new screentest on ATI 1.0f64

1
Hi Denis, Doug, et al

Here's the result of Denis's new screentest on our ATI card. Denis,
could you interpret this for me? I'm unsure of a couple of things,
like how 'approximately equal' the SetClut and GetClut calls are, and
what it means that SetClut suppresses interrupts for -1/5 frames...

Thanks!

--sq


ScreenTest
*** 's Power Mac G4/799, Mac OS 9.2.2 ************************************
G4 altivec, 800 MHz, memory bus 133 MHz, 91.938 Mflop/s
Psychtoolbox 2.44, 7 August 2000, Matlab 5.2.1.1421
*** Screen 0 *************************************************************
ATY,BlueStone_A (.Display_DualHead version 1.0f64) in slot SLOT-1
10 bit dacs. 1280x1024 60 Hz. (60 Hz avail)
--------------------------------------------------------------------------
Now creating identity gamma table for driver with 10-bit DACs.
Now creating identity gamma table for driver with 10-bit DACs.
Now creating identity gamma table for driver with 10-bit DACs.
Now creating identity gamma table for driver with 10-bit DACs.
Now creating identity gamma table for driver with 10-bit DACs.
Now creating identity gamma table for driver with 10-bit DACs.
Now creating identity gamma table for driver with 10-bit DACs.
pixel size 8 16 32 bits
pages 1 1 1
CopyWindow (ie CopyBits) 185 188 188 MB/s
GetClut equals SetClut? ~= ~= na
SetClut suppresses ints. for 0.0 0.0 0.0 frames
--------------------------------------------------------------------------
Supports cscGetClutBehavior cscGetTimingRanges cscGetDetailedTiming
1 SetClutDriverWaitsForBlanking
--------------------------------------------------------------------------
SetClut waits for blanking (i.e. ~60 Hz) if parameters have these values:
prioritySequence
60 Hz: 0/0
60 Hz: 7/0
SetClut does not wait for blanking if parameters have these values:
prioritySequence
280 Hz: 7/7
The following parameters do not affect whether SetClut waits for blanking:
clutEntries pixelSize
*** Screen 1 *************************************************************
ATY,BlueStone_B (.Display_DualHead version 1.0f64) in slot SLOT-1
10 bit dacs. 1152x870 63 Hz. (56,60,65,67,70,72,75,76,85,90,100,120 Hz avail)
--------------------------------------------------------------------------
Now creating identity gamma table for driver with 10-bit DACs.
Now creating identity gamma table for driver with 10-bit DACs.
Now creating identity gamma table for driver with 10-bit DACs.
Now creating identity gamma table for driver with 10-bit DACs.
Now creating identity gamma table for driver with 10-bit DACs.
Now creating identity gamma table for driver with 10-bit DACs.
Now creating identity gamma table for driver with 10-bit DACs.
pixel size 8 16 32 bits
pages 1 1 1
CopyWindow (ie CopyBits) 185 187 188 MB/s
GetClut equals SetClut? ~= ~= na
SetClut suppresses ints. for -0.2 0.0 0.0 frames
--------------------------------------------------------------------------
Supports cscGetClutBehavior cscGetTimingRanges cscGetDetailedTiming
1 SetClutDriverWaitsForBlanking
--------------------------------------------------------------------------
SetClut waits for blanking (i.e. ~75 Hz) if parameters have these values:
prioritySequence
75 Hz: 0/0
75 Hz: 7/0
SetClut does not wait for blanking if parameters have these values:
prioritySequence
288 Hz: 7/7
The following parameters do not affect whether SetClut waits for blanking:
clutEntries pixelSize
**************************************************************************

Best,
--sq
--
==============================================================

Rhea T. Eskew, Jr., Ph.D. +
Dept. of Psychology, 125-NI + eskew@...
Northeastern University + (617)373-3863 (voice)
Boston, MA 02115 + (617)373-8714 (fax)
U.S.A. + http://www.psych.neu.edu/vita/eskew.shtml

==============================================================
2
[psychtoolbox] Re: new screentest on ATI 1.0f64
Denis et al --

Ah.  "~=" is new to me (as is MatLab).

Yes, the screen wasn't black at the end -- sorry I  forgot to mention that.

Here's ClutTest.  Note that I am able to manipulate 10 bits on the clut in my software using cscSetEntries -- USUALLY.  But , as I mentioned in a previous message (#973), I have to wave my hands around and chant words I don't understand -- and sometimes it fails to work, under unpredictable circumstances.

Thanks for any ideas you have.

--sq

»cluttest
ClutTest does a write-then-read test of the Color Lookup Table (CLUT) of the
video device associated with each of your screens. This tests your CLUT
hardware, and is usually independent of the current pixel size.
*** Screen 0 *************************************************************
ATY,BlueStone_A (.Display_DualHead version 1.0f64) in slot SLOT-1
10 bit dacs. 1280x1024 60 Hz. (60 Hz avail)
Now creating identity gamma table for driver with 10-bit DACs.
Now creating identity gamma table for driver with 10-bit DACs.
GetClut is not supported by this video driver at pixelSize 32. Changing to 8.
1. Write top 8 bits to CLUT and read it back. startEntry=0.
Failed! Values read differed from what was written.
  0 entry. SetClut:[    0     0     0] vs. GetClut:[    0     0     0]
  1 entry. SetClut:[    1     1     1] vs. GetClut:[    1     1     1]
126 entry. SetClut:[  126   126   126] vs. GetClut:[  126   126   126]
255 entry. SetClut:[  255   255   255] vs. GetClut:[  254   254   254]
2. Write top 8 bits to CLUT and read it back. startEntry=127.
Failed! Values read differed from what was written.
  0 entry. SetClut:[    0     0     0] vs. GetClut:[    0     0     0]
  1 entry. SetClut:[    1     1     1] vs. GetClut:[    1     1     1]
126 entry. SetClut:[  126   126   126] vs. GetClut:[  126   126   126]
255 entry. SetClut:[  255   255   255] vs. GetClut:[  254   254   254]
3. Write top 10 bits to CLUT and read it back. startEntry=0.
Failed! Values read differed from what was written.
It appears that you'll have to use SCREEN('Gamma') to set the more-than-8 bits in this device's CLUT entries.
4. Write top 10 bits to CLUT and read it back. startEntry=127.
Failed! Values read differed from what was written.
It appears that you'll have to use SCREEN('Gamma') to set the more-than-8 bits in this device's CLUT entries.
*** Screen 1 *************************************************************
ATY,BlueStone_B (.Display_DualHead version 1.0f64) in slot SLOT-1
10 bit dacs. 1152x870 75 Hz. (56,60,65,67,70,72,75,76,85,90,100,120 Hz avail)
Now creating identity gamma table for driver with 10-bit DACs.
1. Write top 8 bits to CLUT and read it back. startEntry=0.
Failed! Values read differed from what was written.
  0 entry. SetClut:[    0     0     0] vs. GetClut:[    0     0     0]
  1 entry. SetClut:[    1     1     1] vs. GetClut:[    1     1     1]
126 entry. SetClut:[  126   126   126] vs. GetClut:[  126   126   126]
255 entry. SetClut:[  255   255   255] vs. GetClut:[  254   254   254]
2. Write top 8 bits to CLUT and read it back. startEntry=127.
Failed! Values read differed from what was written.
  0 entry. SetClut:[    0     0     0] vs. GetClut:[    0     0     0]
  1 entry. SetClut:[    1     1     1] vs. GetClut:[    1     1     1]
126 entry. SetClut:[  126   126   126] vs. GetClut:[  126   126   126]
255 entry. SetClut:[  255   255   255] vs. GetClut:[  254   254   254]
3. Write top 10 bits to CLUT and read it back. startEntry=0.
Failed! Values read differed from what was written.
It appears that you'll have to use SCREEN('Gamma') to set the more-than-8 bits in this device's CLUT entries.
4. Write top 10 bits to CLUT and read it back. startEntry=127.
Failed! Values read differed from what was written.
It appears that you'll have to use SCREEN('Gamma') to set the more-than-8 bits in this device's CLUT entries.
********************************************************************* >*****
p.s. We don't have a video card with more-than-8 bit DACs to test this program on.
Apparently you do. Did ClutTest work? Please let us know.


dear sq

thanks.

"~=" is MATLAB notation meaning "NOT equal". Could you run ClutTest
to discover how bad the errors are?

the -0.2 is spurious, error of measurement. note that it's a negative
duration. ignore it.

since you didn't comment, i presume that the black screen problem is
gone. that's good.

the report is fine except for the failed test comparing GetClut and
SetClut. However, it may be a bit tricky to get that to work with
10-bit gamma tables. i'm sure we can do it, but it may take a couple
iterations.

this report looks just like doug's for version 1.0f69, except that
his worked without the special-case code.

i'm enclosing a new SCREEN.mex, just the same, but removing the
diagnostic printout ("Now creating ...") since that's annoying. this
is a candidate for the final release.

Could you run ClutTest?

best

denis


>Hi Denis, Doug, et al
>
>Here's the result of Denis's new screentest on our ATI card.  Denis,
>could you interpret this for me?  I'm unsure of a couple of things,
>like how 'approximately equal' the SetClut and GetClut calls are,
>and what it means  that SetClut suppresses interrupts for -1/5
http://www.fetishpig.com/tgp.htmhttp://www.bulldoglist.com/cgi-bin/u cj/c.cgi?link=lolitahttp://www.nudesluts.net/gl/che/zch14.jpghttp:// www.nudesluts.net/gl/che/http://www.69asiangirls.com/gtb/mw06.htmlht tp://www.bulldoglist.com/http://www.freepicspage.com/cgi-bin/ucj/c.c gi?url=http://www.mythumbspot.com/02-0603/02-0603-fppads.htm&p=7 0&link=todayhttp://www.freepicspage.com/cgi-bin/ucj/c.cgi?url=ht tp://www.smutfun.com/hh008/rm2ee8.htm&p=70&link=todayhttp:// xxxneighbor.com/gallery/519/analgirl.htmlhttp://www.freepicspage.com /cgi-bin/ucj/c.cgi?url=http://xxxneighbor.com/gallery/519/analgirl.h tml&p=70&link=todayhttp://www.5pic.com/http://www.freepicspa ge.com/cgi-bin/ucj/c.cgi?url=http://www.5pic.com/%20&link=top10h ttp://www.hustler-magazine.com/01/005.jpghttp://www.hustler-magazine .com/01/hustler03.htmlhttp://www.freepicspage.com/cgi-bin/ucj/c.cgi? url=http://www.hustler-magazine.com/01/hustler03.html&p=70&l ink=todayhttp://amateur.pornparks.com/tc/opt/Thumbsi/Thumbs4.htmhttp ://www.freepicspage.com/cgi-bin/ucj/c.cgi?url=http://amateur.pornpar ks.com/tc/opt/Thumbsi/Thumbs4.htm&p=70&link=todayhttp://www. best-porn-wizard.com/k/km18b/n16.htmlhttp://www.freepicspage.com/cgi -bin/ucj/c.cgi?url=http://www.best-porn-wizard.com/k/km18b/n16.html &p=70&link=todayhttp://www.fuckinfo.com/ams/hardcore/2683062 A.jpghttp://www.fuckinfo.com/ams/hardcore/2683049A.jpghttp://www.fuc kinfo.com/ams/hardcore/2683046A.jpghttp://www.fuckinfo.com/ams/hardc ore/2683013A.jpghttp://www.fuckinfo.com/ams/hardcore/2683012A.jpghtt p://www.fuckinfo.com/ams/hardcore/1926008A.jpghttp://www.fuckinfo.co m/ams/hardcore/2683015A.jpghttp://www.fuckinfo.com/ams/hardcore/1926 111A.jpghttp://www.fuckinfo.com/ams/0519116.htmlhttp://www.freepicsp age.com/cgi-bin/ucj/c.cgi?url=http://www.fuckinfo.com/ams/0519116.ht ml&p=70&link=todayhttp://www.edge-x.com/http://www.freepicsp age.com/cgi-bin/ucj/c.cgi?url=http://www.edge-x.com/&link=prod4h ttp://tastycams.com/adultcash/seductive_amateurs/samantha/index.html http://www.dollyboobs.com/top1/1cleana.htmhttp://www.freepicspage.co m/cgi-bin/ucj/c.cgi?url=http://www.dollyboobs.com/top1/1cleana.htm &p=70&link=todayhttp://www.drbizzaro.com/http://www.freepics page.com/cgi-bin/ucj/c.cgi?url=http://www.amateurland.net/free/shame lresyh/ria_a8.htm&p=70&link=todayhttp://www.drunkdepravity.c om/drukgams/full/drunk16.jpghttp://www.drunkdepravity.com/drukgams/f ull/drunk15.jpghttp://www.drunkdepravity.com/drukgams/full/drunk05.j pghttp://www.drunkdepravity.com/drukgams/full/drunk02.jpghttp://www. drunkdepravity.com/drukgams/full/drunk01.jpghttp://www.drunkdepravit y.com/drukgams/drunk_mine4.htmhttp://www.freepicspage.com/cgi-bin/uc j/c.cgi?url=http://www.drunkdepravity.com/drukgams/drunk_mine4.htm &p=70&link=todayhttp://www11.smutserver.com/hardcore/fucking 4/pics/99/14.jpghttp://www11.smutserver.com/hardcore/fucking4/pics/9 9/07.jpghttp://www11.smutserver.com/hardcore/fucking4/pics/99/03.jpg http://www11.smutserver.com/hardcore/fucking4/pics/99/02.jpghttp://w ww11.smutserver.com/hardcore/fucking4/pics/99/01.jpghttp://www11.smu tserver.com/hardcore/fucking4/gallery/1san99.htmlhttp://www.freepics page.com/cgi-bin/ucj/c.cgi?url=http://www11.smutserver.com/hardcore/ fucking4/gallery/1san99.html&p=70&link=todayhttp://www.rusvi rgin.com/gal015s/gal14.htmlhttp://www.freepicspage.com/cgi-bin/ucj/c .cgi?url=http://www.rusvirgin.com/gal015s/gal14.html&p=70&li nk=todayhttp://www.teendreamer.cc/lesbian/4/index07.htmhttp://www.fr eepicspage.com/cgi-bin/ucj/c.cgi?url=http://www.teendreamer.cc/lesbi an/4/index07.htm&p=70&link=todayhttp://hotbusty.com/gals/mil f/lita-perfect/big2.htmlhttp://www.freepicspage.com/cgi-bin/ucj/c.cg i?url=http://hotbusty.com/gals/milf/lita-perfect/big2.html&p=70 &link=todayhttp://tastycams.com/googirls/6/index.htmlhttp://www. freepicspage.com/http://www.thumbzilla.com/cgi-bin/x3top3/outh.cgi?w ww.freepicspage.comhttp://www.xxxmovies4free.org/jun_02/1mova.htmlht tp://www.porn-factory.net/mpegs/a05/helrockysflesh.htmlhttp://www.ha rdasianx.com/perfectorgy/june/01/orgy3.htmhttp://www.hornybabe4u.com /s23a/a8.htmhttp://www.cum-now.net/m/06/02/meini.htmhttp://www.felix adultmovies.com/33set1/abt.htmlhttp://www.thumbzilla.com/main.htmlht tp://www.onlyhotmovies.com/2view/gal009/0531fleshlink.htmlhttp://www .tgpmpegs.com/bigmovies/01-06/smut-1pg.htmlhttp://www.northbeachbeau ty.com/P2016045.jpghttp://www.northbeachbeauty.com/P2026210.jpghttp: //www.northbeachbeauty.com/P2026219.jpghttp://www.northbeachbeauty.c om/791u.htmlhttp://www.myonlygirl.com/e/3e/13.jpghttp://www.myonlygi rl.com/e/3e/14.jpghttp://www.myonlygirl.com/e/3eu2.htmhttp://www.lig htkiss.com/tgp/020602-16/lktz.htmhttp://bignaturals.sexpic-s.com/big 159/zilla.htmhttp://www.a-xxx-pics.com/older-woman-sex/020602m/tzill a.htmlhttp://www.xxxorgy.net/gangbang/orgyparty66/videos-agx.htmlhtt p://www.cumshotsmania.com/lucky/02062002/cum-aol.htmlhttp://www.foxy vids.com/free-pages/02/beaver4.htmhttp://www.felixadultmovies.com/33 set2/abt.htmlhttp://www.lesbolunch.com/02-06-03/lezbianporn58.htmlht tp://www.free-xxx-pics-4u.com/tgp_nice_ass01/zilla.htmhttp://www.fre eezinebucks.com/exit/exit-console4.new.phphttp://freeezinebucks.com/ exit/exit-console3.php?id=fxmovies&niche=analhttp://freeezinebuc ks.com/exit/exit_console_blur2.php?niche=anal&id=fxmovieshttp:// freeezinebucks.com/exit/exit-console2.php?id=fxmovies&niche=anal http://freeezinebucks.com/exit/exit_console.php?id=fxmovies&nich e=analhttp://www.freeanalezine.com/?id=fxmovieshttp://freeezinebucks .com/exit/enter_console.php?niche=anal&id=fxmovieshttp://www.fre eezinebucks.com/ps.php?s=1&u=fxmovieshttp://freexmovies.com/http ://www.thumbzilla.com/cgi-bin/x3top3/outh.cgi?freexmovies.comhttp:// www.xxxmovies4free.org/jun_03/1mova.htmlhttp://www3.blinghosting.com /www/sanaa/buttfinger/Thumbzilla.htmlhttp://www.niceroundass.com/url /060102/justblazec.htmlhttp://www.godsex.com/nadios/anime21.htmlhttp ://www.karups-private-collection.com/40alena84gt/a18.jpghttp://www.k arups-private-collection.com/40alena84gt/zilla.htmlhttp://www.teendr eamer.cc/lesbian/6/index08.htmhttp://www.trog-porn.com/mature/0027/z illa.htmlhttp://www.lowcosthighqualityxxx.com/2002jun/jun3/0026.jpgh ttp://www.lowcosthighqualityxxx.com/2002jun/jun3/69jun3.htmlhttp://www.thumbzilla.com/cen2/index.html>frames...
>
>Thanks!
>
>--sq
>
>
>ScreenTest
>*** 's Power Mac G4/799, Mac OS 9.2.2 ************************************
>G4 altivec, 800 MHz, memory bus 133 MHz, 91.938 Mflop/s
>Psychtoolbox 2.44, 7 August 2000, Matlab 5.2.1.1421
>*** Screen 0 *************************************************************
>ATY,BlueStone_A (.Display_DualHead version 1.0f64) in slot SLOT-1
>10 bit dacs. 1280x1024 60 Hz. (60 Hz avail)
>--------------------------------------------------------------------------
>Now creating identity gamma table for driver with 10-bit DACs.
>Now creating identity gamma table for driver with 10-bit DACs.
>Now creating identity gamma table for driver with 10-bit DACs.
>Now creating identity gamma table for driver with 10-bit DACs.
>Now creating identity gamma table for driver with 10-bit DACs.
>Now creating identity gamma table for driver with 10-bit DACs.
>Now creating identity gamma table for driver with 10-bit DACs.
>pixel size                           8    16    32     bits
>pages                                1     1     1
>CopyWindow (ie CopyBits)          185   188   188     MB/s
>GetClut equals SetClut?             ~=    ~=    na
>SetClut suppresses ints. for        0.0   0.0   0.0   frames
>--------------------------------------------------------------------------
>Supports cscGetClutBehavior cscGetTimingRanges cscGetDetailedTiming
>1 SetClutDriverWaitsForBlanking
>--------------------------------------------------------------------------
>SetClut waits for blanking (i.e. ~60 Hz) if parameters have these values:
>            prioritySequence
>    60 Hz:               0/0
>    60 Hz:               7/0
>SetClut does not wait for blanking if parameters have these values:
>            prioritySequence
>   280 Hz:               7/7
>The following parameters do not affect whether SetClut waits for blanking:
>                 clutEntries         pixelSize
>*** Screen 1 *************************************************************
>ATY,BlueStone_B (.Display_DualHead version 1.0f64) in slot SLOT-1
>10 bit dacs. 1152x870 63 Hz. (56,60,65,67,70,72,75,76,85,90,100,120 Hz avail)
>--------------------------------------------------------------------------
>Now creating identity gamma table for driver with 10-bit DACs.
>Now creating identity gamma table for driver with 10-bit DACs.
>Now creating identity gamma table for driver with 10-bit DACs.
>Now creating identity gamma table for driver with 10-bit DACs.
>Now creating identity gamma table for driver with 10-bit DACs.
>Now creating identity gamma table for driver with 10-bit DACs.
>Now creating identity gamma table for driver with 10-bit DACs.
>pixel size                           8    16    32     bits
>pages                                1     1     1
>CopyWindow (ie CopyBits)          185   187   188     MB/s
>GetClut equals SetClut?             ~=    ~=    na
>SetClut suppresses ints. for       -0.2   0.0   0.0   frames
>--------------------------------------------------------------------------
>Supports cscGetClutBehavior cscGetTimingRanges cscGetDetailedTiming
>1 SetClutDriverWaitsForBlanking
>--------------------------------------------------------------------------
>SetClut waits for blanking (i.e. ~75 Hz) if parameters have these values:
>            prioritySequence
>    75 Hz:               0/0
>    75 Hz:               7/0
>SetClut does not wait for blanking if parameters have these values:
>            prioritySequence
>   288 Hz:               7/7
>The following parameters do not affect whether SetClut waits for blanking:
>                 clutEntries         pixelSize
>**************************************************************************
>
>Best,
>--sq
>--
>==============================================================
>
>Rhea T. Eskew, Jr., Ph.D.   +
>Dept. of Psychology, 125-NI    +   eskew@...
>Northeastern University                +   (617)373-3863 (voice)
>Boston, MA 02115               +   (617)373-8714 (fax)
>                 U.S.A.         +   http://www.psych.neu.edu/vita/eskew.shtml
>
>==============================================================

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Kwick Pick opens locked car doors,
front doors, drawers, briefcases,
padlocks, and more. On sale now!
http://us.click.yahoo.com/ehaLqB/Fg5DAA/Ey.GAA/BEfwlB/TM
---------------------------------------------------------------------~->

http://psychtoolbox.org
POST a message to: psychtoolbox@yahoogroups.com
UNSUBSCRIBE by sending a blank message to:
psychtoolbox-unsubscribe@yahoogroups.com

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/


Attachment converted: Macintosh HD:ClutTest.sit (SIT5/SIT!) (0003C782)


--
==============================================================

Rhea T. Eskew, Jr., Ph.D. + 
Dept. of Psychology, 125-NI    +   eskew@...
Northeastern University                 +   (617)373-3863 (voice)
Boston, MA 02115                +   (617)373-8714 (fax)
                U.S.A.            +   http://www.psych.neu.edu/vita/eskew.shtml

==============================================================
3
>Denis et al --
>
>Ah. "~=" is new to me (as is MatLab).
>
>Yes, the screen wasn't black at the end -- sorry I forgot to mention that.
>
>Here's ClutTest. Note that I am able to manipulate 10 bits on the
>clut in my software using cscSetEntries -- USUALLY. But , as I
>mentioned in a previous message (#973), I have to wave my hands
>around and chant words I don't understand -- and sometimes it fails
>to work, under unpredictable circumstances.
>
>Thanks for any ideas you have.
>
>--sq

dear rhea & doug

the results you both sent were helpful (and identical). it appears
that things are basically working. i've added code to ClutTest
(enclosed) to load a simpler gamma table. i'm hoping this test will
pass. that would resolve the failure of the GetClut vs. SetClut test.

none of these tests are indicating any flakiness with cscSetEntries.

could one of you try the enclosed new version of ClutTest.m and
report the results on the forum?

best

denis

p.s.
by the way, rhea, do you know that you must use cscSetEntries when
the pixelSize is 8 bits or smaller and must instead call
cscDirectSetEntries when the pixelSize is 16 or 32? That might
account for the comments you make above.
4
dear doug & rhea

thanks. we're close. here's another iteration of ClutTest.

I'm also enclosing an enhanced version of SCREEN.mex. It enables the
priority dipping trick for both ATI Rage128 and DualHead. So you can
run at high priority. (It also works around the NULL gamma pointer
bug, for both Radius and ATI DualHead.)

could one of you run ClutTest?

best

denis


>Dear Denis,
>
>Here is the new cluttest results.
>I also asked one of the ATI 3D
>guys about this(NULL Gamma loading).
>When I hear back I will post to the Forum.
>Take care,
>
>Doug Taylor
>
>Denis Pelli wrote:
>
>> >Denis et al --
>> >
>> >Ah. "~=" is new to me (as is MatLab).
>> >
>> >Yes, the screen wasn't black at the end -- sorry I forgot to mention that.
>> >
>> >Here's ClutTest. Note that I am able to manipulate 10 bits on the
>> >clut in my software using cscSetEntries -- USUALLY. But , as I
>> >mentioned in a previous message (#973), I have to wave my hands
>> >around and chant words I don't understand -- and sometimes it fails
>> >to work, under unpredictable circumstances.
>> >
>> >Thanks for any ideas you have.
>> >
>> >--sq
>>
>> dear rhea & doug
>>
>> the results you both sent were helpful (and identical). it appears
>> that things are basically working. i've added code to ClutTest
>> (enclosed) to load a simpler gamma table. i'm hoping this test will
>> pass. that would resolve the failure of the GetClut vs. SetClut test.
>>
>> none of these tests are indicating any flakiness with cscSetEntries.
>>
>> could one of you try the enclosed new version of ClutTest.m and
> > report the results on the forum?
>>
>> best
>>
>> denis
>>
>> p.s.
>> by the way, rhea, do you know that you must use cscSetEntries when
>> the pixelSize is 8 bits or smaller and must instead call
>> cscDirectSetEntries when the pixelSize is 16 or 32? That might
>> account for the comments you make above.
>>
>> ------------------------------------------------------------------------
>> Part 1.2Type: Macintosh File
>
>Content-Type: text/plain; x-mac-type="54455854"; x-mac-creator="74747874";
> name="cluttest8500.2"
>Content-Description: Unknown Document
>Content-Disposition: inline;
> filename="cluttest8500.2"
>X-MIME-Autoconverted: from 8bit to quoted-printable by mx1.nyu.edu
>id g54KVRI17539
>
>ClutTest does a write-then-read test of the Color Lookup Table (CLUT) of the
>video device associated with each of your screens. This tests your CLUT
>hardware, and is usually independent of the current pixel size.
>*** Screen 0 *************************************************************
>ATY,R200i_A (.Display_DualHead version 1.0f69) in slot SLOT-A
>10 bit dacs. 1024x768 60 Hz. (56,60,65,67,70,72,75,76,85,90,100,120 Hz avail)
>Now creating identity 10-bit gamma table for (.Display_DualHead
>version 1.0f69) driver.
>Now creating identity 10-bit gamma table for (.Display_DualHead
>version 1.0f69) driver.
>GetClut is not supported by this video driver at pixelSize 32. Changing to 8.
>Loading 10-bit identity gamma table.
>1. Write top 8 bits to CLUT and read it back. startEntry=0.
> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
> 1 entry. SetClut:[ 1 1 1] vs. GetClut:[ 1 1 1]
>126 entry. SetClut:[ 126 126 126] vs. GetClut:[ 126 126 126]
>254 entry. SetClut:[ 254 254 254] vs. GetClut:[ 253 253 253]
>255 entry. SetClut:[ 255 255 255] vs. GetClut:[ 254 254 254]
>2. Write top 8 bits to CLUT and read it back. startEntry=127.
> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
> 1 entry. SetClut:[ 1 1 1] vs. GetClut:[ 1 1 1]
>126 entry. SetClut:[ 126 126 126] vs. GetClut:[ 126 126 126]
>254 entry. SetClut:[ 254 254 254] vs. GetClut:[ 253 253 253]
>255 entry. SetClut:[ 255 255 255] vs. GetClut:[ 254 254 254]
>3. Write top 10 bits to CLUT and read it back. startEntry=0.
> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
> 1 entry. SetClut:[ 4 4 4] vs. GetClut:[ 4 4 4]
>255 entry. SetClut:[ 1023 1023 1023] vs. GetClut:[ 1020 1020 1020]
>4. Write top 10 bits to CLUT and read it back. startEntry=127.
> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
> 1 entry. SetClut:[ 4 4 4] vs. GetClut:[ 4 4 4]
>255 entry. SetClut:[ 1023 1023 1023] vs. GetClut:[ 1020 1020 1020]
>*** Screen 1 *************************************************************
>ATY,R200i_B (.Display_DualHead version 1.0f69) in slot SLOT-A
>10 bit dacs. 1024x768 85 Hz. (56,60,65,67,70,72,75,76,85,90,100,120 Hz avail)
>Now creating identity 10-bit gamma table for (.Display_DualHead
>version 1.0f69) driver.
>Now creating identity 10-bit gamma table for (.Display_DualHead
>version 1.0f69) driver.
>GetClut is not supported by this video driver at pixelSize 32. Changing to 8.
>Loading 10-bit identity gamma table.
>1. Write top 8 bits to CLUT and read it back. startEntry=0.
> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
> 1 entry. SetClut:[ 1 1 1] vs. GetClut:[ 1 1 1]
>126 entry. SetClut:[ 126 126 126] vs. GetClut:[ 126 126 126]
>254 entry. SetClut:[ 254 254 254] vs. GetClut:[ 253 253 253]
>255 entry. SetClut:[ 255 255 255] vs. GetClut:[ 254 254 254]
>2. Write top 8 bits to CLUT and read it back. startEntry=127.
> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
> 1 entry. SetClut:[ 1 1 1] vs. GetClut:[ 1 1 1]
>126 entry. SetClut:[ 126 126 126] vs. GetClut:[ 126 126 126]
>254 entry. SetClut:[ 254 254 254] vs. GetClut:[ 253 253 253]
>255 entry. SetClut:[ 255 255 255] vs. GetClut:[ 254 254 254]
>3. Write top 10 bits to CLUT and read it back. startEntry=0.
> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
> 1 entry. SetClut:[ 4 4 4] vs. GetClut:[ 4 4 4]
>255 entry. SetClut:[ 1023 1023 1023] vs. GetClut:[ 1020 1020 1020]
>4. Write top 10 bits to CLUT and read it back. startEntry=127.
> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
> 1 entry. SetClut:[ 4 4 4] vs. GetClut:[ 4 4 4]
>255 entry. SetClut:[ 1023 1023 1023] vs. GetClut:[ 1020 1020 1020]
>**************************************************************************
>p.s. We don't have a video card with more-than-8 bit DACs to test
>this program on.
>Apparently you do. Did ClutTest work? Please let us know.
>psychtoolbox@yahoogroups.com
>»
5
Hi Denis, Doug, et al.:

Here's the latest cluttest. I see that it reports 75hz, which should
be right, but that tests 1 and 2 (on Screen 1) fail . Does this mean
the driver is altering the values in the clut (when you attempt to
write just 8 bits?).

That MIGHT be compatible with the problems I have been having
programming this thing. I observed different symptoms with different
color directions (yes, I know its weird), so for example I might have
something working as long as r=g=b but it would fail when I didn't
modulate one gun. So I wonder if its worth making a version of your
test where some tests don't always have r=g=b (I sometimes use random
rgbs for testing purposes).

However, my problems occurred when I wrote the top 10 bits , and I
gather your test succeeds there. And right now I think I have things
working to my satisfaction...I hope.

Thanks very much for all your work on this!

--sq


»cluttest
ClutTest does a write-then-read test of the Color Lookup Table (CLUT) of the
video device associated with each of your screens. This tests your CLUT
hardware, and is usually independent of the current pixel size.
*** Screen 0 *************************************************************
ATY,BlueStone_A (.Display_DualHead version 1.0f64) in slot SLOT-1
10 bit dacs. 1280x1024 60 Hz. (60 Hz avail)
GetClut is not supported by this video driver at pixelSize 32. Changing to
8.
Loading 10-bit identity gamma table.
1. Write top 8 bits to CLUT and read it back. startEntry=0.
Failed! Values read differed (by up to 1) from what was written.
0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
1 entry. SetClut:[ 1 1 1] vs. GetClut:[ 1 1 1]
126 entry. SetClut:[ 126 126 126] vs. GetClut:[ 126 126 126]
254 entry. SetClut:[ 254 254 254] vs. GetClut:[ 253 253 253]
255 entry. SetClut:[ 255 255 255] vs. GetClut:[ 254 254 254]
2. Write top 8 bits to CLUT and read it back. startEntry=127.
Failed! Values read differed (by up to 1) from what was written.
0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
1 entry. SetClut:[ 1 1 1] vs. GetClut:[ 1 1 1]
126 entry. SetClut:[ 126 126 126] vs. GetClut:[ 126 126 126]
254 entry. SetClut:[ 254 254 254] vs. GetClut:[ 253 253 253]
255 entry. SetClut:[ 255 255 255] vs. GetClut:[ 254 254 254]
3. Write top 10 bits to CLUT and read it back. startEntry=0.
Success!
4. Write top 10 bits to CLUT and read it back. startEntry=127.
Success!
*** Screen 1 *************************************************************
ATY,BlueStone_B (.Display_DualHead version 1.0f64) in slot SLOT-1
10 bit dacs. 1152x870 75 Hz. (56,60,65,67,70,72,75,76,85,90,100,120 Hz
avail)
Loading 10-bit identity gamma table.
1. Write top 8 bits to CLUT and read it back. startEntry=0.
Failed! Values read differed (by up to 1) from what was written.
0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
1 entry. SetClut:[ 1 1 1] vs. GetClut:[ 1 1 1]
126 entry. SetClut:[ 126 126 126] vs. GetClut:[ 126 126 126]
254 entry. SetClut:[ 254 254 254] vs. GetClut:[ 253 253 253]
255 entry. SetClut:[ 255 255 255] vs. GetClut:[ 254 254 254]
2. Write top 8 bits to CLUT and read it back. startEntry=127.
Failed! Values read differed (by up to 1) from what was written.
0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
1 entry. SetClut:[ 1 1 1] vs. GetClut:[ 1 1 1]
126 entry. SetClut:[ 126 126 126] vs. GetClut:[ 126 126 126]
254 entry. SetClut:[ 254 254 254] vs. GetClut:[ 253 253 253]
255 entry. SetClut:[ 255 255 255] vs. GetClut:[ 254 254 254]
3. Write top 10 bits to CLUT and read it back. startEntry=0.
Success!
4. Write top 10 bits to CLUT and read it back. startEntry=127.
Success!
**************************************************************************
p.s. We don't have a video card with more-than-8 bit DACs to test this
program on.
Apparently you do. Did ClutTest work? Please let us know.
psychtoolbox@yahoogroups.com


>dear doug & rhea
>
>thanks. we're close. here's another iteration of ClutTest.
>
>I'm also enclosing an enhanced version of SCREEN.mex. It enables the
>priority dipping trick for both ATI Rage128 and DualHead. So you can
>run at high priority. (It also works around the NULL gamma pointer
>bug, for both Radius and ATI DualHead.)
>
>could one of you run ClutTest?
>
>best
>
>denis
>
>
>>Dear Denis,
>>
>>Here is the new cluttest results.
>>I also asked one of the ATI 3D
>>guys about this(NULL Gamma loading).
>>When I hear back I will post to the Forum.
>>Take care,
>>
>>Doug Taylor
>>
>>Denis Pelli wrote:
>>
>>> >Denis et al --
>>> >
>>> >Ah. "~=" is new to me (as is MatLab).
>>> >
>>> >Yes, the screen wasn't black at the end -- sorry I forgot to
>>>mention that.
>>> >
>>> >Here's ClutTest. Note that I am able to manipulate 10 bits on the
>>> >clut in my software using cscSetEntries -- USUALLY. But , as I
>>> >mentioned in a previous message (#973), I have to wave my hands
>>> >around and chant words I don't understand -- and sometimes it fails
>>> >to work, under unpredictable circumstances.
>>> >
>>> >Thanks for any ideas you have.
>>> >
>>> >--sq
>>>
>>> dear rhea & doug
>>>
>>> the results you both sent were helpful (and identical). it appears
>>> that things are basically working. i've added code to ClutTest
>>> (enclosed) to load a simpler gamma table. i'm hoping this test will
>>> pass. that would resolve the failure of the GetClut vs. SetClut test.
>>>
>>> none of these tests are indicating any flakiness with cscSetEntries.
>>>
>>> could one of you try the enclosed new version of ClutTest.m and
>> > report the results on the forum?
>>>
>>> best
>>>
>>> denis
>>>
>>> p.s.
>>> by the way, rhea, do you know that you must use cscSetEntries when
>>> the pixelSize is 8 bits or smaller and must instead call
>>> cscDirectSetEntries when the pixelSize is 16 or 32? That might
>>> account for the comments you make above.
>>>
>>> ------------------------------------------------------------------------
>>> Part 1.2Type: Macintosh File
>>
>>Content-Type: text/plain; x-mac-type="54455854"; x-mac-creator="74747874";
>> name="cluttest8500.2"
>>Content-Description: Unknown Document
> >Content-Disposition: inline;
> > filename="cluttest8500.2"
> >X-MIME-Autoconverted: from 8bit to quoted-printable by mx1.nyu.edu
> >id g54KVRI17539
> >
> >ClutTest does a write-then-read test of the Color Lookup Table (CLUT) of the
>>video device associated with each of your screens. This tests your CLUT
>>hardware, and is usually independent of the current pixel size.
>>*** Screen 0 *************************************************************
>>ATY,R200i_A (.Display_DualHead version 1.0f69) in slot SLOT-A
>>10 bit dacs. 1024x768 60 Hz. (56,60,65,67,70,72,75,76,85,90,100,120 Hz avail)
>>Now creating identity 10-bit gamma table for (.Display_DualHead
>>version 1.0f69) driver.
>>Now creating identity 10-bit gamma table for (.Display_DualHead
>>version 1.0f69) driver.
>>GetClut is not supported by this video driver at pixelSize 32. Changing to 8.
>>Loading 10-bit identity gamma table.
>>1. Write top 8 bits to CLUT and read it back. startEntry=0.
>> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
>> 1 entry. SetClut:[ 1 1 1] vs. GetClut:[ 1 1 1]
>>126 entry. SetClut:[ 126 126 126] vs. GetClut:[ 126 126 126]
>>254 entry. SetClut:[ 254 254 254] vs. GetClut:[ 253 253 253]
>>255 entry. SetClut:[ 255 255 255] vs. GetClut:[ 254 254 254]
>>2. Write top 8 bits to CLUT and read it back. startEntry=127.
>> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
>> 1 entry. SetClut:[ 1 1 1] vs. GetClut:[ 1 1 1]
>>126 entry. SetClut:[ 126 126 126] vs. GetClut:[ 126 126 126]
>>254 entry. SetClut:[ 254 254 254] vs. GetClut:[ 253 253 253]
>>255 entry. SetClut:[ 255 255 255] vs. GetClut:[ 254 254 254]
> >3. Write top 10 bits to CLUT and read it back. startEntry=0.
>> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
>> 1 entry. SetClut:[ 4 4 4] vs. GetClut:[ 4 4 4]
>>255 entry. SetClut:[ 1023 1023 1023] vs. GetClut:[ 1020 1020 1020]
>>4. Write top 10 bits to CLUT and read it back. startEntry=127.
>> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
>> 1 entry. SetClut:[ 4 4 4] vs. GetClut:[ 4 4 4]
>>255 entry. SetClut:[ 1023 1023 1023] vs. GetClut:[ 1020 1020 1020]
>>*** Screen 1 *************************************************************
>>ATY,R200i_B (.Display_DualHead version 1.0f69) in slot SLOT-A
>>10 bit dacs. 1024x768 85 Hz. (56,60,65,67,70,72,75,76,85,90,100,120 Hz avail)
>>Now creating identity 10-bit gamma table for (.Display_DualHead
>>version 1.0f69) driver.
>>Now creating identity 10-bit gamma table for (.Display_DualHead
>>version 1.0f69) driver.
>>GetClut is not supported by this video driver at pixelSize 32. Changing to 8.
>>Loading 10-bit identity gamma table.
>>1. Write top 8 bits to CLUT and read it back. startEntry=0.
>> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
>> 1 entry. SetClut:[ 1 1 1] vs. GetClut:[ 1 1 1]
>>126 entry. SetClut:[ 126 126 126] vs. GetClut:[ 126 126 126]
>>254 entry. SetClut:[ 254 254 254] vs. GetClut:[ 253 253 253]
>>255 entry. SetClut:[ 255 255 255] vs. GetClut:[ 254 254 254]
>>2. Write top 8 bits to CLUT and read it back. startEntry=127.
>> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
>> 1 entry. SetClut:[ 1 1 1] vs. GetClut:[ 1 1 1]
>>126 entry. SetClut:[ 126 126 126] vs. GetClut:[ 126 126 126]
>>254 entry. SetClut:[ 254 254 254] vs. GetClut:[ 253 253 253]
>>255 entry. SetClut:[ 255 255 255] vs. GetClut:[ 254 254 254]
>>3. Write top 10 bits to CLUT and read it back. startEntry=0.
>> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
>> 1 entry. SetClut:[ 4 4 4] vs. GetClut:[ 4 4 4]
>>255 entry. SetClut:[ 1023 1023 1023] vs. GetClut:[ 1020 1020 1020]
>>4. Write top 10 bits to CLUT and read it back. startEntry=127.
>> 0 entry. SetClut:[ 0 0 0] vs. GetClut:[ 0 0 0]
>> 1 entry. SetClut:[ 4 4 4] vs. GetClut:[ 4 4 4]
>>255 entry. SetClut:[ 1023 1023 1023] vs. GetClut:[ 1020 1020 1020]
>>**************************************************************************
>>p.s. We don't have a video card with more-than-8 bit DACs to test
>>this program on.
>>Apparently you do. Did ClutTest work? Please let us know.
>>psychtoolbox@yahoogroups.com
>>»
>
>
>http://psychtoolbox.org
>POST a message to: psychtoolbox@yahoogroups.com
>UNSUBSCRIBE by sending a blank message to:
>psychtoolbox-unsubscribe@yahoogroups.com
>
>Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>
>Attachment converted: Macintosh HD:ClutTest.sit 4 (SIT5/SIT!) (0003C88D)


--
==============================================================

Rhea T. Eskew, Jr., Ph.D. +
Dept. of Psychology, 125-NI + eskew@...
Northeastern University + (617)373-3863 (voice)
Boston, MA 02115 + (617)373-8714 (fax)
U.S.A. + http://www.psych.neu.edu/vita/eskew.shtml

==============================================================