breakthrough(rolv.ai)
4 points by heggenhougen 8mo ago|2 comments
21.
by heggenhougen 8mo ago
Independent validation confirms that ROLV running on commodity CPU systems (Intel Xeon or AMD EPYC) outperforms every major accelerator platform without ROLV — including leading GPUs and TPUs — across the entire sparsity spectrum from 0% to