Re: [Openexr-devel] Optimisation code path for reading images
Piotr Stanczyk <[email protected]> Tue, 18 Jun 2013 16:51:03 +0000
| Newsgroups | gmane.comp.video.openexr.user,gmane.comp.video.openexr.devel |
|---|---|
| Message-ID | <95F01873F66DA54DAD2734EEBF98F86C011092552C@mailbox09.lucas.alllucas.com> |
--===============0847826746896101182== Content-Language: en-GB Content-Type: multipart/alternative; boundary="_000_95F01873F66DA54DAD2734EEBF98F86C011092552Cmailbox09luca_" --_000_95F01873F66DA54DAD2734EEBF98F86C011092552Cmailbox09luca_ Content-Type: text/plain; charset="Windows-1252" Content-Transfer-Encoding: quoted-printable aha - ok. Thanks - Piotr ________________________________ From: Halfdan Ingvarsson [[email protected]] Sent: 18 June 2013 09:49 To: Piotr Stanczyk Cc: Juri Abramov; [email protected]; [email protected] Subject: Re: [Openexr-devel] Optimisation code path for reading images Actually, I should clarify. The corruption occurs on *read*, not *write*. A= ttached is a tiff file converted from a PIZ compressed exr, also attached. It's most noticeable in the alpha channel, although there's a definite glit= ch in the red channel just above the plane. - =BD On 13-06-18 12:44 PM, Piotr Stanczyk wrote: Thanks. Let me see if I can repro this. I presume that simply writing out PIZ files= should flare up. Were you also seeing test failures? - Piotr ________________________________ From: Juri Abramov [[email protected]<mailto:[email protected]>] Sent: 18 June 2013 09:26 To: Piotr Stanczyk; Halfdan Ingvarsson Cc: [email protected]<mailto:[email protected]>; openexr-devel@= nongnu.org<mailto:[email protected]> Subject: RE: [Openexr-devel] Optimisation code path for reading images I=92m afraid mine are under customer=92s NDA. It is rather common though, the corruption was reported for alpha. Juri From: Piotr Stanczyk [mailto:[email protected]] Sent: 18 June, 2013 18:16 To: Juri Abramov; Halfdan Ingvarsson Cc: [email protected]<mailto:[email protected]>; openexr-devel@= nongnu.org<mailto:[email protected]> Subject: RE: [Openexr-devel] Optimisation code path for reading images Is this something that was quite deterministic in behaviour? (Does anyone have an example of a corrupted image?) - Piotr ________________________________ From: Juri Abramov [[email protected]<mailto:[email protected]>] Sent: 18 June 2013 09:13 To: Halfdan Ingvarsson; Piotr Stanczyk Cc: [email protected]<mailto:[email protected]>; openexr-devel@= nongnu.org<mailto:[email protected]> Subject: RE: [Openexr-devel] Optimisation code path for reading images Hi Halfdan, Yep, we ran into this one too. Was already the case for OpenEXR 1.6/1.7. Would be good to know if it is fixed in 4.5. Juri From: [email protected]<mailto:openexr= [email protected]> [mailto:openexr-devel-boun= [email protected]] On Behalf Of Halfdan Ingvarsson Sent: 18 June, 2013 18:06 To: Piotr Stanczyk Cc: [email protected]<mailto:[email protected]>; openexr-devel@= nongnu.org<mailto:[email protected]> Subject: Re: [Openexr-devel] Optimisation code path for reading images Also, compiling the EXR2.0 library using gcc 4.2/4.3/4.4 with -O3 results i= n the PIZ compression code producing occasional garbage data (-O3 is the de= fault for cmake release builds). The workaround was to build with -O2. gcc = 4.6 and up seem ok. I didn't test with gcc 4.5. Unfortunately, I didn't have time to dig any deeper as to whether this was = an actual optimizer bug, or whether there are some spurious assumptions in = the code leading to it. Just something to keep in mind. - =BD On 13-06-17 09:35 PM, Piotr Stanczyk wrote: Hi, There have been a few usage cases reported that exercised parts of the opti= mised reading code path which unfortunately revealed erroneous assumptions = in the source. We have a fix for handling these, which also extents to handling more gener= al cases, in a separate branch and will be releasing that once we have bui= lt up more usage cycles. In the meantime, however, a v2.0.1 release will be available shortly, which= disables the optimisation. - Piotr _______________________________________________ Openexr-devel mailing list [email protected]<mailto:[email protected]> https://lists.nongnu.org/mailman/listinfo/openexr-devel ________________________________ This email message is for the sole use of the intended recipient(s) and may= contain confidential information. Any unauthorized review, use, disclosur= e or distribution is prohibited. If you are not the intended recipient, pl= ease contact the sender by reply email and destroy all copies of the origin= al message. ________________________________ --_000_95F01873F66DA54DAD2734EEBF98F86C011092552Cmailbox09luca_ Content-Type: text/html; charset="Windows-1252" Content-Transfer-Encoding: quoted-printable <html dir=3D"ltr"> <head> <meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3DWindows-1= 252"> </head> <body ocsi=3D"0" fpstyle=3D"1" bgcolor=3D"#FFFFFF"> <div style=3D"direction: ltr;font-family: Tahoma;color: #000000;font-size: = 10pt;">aha - ok.<br> <br> Thanks <br> <div><br> <div style=3D"font-family: Tahoma; font-size: 13px;">- Piotr<br> </div> </div> <div style=3D"font-family: Times New Roman; color: rgb(0, 0, 0); font-size:= 16px;"> <hr tabindex=3D"-1"> <div style=3D"direction: ltr;" id=3D"divRpF673951"><font color=3D"#000000" = face=3D"Tahoma" size=3D"2"><b>From:</b> Halfdan Ingvarsson [halfdan@sidefx.= com]<br> <b>Sent:</b> 18 June 2013 09:49<br> <b>To:</b> Piotr Stanczyk<br> <b>Cc:</b> Juri Abramov; [email protected]; [email protected]<= br> <b>Subject:</b> Re: [Openexr-devel] Optimisation code path for reading imag= es<br> </font><br> </div> <div></div> <div> <div class=3D"moz-cite-prefix">Actually, I should clarify. The corruption o= ccurs on *read*, not *write*. Attached is a tiff file converted from a PIZ = compressed exr, also attached.<br> <br> It's most noticeable in the alpha channel, although there's a definite glit= ch in the red channel just above the plane.<br> <br> - =BD<br> <br> On 13-06-18 12:44 PM, Piotr Stanczyk wrote:<br> </div> <blockquote type=3D"cite"><style>=0A= <!--=0A= @font-face=0A= {font-family:PMingLiU}=0A= @font-face=0A= {font-family:PMingLiU}=0A= @font-face=0A= {font-family:Tahoma}=0A= @font-face=0A= {font-family:Verdana}=0A= @font-face=0A= {font-family:Consolas}=0A= p.MsoNormal, li.MsoNormal, div.MsoNormal=0A= {margin:0in;=0A= margin-bottom:.0001pt;=0A= font-size:12.0pt;=0A= font-family:"Times New Roman","serif";=0A= color:black}=0A= a:link, span.MsoHyperlink=0A= {color:blue;=0A= text-decoration:underline}=0A= a:visited, span.MsoHyperlinkFollowed=0A= {color:purple;=0A= text-decoration:underline}=0A= p=0A= {margin:0in;=0A= margin-bottom:.0001pt;=0A= font-size:12.0pt;=0A= font-family:"Times New Roman","serif";=0A= color:black}=0A= pre=0A= {margin:0in;=0A= margin-bottom:.0001pt;=0A= font-size:10.0pt;=0A= font-family:"Courier New";=0A= color:black}=0A= p.MsoAcetate, li.MsoAcetate, div.MsoAcetate=0A= {margin:0in;=0A= margin-bottom:.0001pt;=0A= font-size:8.0pt;=0A= font-family:"Tahoma","sans-serif";=0A= color:black}=0A= span.HTMLPreformattedChar=0A= {font-family:Consolas;=0A= color:black}=0A= p.msochpdefault, li.msochpdefault, div.msochpdefault=0A= {margin:0in;=0A= margin-bottom:.0001pt;=0A= font-size:10.0pt;=0A= font-family:"Times New Roman","serif";=0A= color:black}=0A= span.htmlpreformattedchar0=0A= {font-family:Consolas;=0A= color:black}=0A= span.emailstyle20=0A= {font-family:"Verdana","sans-serif";=0A= color:#17365D;=0A= font-weight:bold}=0A= span.EmailStyle23=0A= {font-family:"Verdana","sans-serif";=0A= color:#17365D;=0A= font-weight:bold}=0A= span.BalloonTextChar=0A= {font-family:"Tahoma","sans-serif";=0A= color:black}=0A= .MsoChpDefault=0A= {font-size:10.0pt}=0A= @page WordSection1=0A= {margin:1.0in 1.0in 1.0in 1.0in}=0A= -->=0A= </style><style id=3D"owaParaStyle" type=3D"text/css">=0A= <!--=0A= p=0A= {margin-top:0;=0A= margin-bottom:0}=0A= -->=0A= BODY {direction: ltr;font-family: Tahoma;color: #000000;font-size: 10pt;}P = {margin-top:0;margin-bottom:0;}</style> <div style=3D"direction: ltr; font-family: Tahoma; color: rgb(0, 0, 0); fon= t-size: 10pt;"> Thanks. <br> <br> Let me see if I can repro this. I presume that simply writing out PIZ files= should flare up. <br> <br> Were you also seeing test failures?<br> <div><br> <div style=3D"font-family: Tahoma; font-size: 13px;">- Piotr<br> </div> </div> <div style=3D"font-family: Times New Roman; color: rgb(0, 0, 0); font-size:= 16px;"> <hr tabindex=3D"-1"> <div id=3D"divRpF871998" style=3D"direction: ltr;"><font color=3D"#000000" = face=3D"Tahoma" size=3D"2"><b>From:</b> Juri Abramov [<a class=3D"moz-txt-l= ink-abbreviated" href=3D"mailto:[email protected]" target=3D"_blank">gabr= [email protected]</a>]<br> <b>Sent:</b> 18 June 2013 09:26<br> <b>To:</b> Piotr Stanczyk; Halfdan Ingvarsson<br> <b>Cc:</b> <a class=3D"moz-txt-link-abbreviated" href=3D"mailto:openexr-use= [email protected]" target=3D"_blank"> [email protected]</a>; <a class=3D"moz-txt-link-abbreviated" href=3D"= mailto:[email protected]" target=3D"_blank"> [email protected]</a><br> <b>Subject:</b> RE: [Openexr-devel] Optimisation code path for reading imag= es<br> </font><br> </div> <div> <div class=3D"WordSection1"> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Verdana","sans-serif"; color: rgb(23, 54, 93);">I=92m afra= id mine are under customer=92s NDA. </span></b></p> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Verdana","sans-serif"; color: rgb(23, 54, 93);">It is rath= er common though, the corruption was reported for alpha.</span></b></p> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Verdana","sans-serif"; color: rgb(23, 54, 93);"> </sp= an></b></p> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Verdana","sans-serif"; color: rgb(23, 54, 93);">Juri</span= ></b></p> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Verdana","sans-serif"; color: rgb(23, 54, 93);"> </sp= an></b></p> <div> <div style=3D"border-width: 1pt medium medium; border-style: solid none non= e; padding: 3pt 0in 0in;"> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Tahoma","sans-serif"; color: windowtext;">From:</span></b>= <span style=3D"font-size: 10pt; font-family: "Tahoma","sans-= serif"; color: windowtext;"> Piotr Stanczyk [<a class=3D"moz-txt-link-= freetext" href=3D"mailto:[email protected]" target=3D"_blank">mailto:pstanc= [email protected]</a>] <br> <b>Sent:</b> 18 June, 2013 18:16<br> <b>To:</b> Juri Abramov; Halfdan Ingvarsson<br> <b>Cc:</b> <a class=3D"moz-txt-link-abbreviated" href=3D"mailto:openexr-use= [email protected]" target=3D"_blank"> [email protected]</a>; <a class=3D"moz-txt-link-abbreviated" href=3D"= mailto:[email protected]" target=3D"_blank"> [email protected]</a><br> <b>Subject:</b> RE: [Openexr-devel] Optimisation code path for reading imag= es</span></p> </div> </div> <p class=3D"MsoNormal"> </p> <div> <p class=3D"MsoNormal"><span style=3D"font-size: 10pt; font-family: "T= ahoma","sans-serif";">Is this something that was quite deter= ministic in behaviour?<br> (Does anyone have an example of a corrupted image?)</span></p> <div> <p class=3D"MsoNormal"><span style=3D"font-size: 10pt; font-family: "T= ahoma","sans-serif";"> </span></p> <div> <p class=3D"MsoNormal"><span style=3D"font-size: 10pt; font-family: "T= ahoma","sans-serif";">- Piotr</span></p> </div> </div> <div> <div class=3D"MsoNormal" style=3D"text-align: center;" align=3D"center"> <hr align=3D"center" size=3D"3" width=3D"100%"> </div> <div id=3D"divRpF863779"> <p class=3D"MsoNormal" style=3D"margin-bottom: 12pt;"><b><span style=3D"fon= t-size: 10pt; font-family: "Tahoma","sans-serif";">From= :</span></b><span style=3D"font-size: 10pt; font-family: "Tahoma"= ,"sans-serif";"> Juri Abramov [<a class=3D"moz-txt-link-abbreviat= ed" href=3D"mailto:[email protected]" target=3D"_blank">[email protected]= om</a>]<br> <b>Sent:</b> 18 June 2013 09:13<br> <b>To:</b> Halfdan Ingvarsson; Piotr Stanczyk<br> <b>Cc:</b> <a href=3D"mailto:[email protected]" target=3D"_blank">ope= [email protected]</a>; <a href=3D"mailto:[email protected]" target=3D"_blank">openexr-devel= @nongnu.org</a><br> <b>Subject:</b> RE: [Openexr-devel] Optimisation code path for reading imag= es</span></p> </div> <div> <div> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Verdana","sans-serif"; color: rgb(23, 54, 93);">Hi Halfdan= ,</span></b></p> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Verdana","sans-serif"; color: rgb(23, 54, 93);"> </sp= an></b></p> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Verdana","sans-serif"; color: rgb(23, 54, 93);">Yep, we ra= n into this one too. Was already the case for OpenEXR 1.6/1.7.</span></b></= p> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Verdana","sans-serif"; color: rgb(23, 54, 93);">Would be g= ood to know if it is fixed in 4.5.</span></b></p> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Verdana","sans-serif"; color: rgb(23, 54, 93);"> </sp= an></b></p> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Verdana","sans-serif"; color: rgb(23, 54, 93);">Juri</span= ></b></p> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Verdana","sans-serif"; color: rgb(23, 54, 93);"> </sp= an></b></p> <div> <div style=3D"border-width: 1pt medium medium; border-style: solid none non= e; padding: 3pt 0in 0in;"> <p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: &quo= t;Tahoma","sans-serif"; color: windowtext;">From:</span></b>= <span style=3D"font-size: 10pt; font-family: "Tahoma","sans-= serif"; color: windowtext;"> <a href=3D"mailto:openexr-devel-bounces+[email protected]= g" target=3D"_blank"> openexr-devel-bounces+[email protected]</a> [<a href=3D"= mailto:openexr-devel-bounces+[email protected]" target= =3D"_blank">mailto:openexr-devel-bounces+[email protected]= rg</a>] <b>On Behalf Of </b>Halfdan Ingvarsson<br> <b>Sent:</b> 18 June, 2013 18:06<br> <b>To:</b> Piotr Stanczyk<br> <b>Cc:</b> <a href=3D"mailto:[email protected]" target=3D"_blank">ope= [email protected]</a>; <a href=3D"mailto:[email protected]" target=3D"_blank">openexr-devel= @nongnu.org</a><br> <b>Subject:</b> Re: [Openexr-devel] Optimisation code path for reading imag= es</span></p> </div> </div> <p class=3D"MsoNormal"> </p> <div> <p class=3D"MsoNormal">Also, compiling the EXR2.0 library using gcc 4.2/4.3= /4.4 with -O3 results in the PIZ compression code producing occasional garb= age data (-O3 is the default for cmake release builds). The workaround was = to build with -O2. gcc 4.6 and up seem ok. I didn't test with gcc 4.5.<br> <br> Unfortunately, I didn't have time to dig any deeper as to whether this was = an actual optimizer bug, or whether there are some spurious assumptions in = the code leading to it.<br> <br> Just something to keep in mind.<br> <br> - =BD<br> <br> On 13-06-17 09:35 PM, Piotr Stanczyk wrote:</p> </div> <blockquote style=3D"margin-top: 5pt; margin-bottom: 5pt;"> <div> <p class=3D"MsoNormal" style=3D"margin-bottom: 12pt;"><span style=3D"font-s= ize: 10pt; font-family: "Tahoma","sans-serif";"><br> Hi, <br> <br> There have been a few usage cases reported that exercised parts of the opti= mised reading code path which unfortunately revealed erroneous assumptions = in the source. <br> <br> We have a fix for handling these, which also extents to handling more gener= al cases, in a separate branch and will be releasing that once we hav= e built up more usage cycles. <br> <br> In the meantime, however, a v2.0.1 release will be available shortly, which= disables the optimisation.</span></p> <div> <p class=3D"MsoNormal" style=3D"margin-bottom: 12pt;"><span style=3D"font-s= ize: 10pt; font-family: "Tahoma","sans-serif";"> <= /span></p> <div> <p class=3D"MsoNormal"><span style=3D"font-size: 10pt; font-family: "T= ahoma","sans-serif";">- Piotr</span></p> </div> </div> </div> <p class=3D"MsoNormal" style=3D"margin-bottom: 12pt;"><br> <br> </p> <pre>_______________________________________________</pre> <pre>Openexr-devel mailing list</pre> <pre><a href=3D"mailto:[email protected]" target=3D"_blank">Openexr-= [email protected]</a></pre> <pre><a href=3D"https://lists.nongnu.org/mailman/listinfo/openexr-devel" ta= rget=3D"_blank">https://lists.nongnu.org/mailman/listinfo/openexr-devel</a>= </pre> </blockquote> <p class=3D"MsoNormal"> </p> </div> <div> <div class=3D"MsoNormal" style=3D"text-align: center;" align=3D"center"> <hr align=3D"center" size=3D"3" width=3D"100%"> </div> </div> <div> <p class=3D"MsoNormal">This email message is for the sole use of the intend= ed recipient(s) and may contain confidential information. Any unautho= rized review, use, disclosure or distribution is prohibited. If you a= re not the intended recipient, please contact the sender by reply email and destroy all copies of the original message. = </p> </div> <div> <div class=3D"MsoNormal" style=3D"text-align: center;" align=3D"center"> <hr align=3D"center" size=3D"3" width=3D"100%"> </div> </div> </div> </div> </div> </div> </div> </div> </div> </blockquote> <br> </div> </div> </div> </body> </html> --_000_95F01873F66DA54DAD2734EEBF98F86C011092552Cmailbox09luca_-- --===============0847826746896101182== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Openexr-user mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/openexr-user --===============0847826746896101182==--