Buffered Decoding. #1269

JimBobSquarePants · 2020-07-10T16:02:49Z

Prerequisites

I have written a descriptive pull-request title
I have verified that there are no overlapping pull-requests open
I have verified that I am following matches the existing coding patterns and practice as demonstrated in the repository. These follow strict Stylecop rules 👮.
I have provided test coverage for my change (where applicable)

Description

I've been tinkering about with this for a while...

We've used buffered reading via a stream wrapper in our JpegDecoder to reduce the overhead of single byte reads. You end up with a nice performance boost over native reading.

ImageSharp/src/ImageSharp/Formats/Jpeg/JpegDecoderCore.cs

Line 246 in c0594f7

    
           this.InputStream = new DoubleBufferedStreamReader(this.Configuration.MemoryAllocator, stream);

However, the shape of the class doesn't allow for transparent reuse with other decoders which is a must. The mileage this buffering gives for each encoder varies based upon individual implementation but the net benefit is worth the investment.

With that in mind I refactored the class to turn it into an actual stream that internally wraps our input stream and is transparent to any decoder implementation. A great benefit to this is that we can fix our buffered stream to correctly handle when a stream returns less than expected on Read(buffer, offset, count) Fixes #1268 .

The following benchmarks represent our new new implementation BufferedReadStream vs the old jpeg wrapper BufferedReadStreamWrap, reading a MemoryStream preloaded with a buffer, and reading an array directly.

Things to note:

We're a lot faster on .NET Core + than the native stream reader and ~match the wrapping implementation on those frameworks.

The performance loss against the wrapper implementation in .NETFX is due to the jitter being unable to inline the appropriate methods on a sealed, inheriting class, I've experimented with several implementations and there seems to be no way around it. Since we are testing against the fastest possible stream reading scenario and we are measuring in us I'm not concerned. Our decoder benchmarks on NET FX are unaffected.

Method	Runtime	Mean	Error	StdDev	Ratio	RatioSD	Gen 0	Gen 1	Gen 2	Allocated
StandardStreamRead	.NET 4.7.2	126.07 us	126.498 us	6.934 us	0.99	0.08	-	-	-	-
BufferedReadStreamRead	.NET 4.7.2	118.08 us	42.234 us	2.315 us	0.92	0.03	-	-	-	-
BufferedReadStreamWrapRead	.NET 4.7.2	45.33 us	22.833 us	1.252 us	0.35	0.00	-	-	-	-
StandardStreamReadByte	.NET 4.7.2	128.17 us	94.616 us	5.186 us	1.00	0.00	-	-	-	-
BufferedReadStreamReadByte	.NET 4.7.2	143.60 us	92.871 us	5.091 us	1.12	0.08	-	-	-	-
BufferedReadStreamWrapReadByte	.NET 4.7.2	32.72 us	53.708 us	2.944 us	0.26	0.02	-	-	-	-
ArrayReadByte	.NET 4.7.2	19.40 us	12.206 us	0.669 us	0.15	0.01	-	-	-	-

StandardStreamRead	.NET Core 2.1	84.82 us	55.983 us	3.069 us	0.75	0.15	-	-	-	-
BufferedReadStreamRead	.NET Core 2.1	49.62 us	27.253 us	1.494 us	0.44	0.08	-	-	-	-
BufferedReadStreamWrapRead	.NET Core 2.1	67.78 us	87.546 us	4.799 us	0.60	0.10	-	-	-	-
StandardStreamReadByte	.NET Core 2.1	115.81 us	382.107 us	20.945 us	1.00	0.00	-	-	-	-
BufferedReadStreamReadByte	.NET Core 2.1	16.32 us	6.123 us	0.336 us	0.14	0.02	-	-	-	-
BufferedReadStreamWrapReadByte	.NET Core 2.1	16.68 us	4.616 us	0.253 us	0.15	0.03	-	-	-	-
ArrayReadByte	.NET Core 2.1	15.13 us	60.763 us	3.331 us	0.14	0.05	-	-	-	-

StandardStreamRead	.NET Core 3.1	92.44 us	88.217 us	4.835 us	0.94	0.06	-	-	-	-
BufferedReadStreamRead	.NET Core 3.1	36.41 us	5.923 us	0.325 us	0.37	0.00	-	-	-	-
BufferedReadStreamWrapRead	.NET Core 3.1	37.22 us	9.628 us	0.528 us	0.38	0.01	-	-	-	-
StandardStreamReadByte	.NET Core 3.1	98.67 us	20.947 us	1.148 us	1.00	0.00	-	-	-	-
BufferedReadStreamReadByte	.NET Core 3.1	41.36 us	123.536 us	6.771 us	0.42	0.06	-	-	-	-
BufferedReadStreamWrapReadByte	.NET Core 3.1	39.11 us	93.894 us	5.147 us	0.40	0.05	-	-	-	-
ArrayReadByte	.NET Core 3.1	18.84 us	4.684 us	0.257 us	0.19	0.00	-	-	-	-

JimBobSquarePants · 2020-07-10T16:07:05Z

tests/ImageSharp.Tests/Image/ImageTests.Load_FileSystemPath_PassLocalConfiguration.cs

            }

-            [Fact]
+            [Fact(Skip = "TODO: Enable when someone tells me how this mocking stuff works.")]


@tocsoft I really struggled to understand how this worked.

Currently the test verifies that the mock IImageDecoders instances Decode<Rgba32>() method is called with exactly 2 properties that have reference equality (I believe it reference equality anyway) to this.TopLevelConfiguration and this.DataStream (which will have been returned from a mock/fake file system configured in TopLevelConfiguration when its called with MockFilePath).

For your new code which now wraps all the streams with a BufferedReadStream you will need to change the this.DataStream parameter to something like It.Is<Stream>(x=>(BufferedReadStream)x.BaseStream == this.DataStream) which allows the 'verify' to swap in other sorts of checks. (would require the base stream to be exposed from the buffered read stream)

Excellent. That fixed it!

codecov · 2020-07-10T16:16:48Z

Codecov Report

Merging #1269 into master will decrease coverage by 0.01%.
The diff coverage is 82.57%.

@@            Coverage Diff             @@
##           master    #1269      +/-   ##
==========================================
- Coverage   82.68%   82.67%   -0.02%     
==========================================
  Files         697      697              
  Lines       30589    30656      +67     
  Branches     3460     3467       +7     
==========================================
+ Hits        25294    25344      +50     
- Misses       4593     4612      +19     
+ Partials      702      700       -2

Flag	Coverage Δ
#unittests	`82.67% <82.57%> (-0.02%)`	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/ImageSharp/Formats/Bmp/BmpDecoderCore.cs	`94.45% <ø> (ø)`
src/ImageSharp/Formats/Gif/GifDecoderCore.cs	`83.96% <ø> (ø)`
...rmats/Jpeg/Components/Decoder/HuffmanScanBuffer.cs	`94.80% <ø> (ø)`
src/ImageSharp/Formats/Tga/TgaDecoderCore.cs	`91.29% <ø> (ø)`
src/ImageSharp/Formats/Gif/GifDecoder.cs	`50.00% <46.15%> (ø)`
src/ImageSharp/Formats/Png/PngDecoder.cs	`48.14% <50.00%> (+0.52%)`	⬆️
src/ImageSharp/Formats/Tga/TgaDecoder.cs	`46.42% <50.00%> (+0.97%)`	⬆️
src/ImageSharp/Formats/Jpeg/JpegDecoder.cs	`48.38% <53.84%> (+0.23%)`	⬆️
src/ImageSharp/Image.Decode.cs	`87.30% <75.00%> (-1.04%)`	⬇️
src/ImageSharp/IO/BufferedReadStream.cs	`85.60% <85.60%> (ø)`
... and 12 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 7ee0207...e6c78c2. Read the comment docs.

JimBobSquarePants · 2020-07-12T19:06:18Z

@SixLabors/core Any objections to this? I'm getting a twitchy merge finger.

antonfirsov

See my comment below + I think some Jpeg before/after numbers would be also helpful indicators.

antonfirsov · 2020-07-12T21:27:12Z

src/ImageSharp/Formats/Jpeg/Components/Decoder/HuffmanScanBuffer.cs

    internal struct HuffmanScanBuffer
    {
-        private readonly DoubleBufferedStreamReader stream;
+        private readonly Stream stream;


We probably loose a some of the benchmarked performance gain here, since the instance used here is of the unsealed base class Stream, therefore the virtual methods being called on it are not inlined (unlike in benchmarks!).

I didn’t actually see any slowdown at all in jpeg but I’ll run/publish tests tomorrow

@antonfirsov Before and after benchmarks. Short run so plenty of error but nothing that stands out. Ideally I'd be able to use the concrete implementation but I always worry about exposing a new class and breaking changes.

Method TestImage Mean Error StdDev Ratio RatioSD Gen 0 Gen 1 Gen 2 Allocated

'Decode Jpeg - System.Drawing' Jpg/b(...)e.jpg [21] 5.123 ms 7.2206 ms 0.3958 ms 1.00 0.00 - - - 176 B

'Decode Jpeg - ImageSharp' Jpg/b(...)e.jpg [21] 12.219 ms 15.6786 ms 0.8594 ms 2.39 0.02 - - - 15904 B

'Decode Jpeg - System.Drawing' Jpg/b(...)f.jpg [28] 13.339 ms 0.5726 ms 0.0314 ms 1.00 0.00 - - - 176 B

'Decode Jpeg - ImageSharp' Jpg/b(...)f.jpg [28] 29.084 ms 6.2870 ms 0.3446 ms 2.18 0.02 - - - 16880 B

'Decode Jpeg - System.Drawing' Jpg/i(...)e.jpg [43] 329.528 ms 56.3769 ms 3.0902 ms 1.00 0.00 - - - 216 B

'Decode Jpeg - ImageSharp' Jpg/i(...)e.jpg [43] 268.073 ms 179.0279 ms 9.8131 ms 0.81 0.04 - - - 36022424 B

Method TestImage Mean Error StdDev Ratio RatioSD Gen 0 Gen 1 Gen 2 Allocated

'Decode Jpeg - System.Drawing' Jpg/b(...)e.jpg [21] 4.965 ms 0.6894 ms 0.0378 ms 1.00 0.00 - - - 176 B

'Decode Jpeg - ImageSharp' Jpg/b(...)e.jpg [21] 11.881 ms 0.3953 ms 0.0217 ms 2.39 0.02 - - - 15808 B

'Decode Jpeg - System.Drawing' Jpg/b(...)f.jpg [28] 13.548 ms 1.5905 ms 0.0872 ms 1.00 0.00 - - - 176 B

'Decode Jpeg - ImageSharp' Jpg/b(...)f.jpg [28] 28.147 ms 1.6038 ms 0.0879 ms 2.08 0.01 - - - 16816 B

'Decode Jpeg - System.Drawing' Jpg/i(...)e.jpg [43] 332.870 ms 22.0796 ms 1.2103 ms 1.00 0.00 - - - 176 B

'Decode Jpeg - ImageSharp' Jpg/i(...)e.jpg [43] 286.208 ms 732.9871 ms 40.1775 ms 0.86 0.12 - - - 36022360 B

It's ~6% change for the largest image, which is suspicious. These are no microbenchmarks, so the result of a short run is relevant, unless there is other noise, which can ofc happen with local runs.

We have DecodeJpegParseStreamOnly as a relatively quick option reassess this.

Before

Method Job Runtime TestImage Mean Error StdDev Ratio RatioSD Gen 0 Gen 1 Gen 2 Allocated

'System.Drawing FULL' Job-COEKJJ .NET 4.7.2 Jpg/b(...)f.jpg [28] 16.09 ms 0.024 ms 0.001 ms 1.00 0.00 156.2500 - - 777784 B

JpegDecoderCore.ParseStream Job-COEKJJ .NET 4.7.2 Jpg/b(...)f.jpg [28] 17.96 ms 12.218 ms 0.670 ms 1.12 0.04 - - - 12544 B

'System.Drawing FULL' Job-UZXMLV .NET Core 2.1 Jpg/b(...)f.jpg [28] 15.25 ms 0.973 ms 0.053 ms 1.00 0.00 171.8750 - - 775208 B

JpegDecoderCore.ParseStream Job-UZXMLV .NET Core 2.1 Jpg/b(...)f.jpg [28] 12.52 ms 1.801 ms 0.099 ms 0.82 0.01 - - - 12416 B

'System.Drawing FULL' Job-PIDJAF .NET Core 3.1 Jpg/b(...)f.jpg [28] 14.43 ms 3.392 ms 0.186 ms 1.00 0.00 - - - 191 B

JpegDecoderCore.ParseStream Job-PIDJAF .NET Core 3.1 Jpg/b(...)f.jpg [28] 11.55 ms 1.149 ms 0.063 ms 0.80 0.01 - - - 12408 B

After

Method Runtime TestImage Mean Error StdDev Ratio Gen 0 Gen 1 Gen 2 Allocated

'System.Drawing FULL' .NET 4.7.2 Jpg/b(...)f.jpg [28] 13.52 ms 1.621 ms 0.089 ms 1.00 171.8750 - - 777789 B

JpegDecoderCore.ParseStream .NET 4.7.2 Jpg/b(...)f.jpg [28] 13.01 ms 1.466 ms 0.080 ms 0.96 - - - 12416 B

'System.Drawing FULL' .NET Core 2.1 Jpg/b(...)f.jpg [28] 13.64 ms 3.331 ms 0.183 ms 1.00 171.8750 - - 775208 B

JpegDecoderCore.ParseStream .NET Core 2.1 Jpg/b(...)f.jpg [28] 11.89 ms 0.374 ms 0.021 ms 0.87 - - - 12248 B

'System.Drawing FULL' .NET Core 3.1 Jpg/b(...)f.jpg [28] 13.20 ms 1.116 ms 0.061 ms 1.00 - - - 176 B

JpegDecoderCore.ParseStream .NET Core 3.1 Jpg/b(...)f.jpg [28] 12.09 ms 0.310 ms 0.017 ms 0.92 - - - 12246 B

I think I've got a good way now actually.... I updated the internal interface for the core decoders to use the buffered stream. I'll have something to show very soon.

@antonfirsov Less of a difference than I thought. That's with all the core decoder directly referencing the concrete decoder.

Method Job Runtime TestImage Mean Error StdDev Ratio Gen 0 Gen 1 Gen 2 Allocated

'System.Drawing FULL' Job-RSRDJA .NET 4.7.2 Jpg/b(...)f.jpg [28] 13.28 ms 0.976 ms 0.053 ms 1.00 171.8750 - - 777761 B

JpegDecoderCore.ParseStream Job-RSRDJA .NET 4.7.2 Jpg/b(...)f.jpg [28] 13.54 ms 0.792 ms 0.043 ms 1.02 - - - 12416 B

'System.Drawing FULL' Job-QLXCST .NET Core 2.1 Jpg/b(...)f.jpg [28] 13.24 ms 2.348 ms 0.129 ms 1.00 171.8750 - - 775208 B

JpegDecoderCore.ParseStream Job-QLXCST .NET Core 2.1 Jpg/b(...)f.jpg [28] 11.72 ms 1.564 ms 0.086 ms 0.89 - - - 12352 B

'System.Drawing FULL' Job-KQKRGS .NET Core 3.1 Jpg/b(...)f.jpg [28] 13.08 ms 2.922 ms 0.160 ms 1.00 - - - 176 B

JpegDecoderCore.ParseStream Job-KQKRGS .NET Core 3.1 Jpg/b(...)f.jpg [28] 11.49 ms 0.771 ms 0.042 ms 0.88 - - - 12372 B

On .NET Core 3.1:
11.55 (old) VS 12.09 (base Stream) VS 11.49 (concrete Stream)

I think it's quite good, or am I missing something?

I'm looking at the ratios compared to before but error could be creeping in there. It's definitely faster

I'm gonna push with the concrete implementation as it should bring us the best of everything.

@antonfirsov OK. I think this is ready to review again now.

tocsoft · 2020-07-18T08:31:13Z

src/ImageSharp/Formats/Png/PngEncoder.cs

-                await encoder.EncodeAsync(image, stream).ConfigureAwait(false);
-            }
+            using var encoder = new PngEncoderCore(image.GetMemoryAllocator(), image.GetConfiguration(), new PngEncoderOptions(this));
+            return encoder.EncodeAsync(image, stream);


I'm sure this one needs awaiting otherwise the encoder will be disposed before the Task is awaited/activated and things will possibly exploded, but only if we end up having to perform a CopyToAsync which is unlikely to be happening the unit test

I'll switch it back..... Done

antonfirsov

Only one finding, otherwise LGTM.

antonfirsov · 2020-07-21T21:39:48Z

src/ImageSharp/Common/Extensions/StreamExtensions.cs

+            while (count > 0)
            {
-                var foo = new byte[count];
-                while (count > 0)
+                int bytesRead = stream.Read(buffer, 0, count);
+                if (bytesRead == 0)
                {
-                    int bytesRead = stream.Read(foo, 0, count);
-                    if (bytesRead == 0)
-                    {
-                        break;
-                    }
-
-                    count -= bytesRead;
+                    break;
                }
+
+                count -= bytesRead;
            }
+
+            ArrayPool<byte>.Shared.Return(buffer);


Needs a try { ... } finally { ArrayPool<byte>.Shared.Return(buffer); } block.

Sweet, will do.

antonfirsov · 2020-07-21T22:00:02Z

src/ImageSharp/Formats/ImageDecoderUtilities.cs

-            ms.Position = 0;
-            return decoder.Decode<TPixel>(ms);
-        }
+            => Task.FromResult(decoder.Decode<TPixel>(stream));


Was the MemoryStream copying logic an unnecessary duplicate here it's already has been done in Image.WithSeekableStreamAsync ?

Yep, exactly.

antonfirsov · 2020-07-21T22:03:49Z

tests/ImageSharp.Tests/IO/BufferedReadStreamTests.cs

+
+        // Simulates a stream that can only return 1 byte at a time per read instruction.
+        // See https://github.com/SixLabors/ImageSharp/issues/1268
+        private class EvilStream : MemoryStream


antonfirsov · 2020-07-21T23:47:40Z

src/ImageSharp/Formats/ImageDecoderUtilities.cs

+        /// <exception cref="ArgumentNullException"><paramref name="stream"/> is null.</exception>
+        /// <returns>A <see cref="Task"/> representing the asynchronous operation.</returns>
+        public static Task<IImageInfo> IdentifyAsync(this IImageDecoderInternals decoder, BufferedReadStream stream)
+            => Task.FromResult(decoder.Identify(stream));


In case of exception in Identify (and Load), we should return Task.FromException(ex) instead of failing to create the Task.

The fix is on the way, as I'm working on a PR for #1280, so don't bother with it.

Sweet. Will wait to see what you come up with then. 👍

JimBobSquarePants added 10 commits April 17, 2020 19:05

BufferedReadStream WIP - MagickImage not working.

545cf7d

Update BufferedReadStream.cs

bbf85c1

Remove casting, skip mocks

f8a88e4

Cleanup and refactor

5ca6063

Merge branch 'master' into js/buffered-read

219a4fe

Merge branch 'master' into js/buffered-read

48e1e8d

Merge branch 'master' into js/buffered-read

94a3ea5

Merge branch 'master' into js/buffered-read

889dfd5

Handle slow streams. Fix #1268

529d120

Restore missing Guard

96ec070

JimBobSquarePants commented Jul 10, 2020

View reviewed changes

JimBobSquarePants added 2 commits July 10, 2020 17:52

More efficient reading

badae1c

Re-enable skipped tests

a546676

JimBobSquarePants requested a review from a team July 10, 2020 18:03

JimBobSquarePants added area:formats area:performance bug labels Jul 10, 2020

JimBobSquarePants added this to the 1.0.0 milestone Jul 10, 2020

antonfirsov reviewed Jul 12, 2020

View reviewed changes

JimBobSquarePants added 6 commits July 16, 2020 11:13

Internalize unsafe component.

3b94e3d

Use missed offset

4fe04b6

Merge branch 'master' into js/buffered-read

f850e07

Merge branch 'master' into js/buffered-read

2f3b59f

Use explicit stream instance in core decoders.

bbf5b3b

Remove unrequired async/await

d5f7e5a

tocsoft reviewed Jul 18, 2020

View reviewed changes

Update PngEncoder.cs

30092db

antonfirsov approved these changes Jul 21, 2020

View reviewed changes

antonfirsov reviewed Jul 21, 2020

View reviewed changes

Update StreamExtensions.cs

e6c78c2

JimBobSquarePants merged commit ef5e21b into master Jul 22, 2020

JimBobSquarePants deleted the js/buffered-read branch July 22, 2020 08:18

dependabot bot mentioned this pull request Sep 23, 2025

Bump the nuget group with 5 updates norschel/enterJSWebSecurity2025-Demo1#8

Open

dependabot bot mentioned this pull request Oct 19, 2025

Bump the nuget group with 1 update ewdlop/bepuphysics2Fork#1

Open

Method	TestImage	Mean	Error	StdDev	Ratio	RatioSD	Gen 0	Gen 1	Gen 2	Allocated
'Decode Jpeg - System.Drawing'	Jpg/b(...)e.jpg [21]	5.123 ms	7.2206 ms	0.3958 ms	1.00	0.00	-	-	-	176 B
'Decode Jpeg - ImageSharp'	Jpg/b(...)e.jpg [21]	12.219 ms	15.6786 ms	0.8594 ms	2.39	0.02	-	-	-	15904 B

'Decode Jpeg - System.Drawing'	Jpg/b(...)f.jpg [28]	13.339 ms	0.5726 ms	0.0314 ms	1.00	0.00	-	-	-	176 B
'Decode Jpeg - ImageSharp'	Jpg/b(...)f.jpg [28]	29.084 ms	6.2870 ms	0.3446 ms	2.18	0.02	-	-	-	16880 B

'Decode Jpeg - System.Drawing'	Jpg/i(...)e.jpg [43]	329.528 ms	56.3769 ms	3.0902 ms	1.00	0.00	-	-	-	216 B
'Decode Jpeg - ImageSharp'	Jpg/i(...)e.jpg [43]	268.073 ms	179.0279 ms	9.8131 ms	0.81	0.04	-	-	-	36022424 B

Method	TestImage	Mean	Error	StdDev	Ratio	RatioSD	Gen 0	Gen 1	Gen 2	Allocated
'Decode Jpeg - System.Drawing'	Jpg/b(...)e.jpg [21]	4.965 ms	0.6894 ms	0.0378 ms	1.00	0.00	-	-	-	176 B
'Decode Jpeg - ImageSharp'	Jpg/b(...)e.jpg [21]	11.881 ms	0.3953 ms	0.0217 ms	2.39	0.02	-	-	-	15808 B

'Decode Jpeg - System.Drawing'	Jpg/b(...)f.jpg [28]	13.548 ms	1.5905 ms	0.0872 ms	1.00	0.00	-	-	-	176 B
'Decode Jpeg - ImageSharp'	Jpg/b(...)f.jpg [28]	28.147 ms	1.6038 ms	0.0879 ms	2.08	0.01	-	-	-	16816 B

'Decode Jpeg - System.Drawing'	Jpg/i(...)e.jpg [43]	332.870 ms	22.0796 ms	1.2103 ms	1.00	0.00	-	-	-	176 B
'Decode Jpeg - ImageSharp'	Jpg/i(...)e.jpg [43]	286.208 ms	732.9871 ms	40.1775 ms	0.86	0.12	-	-	-	36022360 B

Method	Job	Runtime	TestImage	Mean	Error	StdDev	Ratio	RatioSD	Gen 0	Gen 1	Gen 2	Allocated
'System.Drawing FULL'	Job-COEKJJ	.NET 4.7.2	Jpg/b(...)f.jpg [28]	16.09 ms	0.024 ms	0.001 ms	1.00	0.00	156.2500	-	-	777784 B
JpegDecoderCore.ParseStream	Job-COEKJJ	.NET 4.7.2	Jpg/b(...)f.jpg [28]	17.96 ms	12.218 ms	0.670 ms	1.12	0.04	-	-	-	12544 B

'System.Drawing FULL'	Job-UZXMLV	.NET Core 2.1	Jpg/b(...)f.jpg [28]	15.25 ms	0.973 ms	0.053 ms	1.00	0.00	171.8750	-	-	775208 B
JpegDecoderCore.ParseStream	Job-UZXMLV	.NET Core 2.1	Jpg/b(...)f.jpg [28]	12.52 ms	1.801 ms	0.099 ms	0.82	0.01	-	-	-	12416 B

'System.Drawing FULL'	Job-PIDJAF	.NET Core 3.1	Jpg/b(...)f.jpg [28]	14.43 ms	3.392 ms	0.186 ms	1.00	0.00	-	-	-	191 B
JpegDecoderCore.ParseStream	Job-PIDJAF	.NET Core 3.1	Jpg/b(...)f.jpg [28]	11.55 ms	1.149 ms	0.063 ms	0.80	0.01	-	-	-	12408 B

Method	Job	Runtime	TestImage	Mean	Error	StdDev	Ratio	Gen 0	Gen 1	Gen 2	Allocated
'System.Drawing FULL'	Job-RSRDJA	.NET 4.7.2	Jpg/b(...)f.jpg [28]	13.28 ms	0.976 ms	0.053 ms	1.00	171.8750	-	-	777761 B
JpegDecoderCore.ParseStream	Job-RSRDJA	.NET 4.7.2	Jpg/b(...)f.jpg [28]	13.54 ms	0.792 ms	0.043 ms	1.02	-	-	-	12416 B

'System.Drawing FULL'	Job-QLXCST	.NET Core 2.1	Jpg/b(...)f.jpg [28]	13.24 ms	2.348 ms	0.129 ms	1.00	171.8750	-	-	775208 B
JpegDecoderCore.ParseStream	Job-QLXCST	.NET Core 2.1	Jpg/b(...)f.jpg [28]	11.72 ms	1.564 ms	0.086 ms	0.89	-	-	-	12352 B

'System.Drawing FULL'	Job-KQKRGS	.NET Core 3.1	Jpg/b(...)f.jpg [28]	13.08 ms	2.922 ms	0.160 ms	1.00	-	-	-	176 B
JpegDecoderCore.ParseStream	Job-KQKRGS	.NET Core 3.1	Jpg/b(...)f.jpg [28]	11.49 ms	0.771 ms	0.042 ms	0.88	-	-	-	12372 B

Uh oh!

Buffered Decoding. #1269

Buffered Decoding. #1269

Uh oh!

Conversation

JimBobSquarePants commented Jul 10, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Prerequisites

Description

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

codecov bot commented Jul 10, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

JimBobSquarePants commented Jul 12, 2020

Uh oh!

antonfirsov left a comment

Choose a reason for hiding this comment

Uh oh!

antonfirsov Jul 12, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

JimBobSquarePants Jul 13, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

antonfirsov Jul 14, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

JimBobSquarePants Jul 18, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

antonfirsov left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

antonfirsov Jul 21, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

JimBobSquarePants commented Jul 10, 2020 •

edited

Loading

codecov bot commented Jul 10, 2020 •

edited

Loading

antonfirsov Jul 12, 2020 •

edited

Loading

JimBobSquarePants Jul 13, 2020 •

edited

Loading

antonfirsov Jul 14, 2020 •

edited

Loading

JimBobSquarePants Jul 18, 2020 •

edited

Loading

antonfirsov Jul 21, 2020 •

edited

Loading