curated landing page

.vscode/settings.json

@@ -55,6 +55,7 @@
     "pageleave",
     "postbuild",
     "posthog",
+    "powerx",
     "protobufjs",
     "qwen",
     "renderable",
@@ -65,6 +66,7 @@
     "smallint",
     "smallserial",
     "timestamptz",
+    "togglable",
     "tput",
     "trtllm",
     "ttft",

README.md

@@ -6,7 +6,7 @@ A [Next.js](https://nextjs.org) dashboard for visualizing ML inference benchmark
 
 ## Overview
 
-LLM inference performance is a major concern of providing AI services, but accurate performance analysis remains elusive. Fast cadence of software development and model releases makes comparing performance between setups difficult. Existing performance benchmarks quickly become obsolete due to being static, and participants game the benchmarks with unrealistic, highly specific configurations. InferenceX (formerly InferenceMAX) tackles these issues by benchmarking popular models on major hardware platforms nightly with the latest software. For each model and hardware combination, InferenceX sweeps through different tensor parallel sizes and max concurrent requests, showing the throughput vs. latency graph for the full picture. In terms of software configurations, we ensure they are generally applicable across different serving scenarios, and we open source the repo to welcome community contributions. We hope InferenceX informs the community up-to-date and realistic LLM inference performance.
+LLM inference performance is a major concern of providing AI services, but accurate performance analysis remains elusive. Fast cadence of software development and model releases makes comparing performance between setups difficult. Existing performance benchmarks quickly become obsolete due to being static, and participants game the benchmarks with unrealistic, highly specific configurations. InferenceX tackles these issues by benchmarking popular models on major hardware platforms nightly with the latest software. For each model and hardware combination, InferenceX sweeps through different tensor parallel sizes and max concurrent requests, showing the throughput vs. latency graph for the full picture. In terms of software configurations, we ensure they are generally applicable across different serving scenarios, and we open source the repo to welcome community contributions. We hope InferenceX informs the community up-to-date and realistic LLM inference performance.
 
 ## Architecture
 

packages/app/cypress/component/chart-legend.cy.tsx

@@ -58,9 +58,11 @@ function ChartLegendWrapper({ items = MOCK_ITEMS }: { items?: CommonLegendItemPr
       isLegendExpanded={expanded}
       onExpandedChange={setExpanded}
       variant="sidebar"
-      showResetFilter={itemsWithHandler.some((i) => !i.isActive)}
-      allSelected={itemsWithHandler.every((i) => i.isActive)}
-      onResetFilter={() => setLegendItems(items)}
+      actions={
+        itemsWithHandler.some((i) => !i.isActive)
+          ? [{ id: 'reset-filter', label: 'Reset filter', onClick: () => setLegendItems(items) }]
+          : []
+      }
     />
   );
 }

packages/app/cypress/support/mock-data.ts

@@ -34,50 +34,43 @@ export function createMockHardwareConfig(): HardwareConfig {
       label: 'H100',
       suffix: '',
       gpu: "NVIDIA 'Hopper' H100",
-      color: 'var(--h100)',
     },
     h200: {
       name: 'h200',
       label: 'H200',
       suffix: '',
       gpu: "NVIDIA 'Hopper' H200",
-      color: 'var(--h200)',
     },
     b200: {
       name: 'b200',
       label: 'B200',
       suffix: '',
       gpu: "NVIDIA 'Blackwell' B200",
-      color: 'var(--b200)',
     },
     b200_trt: {
       name: 'b200-trt',
       label: 'B200',
       suffix: '(TRT)',
       gpu: "NVIDIA 'Blackwell' B200 TRT",
-      color: 'var(--b200-trt)',
     },
     mi300x: {
       name: 'mi300x',
       label: 'MI300X',
       suffix: '',
       gpu: 'AMD Instinct MI300X',
-      color: 'var(--mi300x)',
     },
     h100_vllm: {
       name: 'h100-vllm',
       label: 'H100',
       suffix: '(vLLM)',
       gpu: "NVIDIA 'Hopper' H100 vLLM",
-      color: 'var(--h100-vllm)',
       framework: 'vllm',
     },
     b200_sglang: {
       name: 'b200-sglang',
       label: 'B200',
       suffix: '(SGLang)',
       gpu: "NVIDIA 'Blackwell' B200 SGLang",
-      color: 'var(--b200-sglang)',
       framework: 'sglang',
     },
   };
@@ -170,8 +163,10 @@ export function createMockInferenceContext(
     activeHwTypes: new Set(['h100', 'b200', 'b200_trt', 'mi300x', 'h200']),
     hwTypesWithData: new Set(['h100', 'b200', 'b200_trt', 'mi300x', 'h200']),
     toggleHwType: namedStub('toggleHwType'),
+    removeHwType: namedStub('removeHwType'),
     selectAllHwTypes: namedStub('selectAllHwTypes'),
     toggleActiveDate: namedStub('toggleActiveDate'),
+    removeActiveDate: namedStub('removeActiveDate'),
     selectAllActiveDates: namedStub('selectAllActiveDates'),
     activeDates: new Set(['2025-03-01']),
     hardwareConfig: hwConfig,
@@ -310,6 +305,7 @@ export function createMockEvaluationContext(
     unfilteredChartData: [createMockEvaluationChartData()],
     enabledHardware: new Set(['b200_trt', 'h100', 'mi300x']),
     toggleHardware: namedStub('toggleHardware'),
+    removeHardware: namedStub('removeHardware'),
     highContrast: false,
     setHighContrast: namedStub('setHighContrast_eval'),
     showLabels: true,
@@ -373,6 +369,7 @@ export function createMockReliabilityContext(
     setHighContrast: namedStub('setHighContrast_reliability'),
     enabledModels: new Set([Model.DeepSeek_R1]),
     toggleModel: namedStub('toggleModel'),
+    removeModel: namedStub('removeModel'),
     isLegendExpanded: true,
     setIsLegendExpanded: namedStub('setIsLegendExpanded_reliability'),
     modelsWithData: new Set([Model.DeepSeek_R1]),

packages/app/package.json

@@ -2,7 +2,7 @@
   "name": "@semianalysisai/inferencex-app",
   "version": "0.1.0",
   "private": true,
-  "description": "InferenceX (formerly InferenceMAX) - Inference performance benchmarking and visualization",
+  "description": "InferenceX - Inference performance benchmarking and visualization",
   "repository": {
     "type": "git",
     "url": "https://github.com/SemiAnalysisAI/inferencemax-app"

packages/app/src/app/(dashboard)/calculator/page.tsx

@@ -0,0 +1,10 @@
+import type { Metadata } from 'next';
+
+import ThroughputCalculatorDisplay from '@/components/calculator/ThroughputCalculatorDisplay';
+import { tabMetadata } from '@/lib/tab-meta';
+
+export const metadata: Metadata = tabMetadata('calculator');
+
+export default function CalculatorPage() {
+  return <ThroughputCalculatorDisplay />;
+}

packages/app/src/app/(dashboard)/evaluation/page.tsx

@@ -0,0 +1,15 @@
+import type { Metadata } from 'next';
+
+import { EvaluationProvider } from '@/components/evaluation/EvaluationContext';
+import EvaluationChartDisplay from '@/components/evaluation/ui/ChartDisplay';
+import { tabMetadata } from '@/lib/tab-meta';
+
+export const metadata: Metadata = tabMetadata('evaluation');
+
+export default function EvaluationPage() {
+  return (
+    <EvaluationProvider>
+      <EvaluationChartDisplay />
+    </EvaluationProvider>
+  );
+}

packages/app/src/app/(dashboard)/gpu-metrics/page.tsx

@@ -0,0 +1,10 @@
+import type { Metadata } from 'next';
+
+import GpuMetricsDisplay from '@/components/gpu-power/GpuPowerDisplay';
+import { tabMetadata } from '@/lib/tab-meta';
+
+export const metadata: Metadata = tabMetadata('gpu-metrics');
+
+export default function GpuMetricsPage() {
+  return <GpuMetricsDisplay />;
+}

packages/app/src/app/(dashboard)/gpu-specs/page.tsx

@@ -0,0 +1,10 @@
+import type { Metadata } from 'next';
+
+import { GpuSpecsContent } from '@/components/gpu-specs/gpu-specs-content';
+import { tabMetadata } from '@/lib/tab-meta';
+
+export const metadata: Metadata = tabMetadata('gpu-specs');
+
+export default function GpuSpecsPage() {
+  return <GpuSpecsContent />;
+}

packages/app/src/app/(dashboard)/historical/page.tsx

@@ -0,0 +1,15 @@
+import type { Metadata } from 'next';
+
+import { InferenceProvider } from '@/components/inference/InferenceContext';
+import HistoricalTrendsDisplay from '@/components/trends/HistoricalTrendsDisplay';
+import { tabMetadata } from '@/lib/tab-meta';
+
+export const metadata: Metadata = tabMetadata('historical');
+
+export default function HistoricalPage() {
+  return (
+    <InferenceProvider activeTab="historical">
+      <HistoricalTrendsDisplay />
+    </InferenceProvider>
+  );
+}

packages/app/src/app/(dashboard)/inference/page.tsx

@@ -0,0 +1,15 @@
+import type { Metadata } from 'next';
+
+import { InferenceProvider } from '@/components/inference/InferenceContext';
+import InferenceChartDisplay from '@/components/inference/ui/ChartDisplay';
+import { tabMetadata } from '@/lib/tab-meta';
+
+export const metadata: Metadata = tabMetadata('inference');
+
+export default function InferencePage() {
+  return (
+    <InferenceProvider activeTab="inference">
+      <InferenceChartDisplay />
+    </InferenceProvider>
+  );
+}

packages/app/src/app/(dashboard)/layout.tsx

@@ -0,0 +1,5 @@
+import { DashboardShell } from '@/components/dashboard-shell';
+
+export default function DashboardLayout({ children }: { children: React.ReactNode }) {
+  return <DashboardShell>{children}</DashboardShell>;
+}