ClickHouseInstallation YAML file Лицензия: Apache 2.0, Распространяется как Docker image ClickHouse cluster resources kubectl apply K8S API Спецификация ClickHouse on Kubernetes – components Shard 1 Replica com/Altinity/clickhouse-operator/master/manifests/oper ator/clickhouse-operator-install.yaml Single-node “cluster” apiVersion: "clickhouse.altinity.com/v1" kind: "ClickHouseInstallation" metadata: name: STATUS pod/chi-demo-01-demo-0-0-0 1/1 Running NAME TYPE CLUSTER-IP service/chi-demo-01-demo-0-0 ClusterIP None service/clickhouse-demo-01 LoadBalancer

block_size>>>(N, x, y); // … } Ok, about the kernel parameters 10 |Memory“In a typical PC or cluster node today, the memories of the CPU and GPU are physically distinct and separated by the PCI-Express GPU, bridging the CPU-GPU divide. Managed memory is accessible to both the CPU and GPU using a single pointer. The key is that the system automatically migrates data allocated in Unified Memory between and GPU, bridging the CPU-GPU divide. Managed memory is accessible to both the CPU and GPU using a single pointer. The key is that the system automatically migrates data allocated in Unified Memory between

kubernetes cluster in large scale Patrickxie ( 谢谆志） Background Cloud has been the general trend. How to manage so many clusters ,resources and businesses How to ensure load balancing of cluster nodes resources and businesses Cluster1 Cluster2 Cluster2 … Region N … Cluster1 Cluster2 … Cluster2 Business 1 Business 2 Business 3 Business N … How to ensure load balancing of cluster nodes ？ Dynamic-Scheduler Controller of native K8S use only one goroutine to handle the scaling of all businesses in the cluster, and personalization configuration is not supported. Worker HPA1 HPA2 … HPAn HPA1 HPAPlus-controller

tlsCertificate.certificateChain.i nlineBytes' -r | base64 -d | openssl x509 -noout -text -in - ● Part of cluster config in envoy config-dump ○ kubectl exec -c istio-proxy curl localhost:15000/config_dump selector: matchLabels: app: details rules: - from: - source: principals: ["cluster.local/ns/default/sa/bookinfo-productpage"] to: - operation: methods: ["GET"]" apiVersion: selector: matchLabels: app: reviews rules: - from: - source: principals: ["cluster.local/ns/default/sa/bookinfo-productpage"] to: - operation: methods: ["GET"]" apiVersion:

pods via VIP using a load balancer • Two types • ClusterIP provides in-cluster access • NodePort provides out-of-cluster access • Major modes • Iptables • IPVS Iptables mode • How it works configure may differ • Performance of a cluster in different time slot may differ • Due to CPU oversold • Suggestion: • Run the test against the same cluster during near time • Make CPU the bottleneck

BEGINNING … WAS GOTO STRUCTURED PROGRAMMING Structured control flow constructs have a single entry and a single exit, permitting them to be treated like a black box. goto is an unstructured control */ }60 COROUTINES ARE STRUCTURED CONCURRENCY coroutine Awaiting a coroutine has a single entry and a single exit. // This is a coroutine: task compute_helper_async(); task compute_async() move(rec)}; } }; keyclick_sender read_keyclick() { return {}; }88 // Operation state for a single key click: template Rec> struct keyclick_operation : pending_completion

第一个 s 表示标量 (scalar) ，只对 xmm 的最低位进行运算；也可以是 p 表示矢量 (packed) ，一次对 xmm 中所有位进行运算。 3. 第二个 s 表示单精度浮点数 (single) ，即 float 类型；也可以是 d 表示双精度浮点数 (double) ，即 double 类型。 • addss ：一个 float 加法。 • addsd ：一个 double 加法。 %xmm1, %xmm0 addps %xmm1, %xmm0 xmm0 xmm1 xmm0 为什么需要 SIMD ？单个指令处理四个数据 • 这种单个指令处理多个数据的技术称为 SIMD （ single-instruction multiple-data ）。 • 他可以大大增加计算密集型程序的吞吐量。 • 因为 SIMD 把 4 个 float 打包到一个 xmm 寄存器里同时运算，很像数学中矢量的逐元 器 他可以一次存储 4 个 int ，或 4 个 float movups ： move unaligned packed single u 代表 (%rdi) 的地址不一定对齐到 16 字 节 movaps ： move aligned packed single SIMD 指令：敢不敢再宽一点？ 为什么编译器没有用 256 位的 ymm0 ？ 因为他不敢保证运行这个程序的电脑支持 AVX

multi- platform applications from a single codebase. - Single codebase - Hot reload - Rendering Engine 利用 Rust 重塑移动应用开发 跨平台开发的优势和局限性 Pros: - Fast - Single Codebase - Third-party support (Javascript

data is held in different cache lines. On the Intel Xeon Phi coprocessor, a single cache line is 64 bytes, enough to hold 16 single precision floating point values. But what happens when the data is not contiguous When using this ordering, for example in a 16 x 16 matrix, each of the 4 sub matrices can fit into a single cache line. For example, when transposing a simple matrix, if each of the sub-matrices are transposed

No Silver Bullet – Essence and Accident in Software Engineering Mike Ballou“There is no single development, in either technology or management technique, which by itself promises even one order-of- not build, software • Great designersRequirements refinement and rapid prototyping • The hardest single part of building a software system is deciding precisely what to build • The most important function