parallelpatterns Issue Tracker Rss Feed

Closed Issue: Pipeline diagram error [23496]

AdeMiller — Mon, 22 Oct 2012 20:34:08 GMT

On the Pipelines page, the figure 1's arrow between Read Strings and Buffer 1 incorrectly goes from the buffer to the processing element. Instead the arrow must go from the processing element to the buffer, as indicated by the text below the diagram:

For example, the Read Strings task reads from a source and writes to buffer 1.

http://msdn.microsoft.com/en-us/library/ff963548.aspx
Comments:

This is now resolved on the MSDN site. I got the image updated.

David. My comments were intended to convey updates as I worked on the issue, hence "Sent" not "Send". Sorry for any confusion.

Thanks for bringing this to our attention.

Ade

Commented Issue: Pipeline diagram error [23496]

daviburg — Fri, 19 Oct 2012 17:44:52 GMT

How? The MSDN page is redirecting here to CodePlex for community feedback.

Commented Issue: Pipeline diagram error [23496]

AdeMiller — Fri, 19 Oct 2012 04:41:21 GMT

Sent correction to the MSDN team.

Closed Issue: Trend.Predict(double ordinate) [19190]

AdeMiller — Fri, 19 Oct 2012 04:36:08 GMT

The argument of Microsoft.Practices.ParallelGuideSamples.Utilities.Trend.Predict(double ordinate) should be called abscissa according to the mathematical definition.

Comments:

Resolved in changeset 69939.

Commented Issue: Trend.Predict(double ordinate) [19190]

AdeMiller — Fri, 19 Oct 2012 04:05:27 GMT

The argument of Microsoft.Practices.ParallelGuideSamples.Utilities.Trend.Predict(double ordinate) should be called abscissa according to the mathematical definition.

Comments: ** Comment from web user: AdeMiller **

This doesn't appear in the book text so it can be updated in the code.

Closed Issue: Book Errata [18665]

AdeMiller — Fri, 19 Oct 2012 03:09:11 GMT

This issue is designed for tracking errata in the book text. Things like minor typos.
 
*Page 65*
 
On page 65 of the Preliminary Draft there is a typo: "form" should be "from." This is found in the following text: 
 
"Bypassing the Thread Pool If you don’t want a task to use a worker thread **form** the thread pool, you can create a new thread for its dedicated use."

Comments:

All the comments here have been migrated to the O'Reilly errata page.

http://www.oreillynet.com/oreilly/authors/errata.csp?b=0790145310262

Commented Issue: Pipeline diagram error [23496]

AdeMiller — Fri, 19 Oct 2012 02:42:13 GMT

This is only an issue in the MSDN content. They have an incorrect version of the image.

Created Issue: Pipeline diagram error [23496]

daviburg — Mon, 15 Oct 2012 22:43:21 GMT

On the "Pipelines" the figure 1 arrow between "Read Strings" and "Buffer 1" incorrectly goes from the buffer to the processing element. Instead the arrow must go from the processing element to the buffer, as indicated by the text below the diagram:

"For example, the "Read Strings" task reads from a source and writes to buffer 1."

http://msdn.microsoft.com/en-us/library/ff963548.aspx

Commented Issue: Book Errata [18665]

jsh003guardz63 — Tue, 19 Jul 2011 15:15:41 GMT

This issue is designed for tracking errata in the book text. Things like minor typos.

*Page 65*

On page 65 of the Preliminary Draft there is a typo: "form" should be "from." This is found in the following text:

"Bypassing the Thread Pool If you don’t want a task to use a worker thread **form** the thread pool, you can create a new thread for its dedicated use."
Comments: ** Comment from web user: jsh003guardz63 **

The explanation of performance characteristics of pipelines on p121 needs clarification and possibly correction.

In particular, “If there are enough available cores and if all stages of a pipeline take an equal amount of time, the execution time for the pipeline as a whole is the same as the time for just one stage”, would appear to be incorrect.

If:
S = the number of stages in figure 4 (= 4)
N = the number of images in figure 4 (= 6)
T = the duration to process one stage for one image (= t1 - t0 = t2 – t1)
Then:
The elapsed time for a completely parallel and completely efficient pipeline is (S+N-1) * T.
The elapsed time taken for one stage = N * T.
But:
(S+N-1) x T is only equal to N * T when S = 1.

If the sidebar on p121 is referring to the case where S = 1, the phrasing of that sidebar needs changing, as it currently implies that (S+N-1) x T = N * T for any S. If the sidebar isn't referring to the case where S = 1, it needs re-wording as its intended meaning is unclear and/or incorrect. Possibly, it's trying to say that each core has an equal execution time, but I don't think that's the intended meaning...

Commented Issue: Book Errata [18665]

jsh003guardz63 — Tue, 19 Jul 2011 14:31:45 GMT

Minor typo - page 99. "that implied data flow" should be "than implied data flow".

Commented Issue: Book Errata [18665]

AdeMiller — Wed, 12 Jan 2011 04:23:23 GMT

The ISBN you got from O’Reilly, listed below, and which is printed inside the Parallel Programming with Microsoft.NET book on the copyright page is incorrect. If you go to Amazon, for instance, and search by ISBN number, the book you get is Parallel Programming with Microsoft Visual Studio 2010 Step by Step, written by Doris Marshall. The ISBN number for our book is listed on Amazon as 9780735651593. The bar code on the back of our book is the correct number. It’s not a really big deal but just thought you should know.

Commented Issue: Book Errata [18665]

AdeMiller — Thu, 06 Jan 2011 21:55:50 GMT

Additional Errata submitted to O'Reilly:

PDF Minor technical mistake Change "the loop index" to "a loop-indexed variable" Loop-carried dependence. If the body of a parallel for loop performs arithmetic on the loop index, there is likely to be a dependency that is known as loop-carried dependence. Should be Loop-carried dependence. If the body of a parallel for loop performs arithmetic on…
View details and modify status »

Submitted by: Steve Bearman
Send a response »
Needs Review

PDF Minor technical mistake Change "percent" to "fraction": Amdahl�s law says that no matter how many cores you have, the maximum speedup you can ever achieve is (1 / percent of time spent in sequential processing). Should be Amdahl�s law says that no matter how many cores you have, the maximum speedup you can ever achieve…
View details and modify status »

Submitted by: Steve Bearman
Send a response »
Needs Review

Printed, PDF Language change or typo The word consumer is mispelled in the following sentence: Pipelines are expressions of a general technique known as producer/ cotnsumer.
View details and modify status »

Submitted by: genegw
Send a response »
Needs Review

Created Issue: Trend.Predict(double ordinate) [19190]

gfoidl — Sun, 28 Nov 2010 21:54:12 GMT

The argument of Microsoft.Practices.ParallelGuideSamples.Utilities.Trend.Predict(double ordinate) should be called abscissa according to the mathematical definition.

Closed Issue: ImagePipeline UI uses Textboxes for readonly data [18668]

AdeMiller — Wed, 29 Sep 2010 11:40:42 GMT

The content in the textboxes is readonly but the textboxes can be typed in by the user. This is a little odd but doesn't cause any actual issues with the application.
Comments:

Fixed by changeset 54191.

Closed Issue: ToGray GrayLuma not called in ImagePipeline\BitmapExtensions.cs [18669]

AdeMiller — Wed, 29 Sep 2010 11:40:32 GMT

Uncalled code. Remove to make things easier to read.
Comments:

Fixed by changeset 54191.

Closed Issue: F# ImagePipeline throws IndexOutOfRangeException [18703]

AdeMiller — Wed, 29 Sep 2010 11:03:07 GMT

Twice now when I’ve run the ImagePipeline sample I’ve had an IndexOutOfRangeException when I press the “stop” button. The exception is reraised but the stacktrace seems to indicate a problem here in BlockingMultiplexer.fs

let index =
(producers |> Seq.filter (fun info -> info.Collection = waitList.[waitListIndex])
|> Seq.map (fun info -> info.Index)).Single()
let lockId = lockOrderFn (!item)

Tomas, you might like to quickly review the translation from C#. Otherwise I guess it’s a bug in the original C# code.
Comments:

Fixed with changeset 54190.

Commented Issue: F# ImagePipeline throws IndexOutOfRangeException [18703]

AdeMiller — Wed, 29 Sep 2010 11:01:45 GMT

The behavior I see is that the UI locks up.

Commented Issue: F# ImagePipeline throws IndexOutOfRangeException [18703]

AdeMiller — Tue, 07 Sep 2010 21:46:35 GMT

I believe this might be a bug in the port.

The C# code

if (waitList.Length == 0)
{
if (token.IsCancellationRequested)
yield break;
else
throw new InvalidOperationException("Producer blocking collections omitted required value " +
nextLockId.ToString());
}

was ported into F# as
if waitList.Length = 0 then
if token.IsCancellationRequested then
yieldBreak := true
else
let msg = "Producer blocking collections omitted required value " + nextLockId.ToString()
raise (new InvalidOperationException(msg))

I think this is a bug because in F# you have allowed a loop to continue that should be terminated. (I think you need to add some kind of loop break or nonlocal exit to the F# so that you break out of the current iteration.)

Commented Issue: F# ImagePipeline throws IndexOutOfRangeException [18703]

AdeMiller — Tue, 07 Sep 2010 18:40:35 GMT

Doesn't repro on C# version.

Created Issue: ImagePipeline throws IndexOutOfRangeException [18703]

AdeMiller — Tue, 07 Sep 2010 16:08:20 GMT