mirror of
https://github.com/ClusterCockpit/cc-backend
synced 2024-12-26 13:29:05 +01:00
Change to prod data, allow and handle null data
- fix errors regarding render timing - always collect time info in transFormData function - remove size from polar plot
This commit is contained in:
Christoph Kluge
parent
b449b77b95
commit
1b8c4e293c
@ -4,6 +4,7 @@
|
||||
groupByScope,
|
||||
fetchMetricsStore,
|
||||
checkMetricDisabled,
|
||||
transformDataForRoofline
|
||||
} from "./utils.js";
|
||||
import {
|
||||
Row,
|
||||
@ -130,8 +131,8 @@
|
||||
lazyFetchMoreMetrics();
|
||||
|
||||
let plots = {},
|
||||
roofWidth,
|
||||
jobTags,
|
||||
fullWidth,
|
||||
statsTable;
|
||||
$: document.title = $initq.fetching
|
||||
? "Loading..."
|
||||
@ -190,7 +191,6 @@
|
||||
}));
|
||||
</script>
|
||||
|
||||
<div class="row" bind:clientWidth={fullWidth} />
|
||||
<Row>
|
||||
<Col>
|
||||
{#if $initq.error}
|
||||
@ -245,7 +245,6 @@
|
||||
{/if}
|
||||
<Col>
|
||||
<Polar
|
||||
size={fullWidth / 4.1}
|
||||
metrics={ccconfig[
|
||||
`job_view_polarPlotMetrics:${$initq.data.job.cluster}`
|
||||
] || ccconfig[`job_view_polarPlotMetrics`]}
|
||||
@ -254,21 +253,24 @@
|
||||
/>
|
||||
</Col>
|
||||
<Col>
|
||||
<Roofline
|
||||
width={fullWidth / 3 - 10}
|
||||
height={fullWidth / 5}
|
||||
cluster={clusters
|
||||
.find((c) => c.name == $initq.data.job.cluster)
|
||||
.subClusters.find(
|
||||
(sc) => sc.name == $initq.data.job.subCluster
|
||||
)}
|
||||
flopsAny={$jobMetrics.data.jobMetrics.find(
|
||||
(m) => m.name == "flops_any" && m.scope == "node"
|
||||
)}
|
||||
memBw={$jobMetrics.data.jobMetrics.find(
|
||||
(m) => m.name == "mem_bw" && m.scope == "node"
|
||||
)}
|
||||
/>
|
||||
<div bind:clientWidth={roofWidth}>
|
||||
<Roofline
|
||||
width={roofWidth - 10}
|
||||
height={(roofWidth / 2) - 5}
|
||||
renderTime={true}
|
||||
cluster={clusters
|
||||
.find((c) => c.name == $initq.data.job.cluster)
|
||||
.subClusters.find(
|
||||
(sc) => sc.name == $initq.data.job.subCluster
|
||||
)}
|
||||
data={
|
||||
transformDataForRoofline (
|
||||
$jobMetrics.data.jobMetrics.find((m) => m.name == "flops_any" && m.scope == "node").metric,
|
||||
$jobMetrics.data.jobMetrics.find((m) => m.name == "mem_bw" && m.scope == "node").metric
|
||||
)
|
||||
}
|
||||
/>
|
||||
</div>
|
||||
</Col>
|
||||
{:else}
|
||||
<Col />
|
||||
|
||||
@ -31,8 +31,8 @@
|
||||
export let cluster;
|
||||
|
||||
let plotWidths = [],
|
||||
colWidth1 = 0,
|
||||
colWidth2 = 0
|
||||
colWidth1,
|
||||
colWidth2
|
||||
let from = new Date(Date.now() - 5 * 60 * 1000),
|
||||
to = new Date(Date.now());
|
||||
const topOptions = [
|
||||
@ -429,14 +429,14 @@
|
||||
<Roofline
|
||||
width={plotWidths[i] - 10}
|
||||
height={300}
|
||||
colorDots={true}
|
||||
showTime={false}
|
||||
cluster={subCluster}
|
||||
data={transformPerNodeDataForRoofline(
|
||||
$mainQuery.data.nodeMetrics.filter(
|
||||
(data) => data.subCluster == subCluster.name
|
||||
data={
|
||||
transformPerNodeDataForRoofline(
|
||||
$mainQuery.data.nodeMetrics.filter(
|
||||
(data) => data.subCluster == subCluster.name
|
||||
)
|
||||
)
|
||||
)}
|
||||
}
|
||||
/>
|
||||
{/key}
|
||||
</div>
|
||||
@ -444,7 +444,7 @@
|
||||
</Row>
|
||||
{/each}
|
||||
|
||||
<hr style="margin-top: -1em;" />
|
||||
<hr/>
|
||||
|
||||
<!-- Usage Stats as Histograms -->
|
||||
|
||||
|
||||
@ -22,7 +22,6 @@
|
||||
LineElement
|
||||
);
|
||||
|
||||
export let size
|
||||
export let metrics
|
||||
export let cluster
|
||||
export let jobMetrics
|
||||
@ -95,7 +94,7 @@
|
||||
</script>
|
||||
|
||||
<div class="chart-container">
|
||||
<Radar {data} {options} width={size} height={size}/>
|
||||
<Radar {data} {options}/>
|
||||
</div>
|
||||
|
||||
<style>
|
||||
|
||||
@ -4,99 +4,52 @@
|
||||
import { onMount, onDestroy } from 'svelte'
|
||||
import { Card } from 'sveltestrap'
|
||||
|
||||
export let flopsAny = null
|
||||
export let memBw = null
|
||||
export let maxY = null
|
||||
export let data = null
|
||||
export let renderTime = false
|
||||
export let maxY = null // Optional
|
||||
export let cluster = null
|
||||
export let width = 500
|
||||
export let height = 300
|
||||
export let renderTime = false
|
||||
export let data = null
|
||||
|
||||
let plotWrapper = null
|
||||
let uplot = null
|
||||
let timeoutId = null
|
||||
|
||||
// Three Render-Cases:
|
||||
// #1 Single-Job Roofline -> Has Time-Information: Use data, allow renderTime
|
||||
// #2 MultiNode Roofline - > Has No Time-Information: Transform from nodeData, only "IST"-state of nodes, no timeInfo
|
||||
// #3 Multi-Job Roofline as Heatmap -> Keep Original
|
||||
/* Data Format
|
||||
* data = [null, [], []] // 0: null-axis required for scatter, 1: Array of XY-Array for Scatter, 2: Optional Time Info
|
||||
* data[1][0] = [100, 200, 500, ...] // X Axis -> Intensity (Vals up to clusters' flopRateScalar value)
|
||||
* data[1][1] = [1000, 2000, 1500, ...] // Y Axis -> Performance (Vals up to clusters' flopRateSimd value)
|
||||
* data[2] = [0.1, 0.15, 0.2, ...] // Color Code -> Time Information (Floats from 0 to 1) (Optional)
|
||||
*/
|
||||
|
||||
// Start Demo Data
|
||||
|
||||
function randInt(min, max) {
|
||||
return Math.floor(Math.random() * (max - min + 1)) + min;
|
||||
}
|
||||
|
||||
function randFloat(min, max) {
|
||||
return roundTwo(((Math.random() * (max - min + 1)) + min) / randInt(1, 500));
|
||||
}
|
||||
|
||||
function roundTwo(num) {
|
||||
return Math.round((num + Number.EPSILON) * 100) / 100
|
||||
}
|
||||
|
||||
function filledArr(len, val, time) {
|
||||
let arr = Array(len);
|
||||
|
||||
if (typeof val == "function") {
|
||||
for (let i = 0; i < len; ++i)
|
||||
arr[i] = val(i);
|
||||
}
|
||||
else if (time) {
|
||||
for (let i = 0; i < len; ++i)
|
||||
arr[i] = i / 1000;
|
||||
}
|
||||
else {
|
||||
for (let i = 0; i < len; ++i)
|
||||
arr[i] = i;
|
||||
}
|
||||
|
||||
return arr;
|
||||
}
|
||||
|
||||
let points = 1000;
|
||||
|
||||
data = [null, []] // Null-Axis required for scatter
|
||||
data[1][0] = filledArr(points, i => randFloat(1,5000), false) // Intensity
|
||||
data[1][1] = filledArr(points, i => randFloat(1,5000), false) // Performance
|
||||
// data[1][0] = filledArr(points, 0, false) // Intensity
|
||||
// data[1][1] = filledArr(points, 0, false) // Performance
|
||||
data[2] = filledArr(points, 0, true) // Time Information (Optional)
|
||||
|
||||
// End Demo Data
|
||||
// Check
|
||||
// console.assert(data , "you must provide data")
|
||||
|
||||
// Helpers
|
||||
|
||||
function getGradientR(x) {
|
||||
if (x < 0.5) return 0
|
||||
if (x > 0.75) return 255
|
||||
x = (x - 0.5) * 4.0
|
||||
return Math.floor(x * 255.0)
|
||||
}
|
||||
|
||||
function getGradientG(x) {
|
||||
if (x > 0.25 && x < 0.75) return 255
|
||||
if (x < 0.25) x = x * 4.0
|
||||
else x = 1.0 - (x - 0.75) * 4.0
|
||||
return Math.floor(x * 255.0)
|
||||
}
|
||||
|
||||
function getGradientB(x) {
|
||||
if (x < 0.25) return 255
|
||||
if (x > 0.5) return 0
|
||||
x = 1.0 - (x - 0.25) * 4.0
|
||||
return Math.floor(x * 255.0)
|
||||
}
|
||||
|
||||
function getRGB(c) {
|
||||
return `rgb(${getGradientR(c)}, ${getGradientG(c)}, ${getGradientB(c)})`
|
||||
}
|
||||
|
||||
function nearestThousand (num) {
|
||||
return Math.ceil(num/1000) * 1000
|
||||
}
|
||||
|
||||
function lineIntersect(x1, y1, x2, y2, x3, y3, x4, y4) {
|
||||
let l = (y4 - y3) * (x2 - x1) - (x4 - x3) * (y2 - y1)
|
||||
let a = ((x4 - x3) * (y1 - y3) - (y4 - y3) * (x1 - x3)) / l
|
||||
@ -105,12 +58,11 @@
|
||||
y: y1 + a * (y2 - y1)
|
||||
}
|
||||
}
|
||||
|
||||
// End Helpers
|
||||
|
||||
// Dot Renderers
|
||||
const drawColorPoints = (u, seriesIdx, idx0, idx1) => {
|
||||
const size = 5 * devicePixelRatio;
|
||||
|
||||
uPlot.orient(u, seriesIdx, (series, dataX, dataY, scaleX, scaleY, valToPosX, valToPosY, xOff, yOff, xDim, yDim, moveTo, lineTo, rect, arc) => {
|
||||
let d = u.data[seriesIdx];
|
||||
let deg360 = 2 * Math.PI;
|
||||
@ -136,7 +88,6 @@
|
||||
|
||||
const drawPoints = (u, seriesIdx, idx0, idx1) => {
|
||||
const size = 5 * devicePixelRatio;
|
||||
|
||||
uPlot.orient(u, seriesIdx, (series, dataX, dataY, scaleX, scaleY, valToPosX, valToPosY, xOff, yOff, xDim, yDim, moveTo, lineTo, rect, arc) => {
|
||||
let d = u.data[seriesIdx];
|
||||
u.ctx.strokeStyle = getRGB(0);
|
||||
@ -158,113 +109,121 @@
|
||||
return null;
|
||||
};
|
||||
|
||||
function render() {
|
||||
const opts = {
|
||||
title: "",
|
||||
mode: 2,
|
||||
width: width,
|
||||
height: height,
|
||||
legend: {
|
||||
show: false
|
||||
},
|
||||
cursor: { drag: { x: false, y: false } },
|
||||
axes: [
|
||||
{ label: 'Intensity [FLOPS/Byte]' },
|
||||
{ label: 'Performace [GFLOPS]' }
|
||||
],
|
||||
scales: {
|
||||
x: {
|
||||
time: false,
|
||||
range: [0.01, 1000],
|
||||
distr: 3, // Render as log
|
||||
log: 10, // log exp
|
||||
// Main Function
|
||||
function render(plotData) {
|
||||
if (plotData) {
|
||||
const opts = {
|
||||
title: "",
|
||||
mode: 2,
|
||||
width: width,
|
||||
height: height,
|
||||
legend: {
|
||||
show: false
|
||||
},
|
||||
y: {
|
||||
range: [1.0, nearestThousand(cluster.flopRateSimd.value || maxY)],
|
||||
distr: 3, // Render as log
|
||||
log: 10, // log exp
|
||||
},
|
||||
},
|
||||
series: [
|
||||
{},
|
||||
{ paths: renderTime ? drawColorPoints : drawPoints }
|
||||
],
|
||||
hooks: {
|
||||
drawClear: [
|
||||
u => {
|
||||
u.series.forEach((s, i) => {
|
||||
if (i > 0)
|
||||
s._paths = null;
|
||||
});
|
||||
cursor: { drag: { x: false, y: false } },
|
||||
axes: [
|
||||
{
|
||||
label: 'Intensity [FLOPS/Byte]',
|
||||
values: (u, vals) => vals.map(v => formatNumber(v))
|
||||
},
|
||||
],
|
||||
draw: [
|
||||
u => { // draw roofs when cluster set
|
||||
// console.log(u)
|
||||
if (cluster != null) {
|
||||
const padding = u._padding // [top, right, bottom, left]
|
||||
|
||||
u.ctx.strokeStyle = 'black'
|
||||
u.ctx.lineWidth = 2
|
||||
u.ctx.beginPath()
|
||||
|
||||
const ycut = 0.01 * cluster.memoryBandwidth.value
|
||||
const scalarKnee = (cluster.flopRateScalar.value - ycut) / cluster.memoryBandwidth.value
|
||||
const simdKnee = (cluster.flopRateSimd.value - ycut) / cluster.memoryBandwidth.value
|
||||
const scalarKneeX = u.valToPos(scalarKnee, 'x', true), // Value, axis, toCanvasPixels
|
||||
simdKneeX = u.valToPos(simdKnee, 'x', true),
|
||||
flopRateScalarY = u.valToPos(cluster.flopRateScalar.value, 'y', true),
|
||||
flopRateSimdY = u.valToPos(cluster.flopRateSimd.value, 'y', true)
|
||||
|
||||
if (scalarKneeX < width - padding[1]) { // Top horizontal roofline
|
||||
u.ctx.moveTo(scalarKneeX, flopRateScalarY)
|
||||
u.ctx.lineTo(width - padding[1], flopRateScalarY)
|
||||
}
|
||||
|
||||
if (simdKneeX < width - padding[1]) { // Lower horitontal roofline
|
||||
u.ctx.moveTo(simdKneeX, flopRateSimdY)
|
||||
u.ctx.lineTo(width - padding[1], flopRateSimdY)
|
||||
}
|
||||
|
||||
let x1 = u.valToPos(0.01, 'x', true),
|
||||
y1 = u.valToPos(ycut, 'y', true)
|
||||
|
||||
let x2 = u.valToPos(simdKnee, 'x', true),
|
||||
y2 = flopRateSimdY
|
||||
|
||||
let xAxisIntersect = lineIntersect(
|
||||
x1, y1, x2, y2,
|
||||
u.valToPos(0.01, 'x', true), u.valToPos(1.0, 'y', true), // X-Axis Start Coords
|
||||
u.valToPos(1000, 'x', true), u.valToPos(1.0, 'y', true) // X-Axis End Coords
|
||||
)
|
||||
|
||||
if (xAxisIntersect.x > x1) {
|
||||
x1 = xAxisIntersect.x
|
||||
y1 = xAxisIntersect.y
|
||||
}
|
||||
|
||||
// Diagonal
|
||||
u.ctx.moveTo(x1, y1)
|
||||
u.ctx.lineTo(x2, y2)
|
||||
|
||||
u.ctx.stroke()
|
||||
// Reset grid lineWidth
|
||||
u.ctx.lineWidth = 0.15
|
||||
}
|
||||
{
|
||||
label: 'Performace [GFLOPS]',
|
||||
values: (u, vals) => vals.map(v => formatNumber(v))
|
||||
}
|
||||
]
|
||||
},
|
||||
};
|
||||
],
|
||||
scales: {
|
||||
x: {
|
||||
time: false,
|
||||
range: [0.01, 1000],
|
||||
distr: 3, // Render as log
|
||||
log: 10, // log exp
|
||||
},
|
||||
y: {
|
||||
range: [1.0, nearestThousand(cluster.flopRateSimd.value || maxY)],
|
||||
distr: 3, // Render as log
|
||||
log: 10, // log exp
|
||||
},
|
||||
},
|
||||
series: [
|
||||
{},
|
||||
{ paths: renderTime ? drawColorPoints : drawPoints }
|
||||
],
|
||||
hooks: {
|
||||
drawClear: [
|
||||
u => {
|
||||
u.series.forEach((s, i) => {
|
||||
if (i > 0)
|
||||
s._paths = null;
|
||||
});
|
||||
},
|
||||
],
|
||||
draw: [
|
||||
u => { // draw roofs when cluster set
|
||||
// console.log(u)
|
||||
if (cluster != null) {
|
||||
const padding = u._padding // [top, right, bottom, left]
|
||||
|
||||
uplot = new uPlot(opts, data, plotWrapper);
|
||||
u.ctx.strokeStyle = 'black'
|
||||
u.ctx.lineWidth = 2
|
||||
u.ctx.beginPath()
|
||||
|
||||
const ycut = 0.01 * cluster.memoryBandwidth.value
|
||||
const scalarKnee = (cluster.flopRateScalar.value - ycut) / cluster.memoryBandwidth.value
|
||||
const simdKnee = (cluster.flopRateSimd.value - ycut) / cluster.memoryBandwidth.value
|
||||
const scalarKneeX = u.valToPos(scalarKnee, 'x', true), // Value, axis, toCanvasPixels
|
||||
simdKneeX = u.valToPos(simdKnee, 'x', true),
|
||||
flopRateScalarY = u.valToPos(cluster.flopRateScalar.value, 'y', true),
|
||||
flopRateSimdY = u.valToPos(cluster.flopRateSimd.value, 'y', true)
|
||||
|
||||
if (scalarKneeX < width - padding[1]) { // Top horizontal roofline
|
||||
u.ctx.moveTo(scalarKneeX, flopRateScalarY)
|
||||
u.ctx.lineTo(width - padding[1], flopRateScalarY)
|
||||
}
|
||||
|
||||
if (simdKneeX < width - padding[1]) { // Lower horitontal roofline
|
||||
u.ctx.moveTo(simdKneeX, flopRateSimdY)
|
||||
u.ctx.lineTo(width - padding[1], flopRateSimdY)
|
||||
}
|
||||
|
||||
let x1 = u.valToPos(0.01, 'x', true),
|
||||
y1 = u.valToPos(ycut, 'y', true)
|
||||
|
||||
let x2 = u.valToPos(simdKnee, 'x', true),
|
||||
y2 = flopRateSimdY
|
||||
|
||||
let xAxisIntersect = lineIntersect(
|
||||
x1, y1, x2, y2,
|
||||
u.valToPos(0.01, 'x', true), u.valToPos(1.0, 'y', true), // X-Axis Start Coords
|
||||
u.valToPos(1000, 'x', true), u.valToPos(1.0, 'y', true) // X-Axis End Coords
|
||||
)
|
||||
|
||||
if (xAxisIntersect.x > x1) {
|
||||
x1 = xAxisIntersect.x
|
||||
y1 = xAxisIntersect.y
|
||||
}
|
||||
|
||||
// Diagonal
|
||||
u.ctx.moveTo(x1, y1)
|
||||
u.ctx.lineTo(x2, y2)
|
||||
|
||||
u.ctx.stroke()
|
||||
// Reset grid lineWidth
|
||||
u.ctx.lineWidth = 0.15
|
||||
}
|
||||
}
|
||||
]
|
||||
},
|
||||
};
|
||||
uplot = new uPlot(opts, plotData, plotWrapper);
|
||||
} else {
|
||||
console.log('No data for roofline!')
|
||||
}
|
||||
}
|
||||
|
||||
// Copied from Histogram
|
||||
|
||||
// Svelte and Sizechange
|
||||
onMount(() => {
|
||||
render()
|
||||
render(data)
|
||||
})
|
||||
|
||||
onDestroy(() => {
|
||||
if (uplot)
|
||||
uplot.destroy()
|
||||
@ -272,7 +231,6 @@
|
||||
if (timeoutId != null)
|
||||
clearTimeout(timeoutId)
|
||||
})
|
||||
|
||||
function sizeChanged() {
|
||||
if (timeoutId != null)
|
||||
clearTimeout(timeoutId)
|
||||
@ -281,13 +239,10 @@
|
||||
timeoutId = null
|
||||
if (uplot)
|
||||
uplot.destroy()
|
||||
|
||||
render()
|
||||
render(data)
|
||||
}, 200)
|
||||
}
|
||||
|
||||
$: sizeChanged(width, height)
|
||||
|
||||
</script>
|
||||
|
||||
{#if data != null}
|
||||
|
||||
@ -6,8 +6,8 @@ const power = [1, 1e3, 1e6, 1e9, 1e12, 1e15, 1e18, 1e21]
|
||||
const prefix = ['', 'K', 'M', 'G', 'T', 'P', 'E']
|
||||
|
||||
export function formatNumber(x) {
|
||||
if ( isNaN(x) ) {
|
||||
return x // Return if String , used in Histograms
|
||||
if ( isNaN(x) || x == null) {
|
||||
return x // Return if String or Null
|
||||
} else {
|
||||
for (let i = 0; i < prefix.length; i++)
|
||||
if (power[i] <= x && x < power[i+1])
|
||||
|
||||
@ -6,7 +6,7 @@ import {
|
||||
} from "@urql/svelte";
|
||||
import { setContext, getContext, hasContext, onDestroy, tick } from "svelte";
|
||||
import { readable } from "svelte/store";
|
||||
import { formatNumber } from './units.js'
|
||||
// import { formatNumber } from './units.js'
|
||||
|
||||
/*
|
||||
* Call this function only at component initialization time!
|
||||
@ -326,8 +326,11 @@ export function convert2uplot(canvasData) {
|
||||
}
|
||||
|
||||
export function binsFromFootprint(weights, scope, values, numBins) {
|
||||
let min = 0, max = 0
|
||||
let min = 0, max = 0 //, median = 0
|
||||
if (values.length != 0) {
|
||||
// Extreme, wrong peak vlaues: Filter here or backend?
|
||||
// median = median(values)
|
||||
|
||||
for (let x of values) {
|
||||
min = Math.min(min, x)
|
||||
max = Math.max(max, x)
|
||||
@ -364,11 +367,12 @@ export function binsFromFootprint(weights, scope, values, numBins) {
|
||||
}
|
||||
}
|
||||
|
||||
export function transformDataForRoofline(flopsAny, memBw, renderTime) { // Uses Metric Object
|
||||
export function transformDataForRoofline(flopsAny, memBw) { // Uses Metric Objects: {series:[{},{},...], timestep:60, name:$NAME}
|
||||
const nodes = flopsAny.series.length
|
||||
const timesteps = flopsAny.series[0].data.length
|
||||
|
||||
/* c will contain values from 0 to 1 representing the time */
|
||||
let data = null
|
||||
const x = [], y = [], c = []
|
||||
|
||||
if (flopsAny && memBw) {
|
||||
@ -383,24 +387,23 @@ export function transformDataForRoofline(flopsAny, memBw, renderTime) { // Uses
|
||||
|
||||
x.push(intensity)
|
||||
y.push(f)
|
||||
c.push(renderTime ? j / timesteps : 0)
|
||||
c.push(j / timesteps)
|
||||
}
|
||||
}
|
||||
} else {
|
||||
console.warn("transformData: metrics for 'mem_bw' and/or 'flops_any' missing!")
|
||||
}
|
||||
|
||||
return {
|
||||
x, y, c,
|
||||
xLabel: 'Intensity [FLOPS/byte]',
|
||||
yLabel: 'Performance [GFLOPS]'
|
||||
if (x.length > 0 && y.length > 0 && c.length > 0) {
|
||||
data = [null, [x, y], c] // for dataformat see roofline.svelte
|
||||
}
|
||||
return data
|
||||
}
|
||||
|
||||
// Return something to be plotted. The argument shall be the result of the
|
||||
// `nodeMetrics` GraphQL query.
|
||||
export function transformPerNodeDataForRoofline(nodes) {
|
||||
const x = [], y = [], c = []
|
||||
let data = null
|
||||
const x = [], y = []
|
||||
for (let node of nodes) {
|
||||
let flopsAny = node.metrics.find(m => m.name == 'flops_any' && m.scope == 'node')?.metric
|
||||
let memBw = node.metrics.find(m => m.name == 'mem_bw' && m.scope == 'node')?.metric
|
||||
@ -417,12 +420,21 @@ export function transformPerNodeDataForRoofline(nodes) {
|
||||
|
||||
x.push(intensity)
|
||||
y.push(f)
|
||||
c.push(0)
|
||||
}
|
||||
|
||||
return {
|
||||
x, y, c,
|
||||
xLabel: 'Intensity [FLOPS/byte]',
|
||||
yLabel: 'Performance [GFLOPS]'
|
||||
if (x.length > 0 && y.length > 0) {
|
||||
data = [null, [x, y], []] // for dataformat see roofline.svelte
|
||||
}
|
||||
return data
|
||||
}
|
||||
|
||||
// https://stackoverflow.com/questions/45309447/calculating-median-javascript
|
||||
// function median(numbers) {
|
||||
// const sorted = Array.from(numbers).sort((a, b) => a - b);
|
||||
// const middle = Math.floor(sorted.length / 2);
|
||||
|
||||
// if (sorted.length % 2 === 0) {
|
||||
// return (sorted[middle - 1] + sorted[middle]) / 2;
|
||||
// }
|
||||
|
||||
// return sorted[middle];
|
||||
// }
|
||||
|
||||
Loading…
Reference in New Issue
Block a user