onnxruntime/js/web/lib/wasm/jsep/webgpu/gpu-data-manager.ts
Yulong Wang 45ff957973
1.17.3 cherry-picks for ORT Web changes (#19926)
### Description
This PR is a preview of cherry-picks for ort-web to `rel-1.17.3` based
on `rel-1.17.2`.

<details>

<summary>Changes of ort-web to cherry-pick</summary>

The following commits are from main branch.

`o` stands for pick, and `x` stands for skip.
```
o   2e0a388c36 [js/webgpu] Add HardSigmoid support (#19215)
o   d226e40856 [js/webgpu] set query type in onRunStart (#19202)
o   61610ff986 [js/webgpu] Add FusedConv clip test case (#18900)
o   a33b5bd1fa [JS/WebGPU] Added Uniforms to SkipLayerNorm. (#18788)
o   591f90c0b9 [js/webgpu] Fix issue of timestamp query (#19258)
o   7252c6e747 [WebNN EP] Support WebNN async API with Asyncify (#19145)
o   5b06505073 [js/webgpu] Fix Tanh explosion (#19201)
o   656ca66186 [js/webgpu] Support uniforms for conv, conv transpose, conv grouped (#18753)
o   a3f0e2422b [js/webgpu] Support f16 uniform (#19098)
o   9e69606360 fix f16 for attention, enable slice and flatten for more types (#19262)
o   624b4e2063 [js/webgpu] Remove enableShapesUniforms (#19279)
o   90883a366a [js/webgpu] Add hardSigmoid activation for fusedConv (#19233)
o   85cef0af8c [js/webgpu] Support capture and replay for jsep (#18989)
o   d73131cf0f [js/webgpu] Use DataType as uniform cpu type (#19281)
o   dd1f6ccc45 [js/webgpu] resolve codescan alert (#19343)
o   3a2ab1963a [js/webgpu] Refactor createTensorShapeVariables (#18883)
o   efc17e79de [js/webgpu] Fix the undefined push error (#19366)
 x  50806a7dd5 [js/web] support external data in npm test (#19377)
o   ccbe264a39 [js/webgpu] Add LeakyRelu activation for fusedConv (#19369)
o   5ff27ef02a [js/webgpu] support customop FastGelu (#19392)
 x  03be65e064 [js/web] fix types exports in package.json (#19458)
o   06269a3952 [js/webgpu] allow uint8 tensors for webgpu (#19545)
o   dfeda9019c [JS/WebGPU] Add MatMulNBits (#19446)
o   1b48054e1b [js/webgpu] Create Split indices helpers by rank, not by shape (#19554)
o   3fe2c137ee [js] small fix to workaround formatter (#19400)
 x  70567a4b3a [js/web] use ApiTensor insteadof onnxjs Tensor in TensorResultValidator (#19358)
o   6e04e36e3f [js/common] upgrade tsc in common from 4.9.5 to 5.2.2 (#19317)
o   58f4921686 [js] changes to allow Float16Array if any polyfill is available (#19305)
o   57d6819212 [js/web] Fix fused-conv is not included in npm test (#19581)
o   ebd220b073 Misspelling in README.md (#19433)
o   38c3432393 Bump ip from 1.1.8 to 1.1.9 in /js/react_native (#19582)
o   fe82fccf1a [js/webgpu] Fix Conv2DTransposeMatMul f16 compilation failure (#19596)
o   76a2a487a1 Bump ip from 1.1.8 to 1.1.9 in /js/react_native/e2e (#19583)
o   29b1106033 [node] Switch to setImmediate to avoid starving the Node.js event loop (#19610)
o   ae3d73c981 [JS/WebGPU] Fix Split and Where to handle corner cases. (#19613)
o   aec2389ad0 [js/webgpu] allows a ProgramInfo's RunData to use zero sized output (#19614)
o   bb43a0f133 [js/webgpu] minor fixes to make tinyllama work (#19564)
o   0edb035808 [js/web] fix suite test list for zero sized tensor (#19638)
o   3cb81cdde2 [js/common] move 'env.wasm.trace' to 'env.trace' (#19617)
o   e30618d055 [js/webgpu] use Headless for webgpu test by default (#19702)
o   f06164ef8b [js/web] transfer input buffer back to caller thread (#19677)
 x  a788514027 [js/web] dump debug logs for karma for diagnose purpose (#19785)
o   24b72d2613 [JS/WebGPU] Preserve zero size input tensor dims. (#19737)
o   4538d31a8b [js/webgpu] expose a few properties in WebGPU API (#19857)
o   53de2d8cb0 [js/webgpu] Enable GroupedConvVectorize path (#19791)
o   ed250b88c3 [JS/WebGPU] Optimize MatMulNBits (#19852)
 x  e771a763c3 [js/test] align web test runner flags with ort.env (#19790)
o   79e50aeef3 [js/web] rewrite backend resolve to allow multiple EPs (#19735)
o   acb0df2280 Fix #19931 broken Get Started link of "ONNX Runtime JavaScript API" page (#19932)
o   b29849a287 [js/common] fix typedoc warnings (#19933)
o   afdab62f53 Bump follow-redirects from 1.15.4 to 1.15.6 in /js/web (#19949)
o   28ad6c3955 Bump follow-redirects from 1.15.4 to 1.15.6 in /js/node (#19951)
o   7e0d424934 accumulate in fp32 for Reduce* (#19868)
o   4c6a6a37f7 [js/webgpu] Fix NAN caused by un-initialized buffer in instance-norm (#19387)
o   01c7aaf6aa [js/webgpu] allow setting env.webgpu.adapter (#19940)
o   c45cff60cf [js/webgpu] fix maxpool / fp16 (#19981)
```

</details>

<details>
<summary>Cherry-pick commandlines</summary>

```sh
git cherry-pick 2e0a388c36
git cherry-pick d226e40856
git cherry-pick 61610ff986
git cherry-pick a33b5bd1fa
git cherry-pick 591f90c0b9
git cherry-pick 7252c6e747
git cherry-pick 5b06505073
git cherry-pick 656ca66186
git cherry-pick a3f0e2422b
git cherry-pick 9e69606360
git cherry-pick 624b4e2063
git cherry-pick 90883a366a
git cherry-pick 85cef0af8c  #<<<<< Note: conflicts
git cherry-pick d73131cf0f
git cherry-pick dd1f6ccc45
git cherry-pick 3a2ab1963a
git cherry-pick efc17e79de
git cherry-pick ccbe264a39
git cherry-pick 5ff27ef02a
git cherry-pick 06269a3952
git cherry-pick dfeda9019c
git cherry-pick 1b48054e1b
git cherry-pick 3fe2c137ee
git cherry-pick 6e04e36e3f
git cherry-pick 58f4921686
git cherry-pick 57d6819212
git cherry-pick ebd220b073
git cherry-pick 38c3432393
git cherry-pick fe82fccf1a
git cherry-pick 76a2a487a1
git cherry-pick 29b1106033
git cherry-pick ae3d73c981
git cherry-pick aec2389ad0
git cherry-pick bb43a0f133
git cherry-pick 0edb035808
git cherry-pick 3cb81cdde2
git cherry-pick e30618d055
git cherry-pick f06164ef8b
git cherry-pick 24b72d2613
git cherry-pick 4538d31a8b
git cherry-pick 53de2d8cb0
git cherry-pick ed250b88c3
git cherry-pick 79e50aeef3
git cherry-pick acb0df2280
git cherry-pick b29849a287
git cherry-pick afdab62f53
git cherry-pick 28ad6c3955
git cherry-pick 7e0d424934
git cherry-pick 4c6a6a37f7
git cherry-pick 01c7aaf6aa
git cherry-pick c45cff60cf
```
</details>

<details>
<summary>Cherry-pick conflicts</summary>

- 85cef0af8c #18989
this change is for enabling graph capture feature for JSEP, and it is
done after ROCM EP enabled graph capture feature. However, the ROCM EP
graph capture feature is not cherry-picked in rel-1.17.2.
</details>

---------

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: Jiajia Qin <jiajia.qin@intel.com>
Co-authored-by: Xu Xing <xing.xu@intel.com>
Co-authored-by: satyajandhyala <satya.k.jandhyala@gmail.com>
Co-authored-by: Yang Gu <yang.gu@intel.com>
Co-authored-by: Wanming Lin <wanming.lin@intel.com>
Co-authored-by: Jiajie Hu <jiajie.hu@intel.com>
Co-authored-by: Guenther Schmuelling <guschmue@microsoft.com>
Co-authored-by: Matttttt <18152455+martholomew@users.noreply.github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: Segev Finer <segev208@gmail.com>
Co-authored-by: Belem Zhang <belem.zhang@intel.com>
2024-03-29 13:13:39 -07:00

404 lines
14 KiB
TypeScript

// Copyright (c) Microsoft Corporation. All rights reserved.
// Licensed under the MIT License.
import {WebGpuBackend} from '../backend-webgpu';
import {LOG_DEBUG} from '../log';
import {GpuData, GpuDataId, GpuDataType} from './types';
/**
* manages GpuDataId -> GpuBuffer
*/
export interface GpuDataManager {
/**
* copy data from CPU to GPU.
*/
upload(id: GpuDataId, data: Uint8Array): void;
/**
* copy data from GPU to GPU.
*/
memcpy(sourceId: GpuDataId, destinationId: GpuDataId): void;
/**
* create new data on GPU.
*/
create(size: number, usage?: number): GpuData;
/**
* get GPU data by ID.
*/
get(id: GpuDataId): GpuData|undefined;
/**
* release the data on GPU by ID.
*
* @return size of the data released
*/
release(id: GpuDataId): number;
/**
* copy data from GPU to CPU.
*/
download(id: GpuDataId, getTargetBuffer: () => Uint8Array): Promise<void>;
/**
* refresh the buffers that marked for release.
*
* when release() is called, the buffer is not released immediately. this is because we need to wait for the commands
* to be submitted to the GPU. this function is called after the commands are submitted so that the buffers can be
* actually released.
*/
refreshPendingBuffers(): void;
/**
* register an external buffer for IO Binding. If the buffer is already registered, return the existing GPU data ID.
*
* GPU data manager only manages a mapping between the buffer and the GPU data ID. It will not manage the lifecycle of
* the external buffer.
*/
registerExternalBuffer(buffer: GPUBuffer, originalSize: number, previousBuffer?: GPUBuffer): number;
/**
* unregister an external buffer for IO Binding.
*/
unregisterExternalBuffer(buffer: GPUBuffer): void;
/**
* destroy all gpu buffers.
*/
dispose(): void;
/**
* release session related data.
* @param sessionId - specify the session ID.
*/
onReleaseSession(sessionId: number): void;
}
interface StorageCacheValue {
gpuData: GpuData;
originalSize: number;
}
/**
* normalize the buffer size so that it fits the 128-bits (16 bytes) alignment.
*/
const calcNormalizedBufferSize = (size: number) => Math.ceil(size / 16) * 16;
let guid = 1;
const createNewGpuDataId = () => guid++;
/**
* exported standard download function. This function is used by the session to download the data from GPU, and also by
* factory to create GPU tensors with the capacity of downloading data from GPU.
*
* @param backend - the WebGPU backend
* @param gpuBuffer - the GPU buffer to download
* @param originalSize - the original size of the data
* @param getTargetBuffer - optional. If provided, the data will be copied to the target buffer. Otherwise, a new buffer
* will be created and returned.
*/
export const downloadGpuData =
async(backend: WebGpuBackend, gpuBuffer: GPUBuffer, originalSize: number, getTargetBuffer?: () => Uint8Array):
Promise<Uint8Array> => {
const bufferSize = calcNormalizedBufferSize(originalSize);
const gpuReadBuffer = backend.device.createBuffer(
// eslint-disable-next-line no-bitwise
{size: bufferSize, usage: GPUBufferUsage.COPY_DST | GPUBufferUsage.MAP_READ});
try {
const commandEncoder = backend.getCommandEncoder();
backend.endComputePass();
commandEncoder.copyBufferToBuffer(
gpuBuffer /* source buffer */, 0 /* source offset */, gpuReadBuffer /* destination buffer */,
0 /* destination offset */, bufferSize /* size */
);
backend.flush();
await gpuReadBuffer.mapAsync(GPUMapMode.READ);
const arrayBuffer = gpuReadBuffer.getMappedRange();
if (getTargetBuffer) {
// if we already have a CPU buffer to accept the data, no need to clone the ArrayBuffer.
const targetBuffer = getTargetBuffer();
targetBuffer.set(new Uint8Array(arrayBuffer, 0, originalSize));
return targetBuffer;
} else {
// the mapped ArrayBuffer will be released when the GPU buffer is destroyed. Need to clone the
// ArrayBuffer.
return new Uint8Array(arrayBuffer.slice(0, originalSize));
}
} finally {
gpuReadBuffer.destroy();
}
};
class GpuDataManagerImpl implements GpuDataManager {
// GPU Data ID => GPU Data ( storage buffer )
private storageCache: Map<GpuDataId, StorageCacheValue>;
// pending buffers for uploading ( data is unmapped )
private buffersForUploadingPending: GPUBuffer[];
// pending buffers for computing
private buffersPending: GPUBuffer[];
// The reusable storage buffers for computing.
private freeBuffers: Map<number, GPUBuffer[]>;
// The reusable uniform buffers
private freeUniformBuffers: Map<number, GPUBuffer[]>;
// The external buffers registered users for IO Binding.
private externalBuffers: Map<GPUBuffer, GpuDataId>;
// The pendingBuffers for capture graph.
// a SessionID -> GPUBuffer[] mapping.
private capturedPendingBuffers: Map<number, GPUBuffer[]>;
constructor(private backend: WebGpuBackend) {
this.storageCache = new Map();
this.freeBuffers = new Map();
this.freeUniformBuffers = new Map();
this.buffersForUploadingPending = [];
this.buffersPending = [];
this.externalBuffers = new Map();
this.capturedPendingBuffers = new Map();
}
upload(id: GpuDataId, data: Uint8Array): void {
const srcArrayBuffer = data.buffer;
const srcOffset = data.byteOffset;
const srcLength = data.byteLength;
const size = calcNormalizedBufferSize(srcLength);
// get destination gpu buffer
const gpuDataCache = this.storageCache.get(id);
if (!gpuDataCache) {
throw new Error('gpu data for uploading does not exist');
}
if (gpuDataCache.originalSize !== srcLength) {
throw new Error(`inconsistent data size. gpu data size=${gpuDataCache.originalSize}, data size=${srcLength}`);
}
// create gpu buffer
const gpuBufferForUploading = this.backend.device.createBuffer(
// eslint-disable-next-line no-bitwise
{mappedAtCreation: true, size, usage: GPUBufferUsage.MAP_WRITE | GPUBufferUsage.COPY_SRC});
// copy (upload) data
const arrayBuffer = gpuBufferForUploading.getMappedRange();
new Uint8Array(arrayBuffer).set(new Uint8Array(srcArrayBuffer, srcOffset, srcLength));
gpuBufferForUploading.unmap();
// GPU copy
const commandEncoder = this.backend.getCommandEncoder();
this.backend.endComputePass();
commandEncoder.copyBufferToBuffer(gpuBufferForUploading, 0, gpuDataCache.gpuData.buffer, 0, size);
LOG_DEBUG('verbose', () => `[WebGPU] GpuDataManager.upload(id=${id})`);
this.buffersForUploadingPending.push(gpuBufferForUploading);
}
memcpy(sourceId: GpuDataId, destinationId: GpuDataId): void {
// get source gpu buffer
const sourceGpuDataCache = this.storageCache.get(sourceId);
if (!sourceGpuDataCache) {
throw new Error('source gpu data for memcpy does not exist');
}
// get destination gpu buffer
const destinationGpuDataCache = this.storageCache.get(destinationId);
if (!destinationGpuDataCache) {
throw new Error('destination gpu data for memcpy does not exist');
}
if (sourceGpuDataCache.originalSize !== destinationGpuDataCache.originalSize) {
throw new Error('inconsistent source and destination gpu data size');
}
const size = calcNormalizedBufferSize(sourceGpuDataCache.originalSize);
// GPU copy
const commandEncoder = this.backend.getCommandEncoder();
this.backend.endComputePass();
commandEncoder.copyBufferToBuffer(
sourceGpuDataCache.gpuData.buffer, 0, destinationGpuDataCache.gpuData.buffer, 0, size);
}
registerExternalBuffer(buffer: GPUBuffer, originalSize: number, previousBuffer?: GPUBuffer): number {
let id: number|undefined;
if (previousBuffer) {
id = this.externalBuffers.get(previousBuffer);
if (id === undefined) {
throw new Error('previous buffer is not registered');
}
if (buffer === previousBuffer) {
LOG_DEBUG(
'verbose',
() => `[WebGPU] GpuDataManager.registerExternalBuffer(size=${originalSize}) => id=${
id}, buffer is the same, skip.`);
return id;
} else if (this.backend.capturedCommandList.has(this.backend.currentSessionId!)) {
throw new Error(`Registering a different external buffer under graph capture mode is not supported yet.
Please use the previous external buffer!`);
}
this.externalBuffers.delete(previousBuffer);
} else {
id = createNewGpuDataId();
}
this.storageCache.set(id, {gpuData: {id, type: GpuDataType.default, buffer}, originalSize});
this.externalBuffers.set(buffer, id);
LOG_DEBUG(
'verbose',
() => `[WebGPU] GpuDataManager.registerExternalBuffer(size=${originalSize}) => id=${id}, registered.`);
return id;
}
unregisterExternalBuffer(buffer: GPUBuffer): void {
const id = this.externalBuffers.get(buffer);
if (id !== undefined) {
this.storageCache.delete(id);
this.externalBuffers.delete(buffer);
LOG_DEBUG('verbose', () => `[WebGPU] GpuDataManager.unregisterExternalBuffer() => id=${id}`);
}
}
// eslint-disable-next-line no-bitwise
create(size: number, usage = GPUBufferUsage.STORAGE | GPUBufferUsage.COPY_SRC | GPUBufferUsage.COPY_DST): GpuData {
const bufferSize = calcNormalizedBufferSize(size);
let gpuBuffer;
// Currently, only storage buffers are reused.
// eslint-disable-next-line no-bitwise
const isStorage = (usage & GPUBufferUsage.STORAGE) === GPUBufferUsage.STORAGE;
// eslint-disable-next-line no-bitwise
const isUniform = (usage & GPUBufferUsage.UNIFORM) === GPUBufferUsage.UNIFORM;
if (isStorage || isUniform) {
const freeBuffers = isStorage ? this.freeBuffers : this.freeUniformBuffers;
let buffers = freeBuffers.get(bufferSize);
if (!buffers) {
buffers = [];
freeBuffers.set(bufferSize, buffers);
}
if (buffers.length > 0) {
gpuBuffer = buffers.pop() as GPUBuffer;
} else {
// create gpu buffer
gpuBuffer = this.backend.device.createBuffer({size: bufferSize, usage});
}
} else {
// create gpu buffer
gpuBuffer = this.backend.device.createBuffer({size: bufferSize, usage});
}
const gpuData = {id: createNewGpuDataId(), type: GpuDataType.default, buffer: gpuBuffer};
this.storageCache.set(gpuData.id, {gpuData, originalSize: size});
LOG_DEBUG('verbose', () => `[WebGPU] GpuDataManager.create(size=${size}) => id=${gpuData.id}`);
return gpuData;
}
get(id: GpuDataId): GpuData|undefined {
return this.storageCache.get(id)?.gpuData;
}
release(id: GpuDataId): number {
const cachedData = this.storageCache.get(id);
if (!cachedData) {
throw new Error('releasing data does not exist');
}
LOG_DEBUG('verbose', () => `[WebGPU] GpuDataManager.release(id=${id}), gpuDataId=${cachedData.gpuData.id}`);
this.storageCache.delete(id);
this.buffersPending.push(cachedData.gpuData.buffer);
// cachedData.gpuData.buffer.destroy();
return cachedData.originalSize;
}
async download(id: GpuDataId, getTargetBuffer: () => Uint8Array): Promise<void> {
const cachedData = this.storageCache.get(id);
if (!cachedData) {
throw new Error('data does not exist');
}
await downloadGpuData(this.backend, cachedData.gpuData.buffer, cachedData.originalSize, getTargetBuffer);
}
refreshPendingBuffers(): void {
for (const buffer of this.buffersForUploadingPending) {
// upload buffer is only useful in the session creation time. So we don't need to reuse them in session running.
buffer.destroy();
}
this.buffersForUploadingPending = [];
if (this.buffersPending.length === 0) {
return;
}
if (this.backend.sessionStatus === 'default') {
for (const buffer of this.buffersPending) {
// eslint-disable-next-line no-bitwise
if ((buffer.usage & GPUBufferUsage.STORAGE) === GPUBufferUsage.STORAGE) {
// Put the pending buffer to freeBuffers list instead of really destroying it for buffer reusing.
this.freeBuffers.get(buffer.size)!.push(buffer);
// eslint-disable-next-line no-bitwise
} else if ((buffer.usage & GPUBufferUsage.UNIFORM) === GPUBufferUsage.UNIFORM) {
// Put the pending buffer to freeUniformBuffers list instead of really destroying it for buffer reusing.
this.freeUniformBuffers.get(buffer.size)!.push(buffer);
} else {
buffer.destroy();
}
}
this.buffersPending = [];
} else {
// Don't release intermediate tensors in non-default mode.
// TODO: reuse the storage buffers in non-default mode.
let capturedBuffers = this.capturedPendingBuffers.get(this.backend.currentSessionId!);
if (!capturedBuffers) {
capturedBuffers = [];
this.capturedPendingBuffers.set(this.backend.currentSessionId!, capturedBuffers);
}
for (const buffer of this.buffersPending) {
capturedBuffers.push(buffer);
}
this.buffersPending = [];
}
}
dispose() {
this.freeBuffers.forEach((buffers) => {
buffers.forEach(buffer => {
buffer.destroy();
});
});
this.freeUniformBuffers.forEach((buffers) => {
buffers.forEach(buffer => {
buffer.destroy();
});
});
this.storageCache.forEach((storage) => {
storage.gpuData.buffer.destroy();
});
this.capturedPendingBuffers.forEach((buffers) => {
buffers.forEach(buffer => {
buffer.destroy();
});
});
this.storageCache = new Map();
this.freeBuffers = new Map();
this.freeUniformBuffers = new Map();
this.capturedPendingBuffers = new Map();
}
onReleaseSession(sessionId: number) {
// release the captured pending buffers.
const pendingBuffers = this.capturedPendingBuffers.get(sessionId);
if (pendingBuffers) {
pendingBuffers.forEach(buffer => {
buffer.destroy();
});
this.capturedPendingBuffers.delete(sessionId);
}
}
}
export const createGpuDataManager = (...args: ConstructorParameters<typeof GpuDataManagerImpl>): GpuDataManager =>
new GpuDataManagerImpl(...args);