Marktechpost launches GPU Leaderboard and Deploy It Chrome extension
The captured page gives no features, pricing, or availability beyond the launch title.
Stories per day · 30 days
7 RSS
Relatedinference10nvidia8data-center2homomorphic-encryption2rust2simulation25g1agentic-ai1akamai1anomalyco1apple1apple-silicon1bare-metal1benchmark1bf161bit-flip1blackwell1chrome-extension1
67 stories, newest first · one card per story
Stories tagged #gpu
Open guide maps platform engineering skills to inference serving
Figures are dated October 2026 and cover vLLM, SGLang, KV cache, and a four-layer serving stack.
5G CATS certifies AI accelerator location with an in-package 5G radio
Unloaded-cell simulations cut 99.9th-percentile RTT spread to 2.60 ms and reject every 300 km fiber-relay round.
E-means claims much faster k-means at equal or better loss
Authors report beating Lloyd loss in 1.2 to 3 epochs and 10–45× GPU speedups versus Flash-KMeans.
GPU CKKS path speeds encrypted MNIST training about 49 times
In a matched 100-update run, device-resident bootstrapping cut training from 14,268.295 to 293.075 seconds with equal decoded parameters.
Carousel scores a billion encrypted vectors without client index state
On BIGANN SIFT1B it reports recall@100 of 0.9715; 100 million vectors take 3.735 seconds on one GPU.
Rune Rust UI framework for GPU rendering across platforms
New Rust-based UI framework Rune enables native GPU rendering for macOS, Windows, Linux and web.
GPU-optimized PlonK analyzes performance bottlenecks in ZK proof generation
The paper explores parallelization trade-offs on multi-GPU environments using RDMA to maximize performance for large circuit sizes in zero-knowledge proofs.
Nvidia pulls RTX 5080 after 5090 removal, prices rise to 2.2X MSRP
Nvidia quietly removed the RTX 5080 from its store after pulling the RTX 5090, causing street prices to climb to 2.2 times MSRP.
Hacker News users discuss hardware for running local LLMs
Community shares Mac Studios, RTX 3090 systems, and GPU clusters for models like DeepSeek-V4-Flash
Qwen3.5-2B Compiled into qwenGPU with 45 CUDA Layers
The small model now runs 45 layers natively on GPU using CUDA.
Akamai says it is scaling cloud and Blackwell GPU capacity
The October 29 update outlines GPU capacity growth and moving real-time data closer to AI applications.
GearUp hides serverless ML loading using DAG slack time
On NVIDIA L40S GPUs, median latency falls 64.5% and 75.9% versus SMIless on two workflows.
MatAnyone-E cuts 4K video-matting cost with sparse memory matching
Authors report about 7.4x faster inference and peak GPU memory at 7.6% of MatAnyone 2 at 4K.
GPU-unlock measures FP/BF16 rates on Nvidia 3000 cards
The item is only a Hacker News link with 44 points and 25 comments, with no technical write-up.
EXL says the serving harness, not the model, sets AI value
Sponsored piece says latency, concurrency, GPU memory, and cost should drive architecture more than scores.
Rowhammer bit-flip in Triton slows or corrupts LLM inference
LLVMhammer raised single-GPU latency 4-16% on four models and could cut BERTScore to 25%.
NVIDIA Nsight Python tutorial covers profiling GPU kernels
The captured page is a cookie notice, so no profiling steps or measurements are available.
NVIDIA adds Confidential Bare Metal for model builders
Cohere plans to use it in Model Vault, and Google Cloud is extending it to A4X Max on GB300 NVL72.
NVIDIA cuDSS generally available for GPU sparse-system solves
Vendors report speedups from 4x to 10x in chip, plasma, and structural simulation tools that already use cuDSS.
Google Cloud ships llm-d 1.0 with cache-aware routing
The captured page is a cookie notice, so routing, TPU, and GPU claims are not confirmed in the body.
7
SF Compute CEO discusses a market for programmable GPU clusters
Evan Conrad says SF Compute now operates much of a cloud so traded GPU contracts settle as usable compute.
1
Token-Energy Law ties LLM decode joules to active bytes and time
Held-out energy estimates on Hopper and Blackwell stayed within 3.73% of measured means.
Specter speeds single-GPU MoE inference via speculative expert offloading
A quantized on-device draft model predicts future experts, cutting time per output token by 1.50x on average.