- Pharo-dev - lists.pharo.org

Re: [Pharo-project] Cog VM -- Thanks and Performance / Optimization Questions [Micro-Bench Loop]
by John B Thiel Feb. 25, 2011

Feb. 25, 2011

Thanks Eliot, Stef and All for good answers on the Cog VM inlining and implementation details. That is also exciting news about Eliot and MarkusD's work on an adaptive optimizer. By "completely close the gap to C" how close do you mean? The StrongTalk work posits a theoretical best of about 2x, so I suppose you mean in that general 2x-5x neighborhood vs machine code... ? The 20x slower macro algorithm I mentioned is too large to post. Below is a simple micro-loop I made up to explore some aspects of closures, loops and relative VM timings. The C-equivalent machine code takes 400msec on my test machine. In Squeak/Pharo, I see a huge range of speed factors with this, from 7x to 220x (!!) slower than machine code, on current and near-history VMs (Cog and standard Squeak). Consider: "A = to:by:do: loop" benchLooperA |sum| sum:=1e8. sum to: 1 by: -1 do: [:x | sum := (sum bitShift: -1) + x ] . ^sum "B = timesRepeat loop" benchLooperB |sum count| count := sum:=1e8. sum timesRepeat: [ sum := (sum bitShift: -1) + count. count := count - 1 ] . ^sum A and B compute the same except one using #to:by:do:, the other #timesRepeat. The near-machine code for this computation runs about 400msec (0.4 sec) on my test system. (x86 Windows 7, N450 Atom cpu, 1.66GHz) On the following 3 test VM/Images, : 1. Squeak 4.1 stock (vm = Squeak 4.0.2, 2010apr3) 2. Pharo 1.1.1 stock (vm = Cog 2010sep21) 3. Pharo 1.2rc2 + Cog2011feb6 I get the following times for A Time millisecondsToRun: [ Sandbox benchLooperA ] B Time millisecondsToRun: [ Sandbox benchLooperB ] (A, B) execution times rounded to seconds 1. (24, 88) 2. (34, 11) 3. (3, 8) Here we see a 3x-4x difference A to B, with an anomoly that Pharo 1.1.1 is actually much *slower* on #to:by:do: (is it the closure bug (?), see below). (Also we see 8x-11x speedup with the latest Cog-2011feb6, great!) Now here is something odd -- if I invoke the loops via workspace DoIt, the timing changes in Pharo 1.1.1, like this: Time millisecondsToRun: [ |sum| sum:=1e8. sum to: 1 by: -1 do: [:x | sum := (sum bitShift: -1) + x ] . sum ] (highlight and DoIt) (A, B) invoked from workspace, time in seconds 1. (24, 88) 2. (5, 11) 3. (3, 8) The Pharo 1.1.1 timing anomaly for case 2A disappeared - it's now 5 seconds instead of 34. So, 4 questions: * What causes the Pharo 1.1.1. anomaly timing difference for #to:by:do invoked via workspace DoIt vs in the #benchLooper method (5 sec vs 34 sec) ? * Why is the #timesRepeat loop 3x-4x slower than #to:by:do ? Is that simply the difference between inlined and non-inlined methods, or other factors? * Even the best Cog time here is 7x slower vs machine-code (and 20x slower for the timesRepeat case), whereas latest Cog fib: runs at 3x. What factors make this micro-bench loop not as optimizable as fib: for Cog? * Why is the stock Squeak 4.1 VM **220x slower** than machine code for case B = timesRepeat loop, (88 sec vs 0.4 sec) ? Thanks to everyone for insights and comments. -- jbthiel

2 1

[Pharo-project] Pharo mailing list weekly summary #3
by Francois Stephany Feb. 24, 2011

Feb. 24, 2011

Hi guys, I've posted the summary for this week: http://articles.tulipemoutarde.be/pharo-mailing-list-weekly-summary-3 I was away so I certainly forgot some important stuff. Please tell me if I did. I've seen that I'm now in Planet Smalltalk, thanks to the one who registered me ! See you next week, Francois

2 1

[Pharo-project] [update 1.3] #13064
by Marcus Denker Feb. 24, 2011

Feb. 24, 2011

13064 ----- Issue 3350: Network Tests Issue 3694: Balloon message should be opaque for better readability -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

1 0

[Pharo-project] Issue 3515 in pharo: Failing Test: RBInlineMethodFromComponentTest.testInlineEmptyComponentMethod
by pharo＠googlecode.com Feb. 24, 2011

Feb. 24, 2011

Status: Accepted Owner: marcus.denker Labels: Milestone-1.2-DevImage New issue 3515 by marcus.denker: Failing Test: RBInlineMethodFromComponentTest.testInlineEmptyComponentMethod http://code.google.com/p/pharo/issues/detail?id=3515 RBInlineMethodFromComponentTest.testInlineEmptyComponentMethod

1 4

[Pharo-project] Issue 3694 in pharo: Balloon message should be opaque for better readability
by pharo＠googlecode.com Feb. 24, 2011

Feb. 24, 2011

Status: Accepted Owner: hilaire....(a)gmail.com Labels: Milestone-1.2 Difficulty-Easy Importance-High New issue 3694 by hilaire....(a)gmail.com: Balloon message should be opaque for better readability http://code.google.com/p/pharo/issues/detail?id=3694 Currently there are hardly readable

1 6

[Pharo-project] Issue 3350 in pharo: Network Tests
by pharo＠googlecode.com Feb. 24, 2011

Feb. 24, 2011

Status: FixedWaitingToBePharoed Owner: stephane.ducasse Labels: Milestone-1.3 New issue 3350 by stephane.ducasse: Network Tests http://code.google.com/p/pharo/issues/detail?id=3350 Item was changed: ----- Method: Socket>>localAddress (in category 'accessing') ----- localAddress + + self isWaitingForConnection ifFalse: [ + self + waitForConnectionFor: Socket standardTimeout + ifTimedOut: [ ^ByteArray new: 4 ] ]. + ^self primSocketLocalAddress: socketHandle! - self isWaitingForConnection - ifFalse: [[self waitForConnectionFor: Socket standardTimeout] - on: ConnectionTimedOut - do: [:ex | ^ ByteArray new: 4]]. - ^ self primSocketLocalAddress: socketHandle! Item was changed: ----- Method: Socket>>localPort (in category 'accessing') ----- localPort + + self isWaitingForConnection ifFalse: [ + self + waitForConnectionFor: Socket standardTimeout + ifTimedOut: [ ^0] ]. - self isWaitingForConnection - ifFalse: [[self waitForConnectionFor: Socket standardTimeout] - on: ConnectionTimedOut - do: [:ex | ^ 0]]. ^ self primSocketLocalPort: socketHandle! Item was changed: ----- Method: Socket>>waitForAcceptFor: (in category 'waiting') ----- waitForAcceptFor: timeout "Wait and accept an incoming connection. Return nil if it falis" + self waitForConnectionFor: timeout ifTimedOut: [^ nil]. + ^ self isConnected - [self waitForConnectionFor: timeout] on: ConnectionTimedOut do: [:ex | ^nil]. - ^self isConnected ifTrue:[self accept] ! Item was changed: ----- Method: Socket>>waitForConnectionFor:ifTimedOut: (in category 'waiting') ----- waitForConnectionFor: timeout ifTimedOut: timeoutBlock "Wait up until the given deadline for a connection to be established. Return true if it is established by the deadline, false if not." + | startTime msecsDelta msecsEllapsed status | + startTime := Time millisecondClockValue. + msecsDelta := (timeout * 1000) truncated. - | status deadline | - deadline := Socket deadlineSecs: timeout. status := self primSocketConnectionStatus: socketHandle. + [(status = WaitingForConnection) and: [(msecsEllapsed := Time millisecondsSince: startTime) < msecsDelta]] - [(status = WaitingForConnection) and: [Time millisecondClockValue < deadline]] whileTrue: [ + semaphore waitTimeoutMSecs: msecsDelta - msecsEllapsed. - semaphore waitTimeoutMSecs: (deadline - Time millisecondClockValue). status := self primSocketConnectionStatus: socketHandle]. + status = Connected ifFalse: [^timeoutBlock value]. + ^ true! - status = Connected ifFalse: [^timeoutBlock value] - ! Item was changed: ----- Method: Socket>>waitForDataFor:ifClosed:ifTimedOut: (in category 'waiting') ----- waitForDataFor: timeout ifClosed: closedBlock ifTimedOut: timedOutBlock "Wait for the given nr of seconds for data to arrive." + + | startTime msecsDelta | + startTime := Time millisecondClockValue. + msecsDelta := (timeout * 1000) truncated. + [(Time millisecondsSince: startTime) < msecsDelta] whileTrue: [ + (self primSocketReceiveDataAvailable: socketHandle) + ifTrue: [^self]. + self isConnected + ifFalse: [^closedBlock value]. + self readSemaphore waitTimeoutMSecs: + (msecsDelta - (Time millisecondsSince: startTime) max: 0). + ]. - | deadline | - deadline := Socket deadlineSecs: timeout. - - [Time millisecondClockValue < deadline] - whileTrue: [ - (self primSocketReceiveDataAvailable: socketHandle) - ifTrue: [^self]. - self isConnected - ifFalse: [^closedBlock value]. - self readSemaphore waitTimeoutMSecs: (deadline - Time millisecondClockValue)]. - (self primSocketReceiveDataAvailable: socketHandle) ifFalse: [ self isConnected ifTrue: [^timedOutBlock value] + ifFalse: [^closedBlock value]].! - ifFalse: [^closedBlock value]]! Item was changed: ----- Method: Socket>>waitForDisconnectionFor: (in category 'waiting') ----- waitForDisconnectionFor: timeout "Wait for the given nr of seconds for the connection to be broken. Return true if it is broken by the deadline, false if not. The client should know the connection is really going to be closed (e.g., because he has called 'close' to send a close request to the other end) before calling this method." + | startTime msecsDelta status | + startTime := Time millisecondClockValue. + msecsDelta := (timeout * 1000) truncated. - | status deadline | status := self primSocketConnectionStatus: socketHandle. - deadline := Socket deadlineSecs: timeout. [((status == Connected) or: [(status == ThisEndClosed)]) and: + [(Time millisecondsSince: startTime) < msecsDelta]] whileTrue: [ - [Time millisecondClockValue < deadline]] whileTrue: [ self discardReceivedData. + self readSemaphore waitTimeoutMSecs: + (msecsDelta - (Time millisecondsSince: startTime) max: 0). - self readSemaphore waitTimeoutMSecs: (deadline - Time millisecondClockValue). status := self primSocketConnectionStatus: socketHandle]. + ^ status ~= Connected! - - ^ status ~= Connected - ! Item was changed: ----- Method: Socket>>waitForSendDoneFor: (in category 'waiting') ----- waitForSendDoneFor: timeout "Wait up until the given deadline for the current send operation to complete. Return true if it completes by the deadline, false if not." + | startTime msecsDelta msecsEllapsed sendDone | + startTime := Time millisecondClockValue. + msecsDelta := (timeout * 1000) truncated. + [(sendDone := self primSocketSendDone: socketHandle) not and: [ self isConnected - | sendDone deadline | - deadline := Socket deadlineSecs: timeout. - [self isConnected & (sendDone := self primSocketSendDone: socketHandle) not "Connection end and final data can happen fast, so test in this order" + and: [(msecsEllapsed := Time millisecondsSince: startTime) < msecsDelta]]] whileTrue: [ + self writeSemaphore waitTimeoutMSecs: msecsDelta - msecsEllapsed]. - and: [Time millisecondClockValue < deadline]] whileTrue: [ - self writeSemaphore waitTimeoutMSecs: (deadline - Time millisecondClockValue)]. ^ sendDone! Item was changed: ----- Method: SocksSocket>>connectTo:port: (in category 'connection open/close') ----- connectTo: hostAddress port: port self initializeNetwork. self shouldUseSocks ifFalse: [^super connectTo: hostAddress port: port]. super connectTo: socksIP port: socksPort. + self waitForConnectionFor: Socket standardTimeout. - self waitForConnectionUntil: Socket standardDeadline. dstIP := hostAddress. dstPort := port. vers = 4 ifTrue: [self connectSocks4] ifFalse: [self connectSocks5] ! Item was changed: ----- Method: SocksSocket>>connectToHostNamed:port: (in category 'connection open/close') ----- connectToHostNamed: hostName port: port super connectTo: socksIP port: socksPort. + self waitForConnectionFor: Socket standardTimeout. - self waitForConnectionUntil: Socket standardDeadline. dstName := hostName. dstPort := port. vers = 4 ifTrue: [self connectSocks4] ifFalse: [self connectSocks5] !

1 5

[Pharo-project] OpenCroquet Website
by Torsten Bergmann Feb. 24, 2011

Feb. 24, 2011

Bill wrote (on pharo list): >They appear to back full tilt on the Smalltalk front, using Squeak. >http://www.opencobalt.org Yes, I know OpenCobalt and I thought of it as a fork. According to the wikipedia page [1] (which only links to Croquet) it sits on top of the Croquet SDK. Similar to the Krestianstvo SDK. However - would be nice if http://www.opencroquet.org would point also to - http://www.opencobalt.org and - http://www.krestianstvo.org/ if anyone knows a person who is able to change the page. Bye T. [1] http://en.wikipedia.org/wiki/Croquet_Project -- NEU: FreePhone - kostenlos mobil telefonieren und surfen! Jetzt informieren: http://www.gmx.net/de/go/freephone

3 2

[Pharo-project] [update 1.2] #12337
by Marcus Denker Feb. 24, 2011

Feb. 24, 2011

[update 1.2] #12337 12337 ----- Issue 3694: Balloon message should be opaque for better readability Issue 3716: Different return value for adoptInstance: depending on VM [Pharo-project] Browser class list/switches -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

1 0

[Pharo-project] Issue 3716 in pharo: Different return value for adoptInstance: depending on VM
by pharo＠googlecode.com Feb. 24, 2011

Feb. 24, 2011

Status: Accepted Owner: ryd...(a)gmail.com Labels: Milestone-1.2 Milestone-1.3 New issue 3716 by ryd...(a)gmail.com: Different return value for adoptInstance: depending on VM http://code.google.com/p/pharo/issues/detail?id=3716 This method which is new in 1.2, returns differently depending on whether the primitive is present in VM, as the primitive returns the class, while the fallback changeClassTo: returns the instance whose class was changed. Slices for 1.2 and 1.3 will be in inbox shortly for consistently returning the class on all VM's.

1 4

[Pharo-project] [update 1.3] #13063
by Marcus Denker Feb. 24, 2011

Feb. 24, 2011

13063 ----- Issue 3749: new TimeProfiler integration Issue 3740: Provide Settings to directly open the full Debugger View Issue 3723: The Finder tool does not alphabetize methods. [Pharo-project] Browser class list/switches -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

1 0