---
title: "Performance"
description: "817 posts about Performance, summarised, each linking to the original."
---

# Performance
> 817 posts about Performance, summarised, each linking to the original.

## Articles

### [Speculation Is All You Need](https://yomu.fyi/post/speculation-is-all-you-need.md)
- Company: [Modal](https://yomu.fyi/company/modal.md)
- Published: Jun 19, 2026
- Type: Explainer

Speculative decoding is presented as the most important optimization for LLM inference, delivering speedups of 2x-3x rather than small percentage gains from kernel tuning. The technique accelerates the decode phase by using a lightweight draft model to generate candidate tokens that the target model processes in parallel, with lossless resampling. The authors released DFlash speculators for multiple Qwen models, achieving an additional 5-20% speedup over existing baselines, and worked with SGLang to integrate them. On a B200 node, the Qwen 3.5 122B-A10B model with speculation reaches over 1000 tok/s at concurrency 1, compared to 250 tok/s without speculation, demonstrating a 4x speedup. Training these speculators is simplified because the data-generating process is the target model itself, making data collection trivial, and the post envisions a flywheel where custom speculators and distillation enable continuous improvement in latency, cost, and quality for self-hosted inference.


### [Beyond the warehouse: How METRO Markets built a do-it-all data platform on ClickHouse Cloud](https://yomu.fyi/post/beyond-the-warehouse-how-metro-markets-built-a-do-it-all-data-platform.md)
- Company: [Clickhouse](https://yomu.fyi/company/clickhouse.md)
- Published: Jun 18, 2026
- Type: Problem & solution

METRO Markets outgrew its self-hosted Impala and Kudu-based Hadoop stack as the marketplace expanded, with parallel stores, inconsistent reporting, and too much maintenance burdening a lean data team. They evaluated BigQuery, Snowflake, and ClickHouse against production queries, finding faster ClickHouse performance, including one response falling from four seconds to under 200 milliseconds, while its pricing supported more predictable costs. Airbyte moves CDC data from MySQL and Postgres into ClickHouse; ReplacingMergeTree handles deduplication during background merges, and BigQuery-hosted Google Analytics data enters through cloud storage and external tables. The resulting warehouse consolidates company data and supports seller analytics, credit risk features through Query API endpoints, observability, and emerging AI use cases, while teams query a shared dataset through compatible tools; planned Kafka and Flink integration would extend it to streaming workloads.


### [Why everyone is talking about real-time analytics (a note from “the yellow company”)](https://yomu.fyi/post/why-everyone-is-talking-about-real-time-analytics-a-note-from-the-yell.md)
- Company: [Clickhouse](https://yomu.fyi/company/clickhouse.md)
- Published: Jun 18, 2026
- Type: Explainer

Real-time analytics is presented as a foundational workload for customer-facing analytics, observability, fraud detection, operational applications, and AI agents rather than a niche use case. The discussion argues that performance claims require open, transparent, reproducible benchmarks that disclose methodology, datasets, queries, results, and operating conditions, including continuous ingestion, concurrency, memory pressure, and fresh data. A real-time database should support queryable data within milliseconds to a couple of seconds, incremental transformations, bounded tail latency as writes increase, configurable read-time versus write-time work, unified support for text, vector, JSON, and structured analytics, and efficiency as workload scales. The article connects renewed attention to agentic workloads, whose continuous, high-volume querying makes low latency, concurrency, and fresh context increasingly necessary, and calls for shared standards beyond a single latency number.


### [Appcues delivers personalized customer engagement with ClickHouse Cloud](https://yomu.fyi/post/appcues-delivers-personalized-customer-engagement-with-clickhouse-clou.md)
- Company: [Clickhouse](https://yomu.fyi/company/clickhouse.md)
- Published: Jun 18, 2026
- Type: Problem & solution

Appcues replaced an analytics stack that included Snowflake and Airflow after rising scale pushed queries to 20 seconds or more and limited reports to 10-minute windows. The company evaluated ClickHouse Cloud on AWS for real-time segmentation, exporting Snowflake data to Parquet in S3 and testing complex email segments with nested conditions, historical events, and negative logic. Production events now pass through Kafka and a clickwriter consumer, while a clickclient library built on Elixir’s Ecto generates controlled ClickHouse queries; SharedReplacingMergeTree, table engines, and materialized views were tuned during a one-year migration. Across 1.31 PB and 410 billion events, P95 queries fell below two seconds, ingestion latency dropped to about five seconds, and analytics costs decreased 23%, while new funnel reports and AI-powered insights were added.


### [What's new in Postgres Managed by ClickHouse: RBAC, Terraform, ClickPipes, extensions, and more](https://yomu.fyi/post/what-s-new-in-postgres-managed-by-clickhouse-rbac-terraform-clickpipes.md)
- Company: [Clickhouse](https://yomu.fyi/company/clickhouse.md)
- Published: Jun 18, 2026
- Type: Announcement

Postgres managed by ClickHouse, which entered public beta in May, received a broad update covering console operations, per-service RBAC, Terraform, Stripe provisioning, ClickPipes migrations, billing, capacity reservations, and extensions. The redesigned console adds service health, resource, and query views, while RBAC provides View and Manage roles across the console and public REST API; the alpha Terraform provider covers lifecycle operations including resizing, HA, replicas, restore, and password rotation. Postgres-to-Postgres ClickPipes now automate schema migration, usage is integrated into ClickHouse Cloud’s metering and billing, and AWS capacity reservations are allocated dynamically across availability zones. Underneath, pg\_clickhouse and pg\_stat\_ch moved to the header-only clickhouse-c library, removing a C++ exception conflict, using Postgres memory contexts, and reducing the x86-64 Linux extension binary from 4.9 MB to 1.4 MB; the update shipped less than a month after beta, with Terraform and extensions still evolving.


### [Beyond LoRA: Can you beat the most popular fine-tuning technique?](https://yomu.fyi/post/beyond-lora-can-you-beat-the-most-popular-fine-tuning-technique.md)
- Company: [Hugging Face](https://yomu.fyi/company/hugging-face.md)
- Author: Benjamin Bossan, Sayak Paul, Marian Tietz, Kashif Rasul
- Published: Jun 18, 2026
- Type: Explainer

Parameter-efficient fine-tuning (PEFT) reduces the memory required to adapt open models, enabling fine-tuning with a fraction of the memory and, in some cases, quantized models. LoRA is dominant because it freezes base weights and trains a small set of added parameters, but its popularity may also reflect ecosystem momentum and uneven comparisons. The post presents shared-condition benchmarks for language-model math reasoning and image generation, using identical models, datasets, code, hardware, and measures including test performance, VRAM, forgetting or drift, runtime, and checkpoint size. The results show that LoRA performs well but is not consistently best: other techniques can outperform it on one or more axes, especially in the image-generation benchmark. The recommended choice is therefore to compare techniques against a use case’s needs and constraints, including quantization support, layer coverage, adapter merging, and downstream serving compatibility.


### [Is it agentic enough? Benchmarking open models on your own tooling](https://yomu.fyi/post/is-it-agentic-enough-benchmarking-open-models-on-your-own-tooling.md)
- Company: [Hugging Face](https://yomu.fyi/company/hugging-face.md)
- Author: Lysandre, Nathan Habib, Pedro Cuenca
- Published: Jun 18, 2026
- Type: Benchmark

Final-answer checks can show whether an agent completed a task, but they miss the turns, tokens, seconds, errors, and command path required. The proposed tool-specific benchmark measures that work on transformers, comparing models, revisions, tasks, and three access tiers: bare installation, source clone, and a packaged Skill containing CLI documentation and examples. It uses deterministic tasks with exact matching, launches one Hugging Face Job per model–revision–task combination on identical hardware, and records traces in a Hugging Face Bucket. Reports include match rate, median new, cached, and generated tokens, time, error rate, and marker adoption, while traces expose the agent's commands. On transformers, the CLI plus Skill helped large open models but caused Qwen3-14B to misinterpret documentation as an executable tool, dropping classify-sentiment match from 100% to 0% in the cited comparison.


### [June 2026 newsletter](https://yomu.fyi/post/june-2026-newsletter.md)
- Company: [Clickhouse](https://yomu.fyi/company/clickhouse.md)
- Published: Jun 18, 2026
- Type: Announcement

ClickHouse’s June 2026 newsletter marks the project’s ten years as open source, noting more than 2,000 contributors and community projects alongside engineering updates. It profiles Vasily Chekalkin’s 26.5 improvements to WebAssembly user-defined functions, including the DETERMINISTIC keyword and visibility in system.functions, and covers native Jaeger storage, the ClickCannon benchmark tool, and Open House announcements. The 26.5 release reports Top-N optimizations that make ORDER BY … LIMIT through JOIN 20.4× faster with 175× less memory on TPC-H SF100, while GROUP BY … LIMIT is 11.9× faster with 185× less peak memory. Other featured systems include a 15-million-rows-per-hour KPI pipeline, join-engineering work yielding a 26× improvement on join-heavy TPC-H workloads, and community guidance on inserts, columnar storage, ArrowStream, and RAG ingestion.


### [92x faster queries: How Open Electricity uses ClickHouse to track Australia’s energy transition in real time](https://yomu.fyi/post/92x-faster-queries-how-open-electricity-uses-clickhouse-to-track-austr.md)
- Company: [Clickhouse](https://yomu.fyi/company/clickhouse.md)
- Published: Jun 17, 2026
- Type: Problem & solution

Open Electricity uses ClickHouse to serve live Australian energy data through a public platform and API spanning roughly one billion records. Migrating from Postgres and TimescaleDB addressed queries that could take more than five minutes, reducing a benchmark to just over three seconds and to 60 milliseconds with a materialized view. The new setup keeps heavy relational workloads in Postgres while ClickHouse handles analytics, using ReplacingMergeTree base tables and more than 20 materialized views for different aggregation patterns. It also uses substantially less hardware than the previous configuration: four cores, four GB RAM, and 200 GB disk versus 16 cores, 32 GB RAM, and 1.6 TB, while serving data added daily and API users.


### [GLM-5.2: Built for Long-Horizon Tasks](https://yomu.fyi/post/glm-5-2-built-for-long-horizon-tasks.md)
- Company: [Hugging Face](https://yomu.fyi/company/hugging-face.md)
- Author: Z.AI
- Published: Jun 17, 2026
- Type: Announcement

GLM-5.2 is presented as a flagship model for long-horizon tasks, extending GLM-5.1’s capability to a solid 1M-token context for sustained engineering work. It was trained extensively on 1M-context coding-agent scenarios involving large-scale implementation, automated research, performance optimization, and complex debugging, and adds effort-level controls to balance capability, latency, and computational cost. Its architecture uses IndexShare, sharing one lightweight indexer across every four transformer layers and reducing per-token FLOPs by 2.9× at 1M context; MTP changes for speculative decoding increase acceptance length by up to 20%. On reported long-horizon and standard coding benchmarks, GLM-5.2 is described as the highest-ranked open-source model: it trails Opus 4.8 by 1% on FrontierSWE, ranks second only to Opus 4.8 on PostTrainBench, and scores 81.0 on Terminal-Bench 2.1 versus GLM-5.1’s 63.5.


### [Build Your Own Transaction Foundation Model for Financial Intelligence](https://yomu.fyi/post/build-your-own-transaction-foundation-model-for-financial-intelligence.md)
- Company: [NVIDIA](https://yomu.fyi/company/nvidia-developer-blog.md)
- Author: Benjamin Wu
- Published: Jun 16, 2026
- Type: Tutorial

Production financial intelligence systems frequently rely on hand-engineered tabular features and rule sets that are brittle, expensive to maintain, and blind to historical sequential patterns. NVIDIA demonstrates an accelerated reference pipeline that replaces standard BPE tokenization with a GPU-based domain tokenizer, converting raw transactions into semantic tokens with an 8,192-token context window. Using the NeMo AutoModel library, a compact 29-million parameter decoder-only transformer is pretrained from scratch on unlabeled transaction sequences using causal language modeling. Learned sequence embeddings are extracted, compressed using PCA, and concatenated with raw tabular features to train a downstream GPU-accelerated XGBoost fraud detection model. On the IBM TabFormer benchmark dataset, this combined approach achieves a 41.76% lift in Average Precision over the baseline model relying solely on raw tabular features.


### [NVIDIA Blackwell Tops MLPerf Training 6.0 with Industry-Leading Scale and Performance](https://yomu.fyi/post/nvidia-blackwell-tops-mlperf-training-6-0-with-industry-leading-scale.md)
- Company: [NVIDIA](https://yomu.fyi/company/nvidia-developer-blog.md)
- Author: Farshad Ghodsian
- Published: Jun 16, 2026
- Type: Benchmark

NVIDIA submitted results across all MLPerf Training v6.0 benchmarks, setting performance records on workloads including the DeepSeek-V3 and GPT-OSS-20B Mixture of Experts models. To scale training up to 8,192 Blackwell GPUs, NVIDIA combined hardware cluster designs with scale-out networking via Spectrum-X Ethernet and Quantum InfiniBand. The engineering team eliminated CPU-GPU synchronization bottlenecks in token-dropless MoEs by implementing full-iteration CUDA graphs with synchronization-free operators and paged stashing. Further software optimizations included CuTe DSL kernel fusions, an MXFP8 attention block in the Transformer Engine, and transitioning router elementwise math to FP32. Across these benchmarks, the GB300 NVL72 platform achieved the fastest time to train at scale and the highest normalized per-accelerator performance.


### [Build On-Device AI Companions with the NVIDIA ACE Game Agent SDK and Unreal Engine 5 Plugins](https://yomu.fyi/post/build-on-device-ai-companions-with-the-nvidia-ace-game-agent-sdk-and-u.md)
- Company: [NVIDIA](https://yomu.fyi/company/nvidia-developer-blog.md)
- Author: Phillip Singh
- Published: Jun 16, 2026
- Type: Announcement

Game developers integrating on-device artificial intelligence companions face challenges such as conversational latency, state synchronization, and loop prevention. NVIDIA announced the open-source ACE Game Agent SDK alongside a suite of Unreal Engine 5 plugins to streamline native, hardware-accelerated character pipelines. The SDK offers Agent, Chat, and RAG APIs, enabling characters to execute multi-step tool-assisted reasoning and query game databases. Unreal Engine 5 plugins introduce local runtime models including nemo-conformer-ctc-120m for speech recognition, Qwen 3.5 4B for language generation, and Chatterbox Turbo 350M for speech synthesis. Additional tooling expands to motion generation via Animotive Kimodo and rendering updates with the DLSS 4.5 Unreal Engine plugin.


### [How to Optimize Transformer-Based Models for Low-Precision Training](https://yomu.fyi/post/how-to-optimize-transformer-based-models-for-low-precision-training.md)
- Company: [NVIDIA](https://yomu.fyi/company/nvidia-developer-blog.md)
- Author: Jonathan Mitchell
- Published: Jun 16, 2026
- Type: Benchmark

Accelerating transformer training requires understanding how low-precision formats such as FP8 and NVFP4 impact specific general matrix multiplication (GEMM) workloads. Transformer configurations do not explicitly reveal these shapes, making microbenchmarks necessary before committing to full training runs. NVIDIA Transformer Engine enables quantization and kernel dispatch across precisions, which can be evaluated in realistic autocast mode or kernel-only prequantized mode. Profiling the ESM2-15B model on NVIDIA B300 GPUs demonstrated that NVFP4 achieved a 1.79x blended forward propagation speedup over MXFP8 and up to 4.01x over BF16 in prequantized execution. Although large GEMM dimensions successfully overcome quantization overheads, dynamic scaling, Hadamard transforms, and kernel selection asymmetries moderate real-world gains.


### [Introducing Lakehouse//RT: Real-Time Performance on a Unified Lakehouse](https://yomu.fyi/post/introducing-lakehouse-rt-real-time-performance-on-a-unified-lakehouse.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Nong Li, Shoumik Palkar, Shant Hovsepian, Mostafa Mokhtar, Reynold Xin
- Published: Jun 16, 2026
- Type: Announcement

Databricks introduces Lakehouse//RT, a real-time data warehouse designed for operational analytics, BI, app serving, and observability workloads. It is powered by Reyden and is intended to deliver millisecond performance directly on lakehouse data without copying it into a separate serving layer. Preview participants saw up to 16x better performance, with response times as low as 10ms on smaller datasets, sub-100ms on larger ones, and sub-100ms latency at 12,000 queries per second on standard analytical benchmarks. Tests covering concurrency, dataset scale, and complex TPCDS queries report low latency where alternatives slowed or failed. Lakehouse//RT is in Beta for select read-only workloads, with incremental autoscaling and automatic baseline compute sizing.


### [What a difference 10 years of open source makes](https://yomu.fyi/post/what-a-difference-10-years-of-open-source-makes.md)
- Company: [Clickhouse](https://yomu.fyi/company/clickhouse.md)
- Published: Jun 16, 2026
- Type: Explainer

ClickHouse marks ten years since becoming open source under the Apache 2.0 license, crediting more than 2,600 contributors and a decade of monthly releases for its growth into one of the world’s most widely deployed analytical databases. Recent engineering work spans joins, updates and deletes, JSON, full-text search, vector search, query acceleration, integrations, and developer tooling. Join-focused changes made the TPC-H SF100 workload 26× faster between versions 22.4 and 26.4, while statistics-based reordering reduced one six-table query from 3,903.7 seconds to 2.7 seconds. The project also publishes open benchmarks, including ClickBench, JSONBench, CostBench, and PostgresBench, with scripts, configurations, data, and results intended for independent validation. Its next phase continues to emphasize community contributions and broader ways to use ClickHouse.


### [Ten years of ClickHouse in open source](https://yomu.fyi/post/ten-years-of-clickhouse-in-open-source.md)
- Company: [Clickhouse](https://yomu.fyi/company/clickhouse.md)
- Published: Jun 15, 2026
- Type: Explainer

ClickHouse’s tenth-anniversary history describes its path from an internal web-analytics experiment to an open-source analytical database with more than 2,000 contributors. The original problem was processing rapidly growing, real-time logs while answering flexible reports quickly; tested MySQL extensions and standalone analytical databases did not meet the stated scale of 100 billion records per day and 500 columns. The project built column-oriented storage, lightweight compression, custom aggregation structures, background merges, and query support for grouping, aggregation, filtering, and sorting. Released publicly on June 15, 2016, ClickHouse is presented as a level-three open-source project with public contribution guidelines, code review, testing, CI, documentation, and a roadmap, while contributors receive credit even when implementations are substantially rewritten.


### [Cisco Talos empowers threat researchers while reducing TCO by 75% with ClickHouse Cloud](https://yomu.fyi/post/cisco-talos-empowers-threat-researchers-while-reducing-tco-by-75-with.md)
- Company: [Clickhouse](https://yomu.fyi/company/clickhouse.md)
- Published: Jun 15, 2026
- Type: Problem & solution

Cisco Talos operates a threat-intelligence reputation service that classifies file hashes from Cisco product telemetry, enabling researchers to make security decisions about benign and potentially malicious artifacts. Previously, a custom pipeline moved data from Databricks Delta and S3 through EKS, where staging and aggregated tables supported an API; the self-managed design retained about a petabyte of compressed, replicated data on 32 high-compute nodes. The team rebuilt the system in four phases, using Terraform, SSO/SAML, sharded ClickPipes for a 12-month backfill, shadow traffic, and a validated cutover to ClickHouse Cloud on AWS. The resulting architecture uses EventBridge, SQS, ClickPipes, SharedMergeTree, object storage, and AWS PrivateLink, delivering less than a minute of queue lag, zero data loss, a 100% response match across more than 1,000 hashes, zero downtime, roughly 90% lower storage costs, and about 75% lower TCO while reclaiming over 300 engineer hours annually.


### [ClickHouse achieves AWS Retail Competency](https://yomu.fyi/post/clickhouse-achieves-aws-retail-competency.md)
- Company: [Clickhouse](https://yomu.fyi/company/clickhouse.md)
- Published: Jun 12, 2026
- Type: Announcement

ClickHouse has achieved the AWS Retail Competency in the Advanced Data Insights category, a designation recognizing validated technical expertise and demonstrated customer success in retail real-time analytics. Retail workloads described here combine high-volume, high-cardinality, high-concurrency data with hard freshness requirements, including inventory, pricing, personalization, and peak seasonal demand. The post attributes the platform's fit to compressed columnar storage, parallel queries, continuous Kafka and Amazon S3 ingestion through ClickPipes, compute-compute separation, and AWS support for private networking, multi-region deployment, and Marketplace billing. Examples include Mercado Libre's 90-TB-per-day observability ingestion and trace lookups reduced from 15–20 minutes to about two seconds, while other customers report cost, speed, and analyst-efficiency gains.


### [Better, faster, less wrong: Enhancing issue grouping](https://yomu.fyi/post/better-faster-less-wrong-enhancing-issue-grouping.md)
- Company: [Sentry](https://yomu.fyi/company/sentry.md)
- Author: Kush Dubey, Yuval Mandelboum
- Published: Jun 12, 2026
- Type: Problem & solution

Sentry upgraded its AI-driven issue grouping system to better prevent duplicate issues without merging distinct application errors. The team trained lightonai/modernbert-embed-large using Matryoshka Representation Learning on hundreds of thousands of stacktrace pairs labeled by Claude Sonnet 4.5. To optimize the high-throughput ingestion pipeline, embeddings were truncated from 768 to 64 dimensions, combined with bfloat16 precision, PyTorch SDPA, and CUDA graph compilation. A phased live rollout used threshold-gated index backfilling and automatic fallback to prevent issue spikes during migration. In production, the v2 model cuts the overgrouping rate from 8% to 4%, increases prevented duplicate issues to 70%, and delivers 6x faster inference.


[Newer posts](https://yomu.fyi/topic/performance/page/15.md) · [Older posts](https://yomu.fyi/topic/performance/page/17.md)