Exupery

exupery@lists.squeakfoundation.org

107 discussions

Preparing for 0.15 and planning 0.16
by bryce＠kampjes.demon.co.uk 01 Mar '09

01 Mar '09

I'm doing the final preparation for the 0.15 release. There's a Linux VM and a Windows VM built so far. The Windows VM lacks freetype but should otherwise be OK, a proper Windows VM with Freetype would be nice especially as these VMs have Eliot's closure bytecodes in them.. I'll build an image in the next few days. The code on SqueakSource should be very close to the actual release code. There's no Mac VM. It would be nice if someone built one. I may add full method inlining in the next release, or at least an initial implementation. Inlining is a fairly simple feature except for blocks and de-optimisation. The problem with blocks is they can be accessed from external contexts so need to exist and can be returned to which complicates inlining them. The problem with de-optimisation is almost another blocks problem as de-optimisation needs to be thought about a little more carefully when de-optimising a method context may impact the execution of the block. The solution is probably to create contexts for both blocks and methods that create blocks. It'll make the inlined block situation a little slower but will make it possible to de-optimise one context at a time without risking crashes. Block contexts can always be removed if they're inlined from creation to execution, i.e. for the entire loop. The exact block inlining design doesn't need to be decided yet, it can wait until after the simple case of inlining a method into another method is implemented and debugged. Starting on full method inlining is a departure from plan. 0.15 looks solid enough to enable easy debugging of blocks. The original plan was to continue tuning for a 1.0 before adding full method inlining. the advantage of adding full method inlining now is it's a key part of the original design and will provide a large performance impact both in performance gains and also in changing relative costs. Bryce

1 0

Another bug fixed.
by bryce＠kampjes.demon.co.uk 22 Feb '09

22 Feb '09

It look's like I've finally fixed the bug I've been chasing for the last few months. A fix will be published to SqueakSource in the next few days then final release testing can begin. It'll be interesting to see how reliable Exupery is now. It's possible that this bug has been the main reason for crashes for the last few releases. It's an old bug that's most often triggered by the first method compiled which is Delay>>wait, it's just the conditions required to trigger it are rare. The bug was Exupery was marking the home context of a block as a root but not the block itself. So if the block was in old space then memory corruption could be caused by an incremental GC if the block became a root. Live contexts in Squeak are always mapped as roots to avoid needing to use a write barrier for every context write which includes every stack push. Even if this isn't the last major reliability bug, my VM debugging skills have improved so later bugs should take less effort to fix. There's a few useful debugging functions in the VM that can be exploited or modified. allAccessibleObjectsOkay is much slower than validateRoots even though both do a full memory scan. I'll probably merge in VMMaker-dtl.115 which has the latest version of Eliot's closure bytecodes. The closure bytecodes are already in the Exupery VM. Bryce

1 0

Obama -- The Judas Goat
by Lawrence Auster 25 Jan '09

25 Jan '09

Obama -- The Judas Goat 1/25/2009 By David Duke Judas Goat\ A goat that leads other goats or sheep to slaughter. Also, one who entices into danger and betrays others. The name is an allusion to Judas Iscariot, who betrayed Jesus for 30 pieces of slaughter. (From Merriam Webster�s Dictionary) Barack Hussein Obama is a Judas Goat. Extremist Jews guided Barack Hussein Obama�s career from day one, even all the way back to Harvard Law School. Radical Zionist hitman, David Axelrod previously orchestrated the Jewish-financed and organized defeat of perceived anti-Zionist Sen. Charles Percy. He is the man who ran Obama�s campaign for President and who is his chief handler. Obama�s campaign was overwhelmingly financed by the most powerful Zionist bankers in the world. His campaign�s largest contribution source was the Zionist international banking firm of Goldman Sachs. (FEC campaign records). In both Obama�s Senate and Presidential campaign he prostrated himself before AIPAC (American Israel Public Affairs Committee) promising even more money and blood for Israel�s terrorism than even the supine John McCain, and even more money and blood than the previous Shabbez Goy in the White House, George Bush. Before the Israeli terrorism and mass murder in Gaza, he went to Israel and said that he supported Israel�s planned murderous terrorism against the men, women and children of Gaza. His first act as President-elect was to appoint a rabid Zionist, Israeli dual citizen who served in the Israeli Army as his Chief of Staff, Rahm Emmanuel. As thousands of women and children in Gaza were killed or maimed, Obama remained silent. Within a few days as President Obama supported indiscriminate American missile strikes in villages of our ally Pakistan, a clear continuation of Bush�s policies. He completely supports the theft of trillions of American taxpayer dollars to the Zionist international bankers. Eighty percent of American Jews voted for Obama, and all the main leaders of the Jewish Supremacist state of Israel have proclaimed Obama as the perfect man for U.S. President. What better for the Zionists to have their own servant perceived as a clean break from the Jewish extremist-controlled Bush administration? What better than for the Zionists than to have their Shabbez Goy be treated by the world�s press and even by much of the anti-Zionist community as a man of �real change?� What better for the Zionists than for the world to think that Obama will be a change from the Zionist-controlled policies when he willing to do anything that Israel demands? What better for the Zionist murderers to have their craven puppet be looked at by the whole world as a man of honor and integrity and fairness. In the slaughterhouse, the Judas Goat is often painted with bright colors, adorned with strong, sweet scents to lure the sheep to their pens and to their death. Obama, the Judas Goat of our time, is looked upon by millions of Zionist-propagandized sheep as the man who will lead them to salvation. Instead, he leads America, Palestine and the world to the bloody altar of Jewish Supremacism. Any supposed anti-Zionist who praises Barack Obama is actually aiding this Judas Goat to lead us all to slaughter. Every person who truly opposes Jewish extremism must speak out and expose the Judas Goat named Barack Obama! -- Dr. David Duke Former Member of the House of Representatives State of Louisiana United States of America It is up to you the people of the United States who can still think freely, and up to all of the people of the world who are able to see through the deceptions of the Zionist-influenced Global media -- to get this simple, powerful message to everyone on earth. Email this message to your friends, post it on forums and websites and put links to it on every website, facebook or other media in the world. Go and seek out media and Internet sites not controlled by Zionist power. Make youtube videos of this message, (use the short audio and find good illustrative pictures) and post it untill your fingers are worn and tired, print it and mail it to newspapers or any media outlet that has still not fallen under the propaganda of the Zionists. Let the world know the truth. Person by person, in the USA, Canada, UK, France, Germany, Italy, japan, Russia, China, Africa, Palestine, South America and across the whole world this Judas Goat must be exposed for what he is, so that when begins to do his evil for Israel, the whole world will know exactly what is going on and resist! Source : http://www.davidduke.com/general/obama-the-judas-goat_7317.html ------------------------ Why is the President of Israel, the terrorist who just oversaw the Zionist mass murder and maiming of thousands of Palestinians so happy that Obama is President of the USA? by David Duke Read the excerpt from the Israeli News about how President Perez and Israel think that Obama�s becoming U.S. President is great day for Israel. "Israel�s President Shimon Peres ecstatic over the election of Obama" Ronen Medzini Israel News Jan. 21 �Today is a great day not only for the United States of America, but for the entire world,� President Shimon Peres wrote in a letter addressed to Barack Obama on the day of his inauguration as president of the United States. �Obama was elected by the United States, but as a matter of fact, he was chosen by the whole of humankind,� Why is Peres so ecstatic? Why shouldn�t he be, he knows that Obama is completely in the grip of the extremist Jewish Zionists in America, and he knows that the greater Obama�s popularity and idol worship, the more Obama can do for the International Zionist Cause. Any thinking and caring human being who realizes that the Zionist-controlled American foreign policy has been a disaster for the robbed and murdered people of Palestine, Lebanon, Iraq and a catastrophe for the 50,000 American wounded in Iraq and Afghanistan, as well as an economic catastrophe for the hardworking Americans who pay trillions to finance these wars for Israel -- must wake up the fact that supporting Obama and increasing his popularity will only aid Zionist terrorism, war, and their murder and oppression of the Palestinians. It will also hasten the economic suffering of billions of people around the world as his popularity enables him to more easily aid the Zionist International Bankers steal the wealth of the United States, Europe and the world. Obama is totally in the bloodstained and green ink- stained hands of the Zionists. The hard truth is that the more good will and support Obama has also gives more power to support the Zionist agenda! Mark my words. The Obama Presidency will be disaster for America and for the world. Obama was put into office by the Zionists. His top two cohorts for years have been the radical Jews David Axelrod and Rahm Emmanuel. Both have long records of radical Zionism and have been attack dogs against anyone perceived as having the slightest opposition to Israel. One such victim was Sen. Charles Percy, who both men worked to defeat and destroy because he dared to only be 99 percent rather than 100 percent pro-Israel. Rahm Emmanuel, a dual citizen of Israel who went to fight for Israel, he has a long pedigree of Jewish extremism. His father served in the Irgun Terrorist Gang and he himself is named after an Irgun terrorist. Zionist leaders in Chicago actually call Obama �the first Jewish President� and boast that Jews were key players in Obama�s every step up the ladder to President. from the very earliest days, extremist Jews were the largest contributors to his campaign. In the beginning of his Presidential bid, three Hollywood Jews that constantly make movies about Jewish suffering, but never about the Zionist terrorism and theft against the Palestinian people, Steven Spielberg, David Geffen, and Jeffrey Katzenberg raised 1.2 million for Obama in a single Hollywood party. By the time Obama�s campaign was in full swing, he had huge support from the criminal Zionist International Banking firms such as Goldman Sachs and Lehman brothers. Goldman Sachs was Obama�s biggest single contributor, and his vast war chest came not from American manufacturing firms like GM or even American oil companies, (not one was in his top twenty) it was overwhelmingly dominated by Zionist international bankers, the same ones whose thievery and fraud are giving the world this economic depression. For those looking for meaningful social and political change, do you really think it will come from this man who has already been bought heart, head and soul by the most powerful czars of the international financial establishment and the biggest globalists in the world? I know that many are desperate for change, so desperate that you want to believe anything. But in the face of these facts can�t you see that Obama will be even more dangerous to freedom and justice than even George Bush and his band of Neocons were. What better way to wipe out George Bush�s hated legacy and make the world believe that America has really changed than with the election of Obama. But, all the real Zionist power, Zionist media power, and Zionist financial power in America is still in place, even stronger than ever. Many Americans and others around the world who want to do good are now telling us how wonderful Obama will be as president. What a great change it will be from the old policies. This is because of the Zionist-Controlled media hype, promoting Obama. The fact is that these poor sods are ignorantly helping the radical Zionist agenda in Israel and around the world. Every day that you don�t help expose Obama for the Zionist servant that he actually is, his popularity will be a greater danger to peace and freedom. If the Zionist terrorist Shimon Perez is happy about the coronation of Obama, then why in the hell should you be? �David Duke Source : http://www.davidduke.com/general/7303_7303.html ------------------------------------- You or someone using your email adress is currently subscribed to the Lawrence Auster Newletter. If you wish to unsubscribe from our mailing list, please let us know by calling to 1 212 865 1284 Thanks, Lawrence Auster, 238 W 101 St Apt. 3B New York, NY 10025 Contact: lawrence.auster(a)att.net -------------------------------------

1 0

Debugging Exupery
by bryce＠kampjes.demon.co.uk 10 Nov '08

10 Nov '08

I've just found one of the old bugs that's causing crashes every hour or so. I'm not sure if it's the causes of all the crashes I can reproduce with the current slow reproduction or just one of the crashes. Here's a quick walk through debugging real crashes in Exupery. First try running the background compiler to see if it crashes: Exupery initialiseExupery. Exupery startBackgroundCompilation. Now just use Squeak as normal until it crashes. It takes about ten minutes on my machine for it to finish compiling itself. Load the logs using "explore it": ExuperyLogReader new readFile: 'Exupery.log'. Exupery is a dynamic compiler, it chooses what to compile based on profiling, so it can be very hard to reproduce bugs as there's some luck in what's compiled. Then replay the logs to reproduce the code cache as it was when it crashed. Execute the following in the explorer on the logs: self compile Then try to make the crash reproducible by running the last thing many times over because it looks like a race. 10000 timesRepeat: [Exupery compileMethod: #visitMov: class: InstructionSelector] Now reduce the amount of code in the code cache to make the bug more understandable and normally faster to reproduce. The following expressions are useful on the logs: 3 to: 4 do: [:each| (self at: each) shouldCompile: false] 51 to: 100 do: [:each| (self at: each) shouldCompile: false] 3 to: 4 do: [:each| (self at: each) shouldCompile: true] (self at: 3) shouldCompile: false (self at: 3) shouldCompile: true Then refine the reproduction to something easier to debug and hopefully that can be turned into an SUnit test: Exupery initialiseExupery. Exupery compileMethod: #wait class: Delay 10000 timesRepeat: [Exupery log: 'Compiling example'. Exupery compileMethod: (Exupery definingMethodForSelector: #visitMov: class: InstructionSelector) inlining: #() receiver: InstructionSelector] Then debug it using logging and gdb. There's only been a few Exupery bugs that were debug-gable with the Smalltalk debugger, bugs normally crash the image. In this case it's curious that the bug is triggered by compiling Delay>>wait, the first method that the background compiler normally compiles as often most of the time is spent waiting in the idle loop. One bug is Exupery isn't checking if the context it's returning into has already returned, it's PC is nil, and is crashing in Exupery's return code when it fails to calculate the address to place it's result in as the stack pointer is nil when it's expecting a very small integer. I'm not sure if that's the only bug that the reproduction above is producing, only in one case was the crash at a time when it was definitely due to only trying to return into a dead context. Most crashes are however in the return code, and the context being returned to isn't good. The next step is to produce an SUnit test for returning from a compiled method into a context with a nil PC and stack pointer. Then after fixing it, checking if that fixes the slow bug reproduction as well. Bryce

2 2

native CompiledMethod format
by Gerardo Richarte 29 Oct '08

29 Oct '08

Hi, for some time I've been wondering how the native code for CompiledMethods is arranged, and I have a few specific questions. Any pointers to info will be very much welcome Are the native opcodes maintained in Smalltalk objects or in external memory? and then: Generally, how are CompiledMethods fixed after a Garbage Collection? for example, how are references to other objects handled? are calls to other methods somehow inlined (I mean, direct calls, not through lookups), and if so, how are this inlined direct calls solved when native code moves around (if it's stored in Object Memory) I know that by reading the code I can learn all this, but my need for info is not so serious :) thanks a lot! richie

2 1

( picoVerse-:( How do I get Exupery? ) )
by Kjell Godo 28 Oct '08

28 Oct '08

I tried http://mfrouter.dyndns.org/Exupery_VM.7z but nothing happened. I am making an assembler in Dolphin Smalltalk. It is trying to do all the addressing modes. It is part of the picoLARC project on SourceForge.net but the assembler stuff is not uploaded yet. It uses a Lispy syntax that looks like ( add ( into:From: EAX ( at ( + EBP 4 ) ) ) ) or ( add ( into:From: EAX ( at ( scale:index:base: 4 EBX EBP ) ) ) ) or ( push ( at EBP ) ) Doing it this way seems to minimize the compiling complexity. Plus it can participate in the picoLARC OOP macro system. Your Exupery way is more like regular assembler. It's brief and that's nice. But you also do long function names for machine language instructions which is good. All the addressing modes are mostly finished coding but untested. I would like to use Exupery to test against my picoLARC assembler. If possible.

3 11

That wasn't the bug
by bryce＠kampjes.demon.co.uk 09 Oct '08

09 Oct '08

Unfortunately the VM always marks the contexts before exuperyReturn is called, so the bug wasn't what I thought it was. Pity, that means I don't know what the remaining bugs are. Time for more testing. Bryce

1 0

[ANN] Exupery 0.14 is released
by bryce＠kampjes.demon.co.uk 06 Oct '08

06 Oct '08

The major improvement is to the speed of register allocation. I've fixed a couple of major performance problems so register allocation takes 50% of compilation time even for the largest methods. Register allocation now appears to take roughly linear time. There's still plenty of room to improve compilation time. Every stage copies the entire intermediate graph to produce the input to the next stage which is redundant, most stages only change a few places. The register allocator's liveness analyser still uses Sets to represent which variables are live rather than bit vectors. This release can compile cascades, the last missing core language feature. Exupery still can only compile a handful of the core primitives, it only compiles #at: for pointer objects. Cascades were added because they were used more in 3.10, the choice was either delete 2 system tests or add cascades. I delayed the release to add cascades. There's a few bug fixes of old bugs but this release is not noticeably more reliable than the previous release. It should be a bit more reliable though, especially when running the new Exupery VM. The new Exupery VM just has a single bug fix in it. Here's the benchmarks: arithmaticLoopBenchmark 414 compiled 94 ratio: 4.404 bytecodeBenchmark 726 compiled 264 ratio: 2.750 sendBenchmark 707 compiled 454 ratio: 1.557 doLoopsBenchmark 388 compiled 398 ratio: 0.975 pointCreation 433 compiled 423 ratio: 1.024 largeExplorers 257 compiled 258 ratio: 0.996 compilerBenchmark 248 compiled 249 ratio: 0.996 Cumulative Time 419 compiled 275 ratio 1.519 ExuperyBenchmarks>>arithmeticLoop 105ms SmallInteger>>benchmark 362ms InstructionStream>>interpretExtension:in:for: 6051ms Average 612.691 The key benchmark for this release is the last one, compiling interpretExtension:in:for: which now only takes 6 seconds. With previous version of the compiler it used to take over 2 minutes to compile. The change in times in the other benchmarks are mostly due to me upgrading from an Athlon 64 2.2GHz to a Core 2 3.0GHz. The register allocator should be slightly more efficient especially when compiling send heavy code. Reliability and compile time performance has dominated the last few releases. Now run time performance and reliability are the biggest issues. Exupery still can crash after about an hour's active use depending on what's being done. Bryce P.S. I don't think it'll be that useful to have VMs for all platforms. The old VM 0.12 should be fine for messing around. The new VM would be useful if anyone felt like trying to capture bugs though. If anyone want's to rebuild the VM for non-linux platforms, I'll delay announcing the release for a few days. I doubt it's worth the effort of re-creating a VM build environment though. I suspect that the major remaining bug has to do with de-compilation of contexts and may require VM changes to fix.

2 9

Exupery method cache
by Markus Fritsche 05 Oct '08

05 Oct '08

Hi, I just fired up an image where I compiled a few methods. When I start the image again, I can't compile because Exupery complains, that the plugin is missing. Is this how it's supposed to be or is there a bug in my vm-compilation/ image-compilation?

2 3

An update on progress
by bryce＠kampjes.demon.co.uk 05 Oct '08

05 Oct '08

The three things that have been added since the last release are: * More x86 addressing modes * Generic primitives (calling out to the interpreter's primitives) * Eliot's closure bytecodes and Freetype support merged into the VM More x86 addressing modes and generic primitives have helped out with performance. The addressing modes help out in send/return code which accesses VM variables. Previously the address of the variable had to be loaded into a register, then that used to access it. Now it can be done in a single instruction. Generic primitive support improves the number of cases Exupery can increase performance. If a primitive is called now that Exupery doesn't know how to compile, it can now dispatch to it via a PIC which is much faster than running through the interpreters lookup code. Eliot's closure bytecodes have been merged in because the Pharo developers would like to start using them. I'd like to see proper closure support. At the moment the VM will run closure code but Exupery will not be able to compile it. That will take another VM change, and an Exupery compiler change to free up the "unused" slot in MethodContext. The subpixel changes to bitblt have been merged into the VM, and the exupery-dev project on the Universes modified to load the packages required to build the FreeType plugin. I've got makefiles that work, they're not ideal but I ran into autoconf/automake issues that prevented using a better solution. Also, I think I know what one of the remaining bugs is. When returning into compiled code from interpreted code the compiled context is not being marked as a root. Contexts are real objects, and need to keep the garbage collector book-keeping. Any objects in old space must be marked as roots if they point to new space. To avoid write barrier checks on context writes (all stack operations, and temporary assignments), all contexts are marked as roots if they're in old space when they're entered. Exupery's code does the marking in the return bytecode because there are less returns than points to return to. Every send, or potential send has an entry point, which is much more than the number of return bytecodes. Next, I'll investigate this potential bug then fix it if it's real. Bryce

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Exupery