Hi All,<div><br></div><div> so you'll probably have seen that I actually got round to implementing <span class="Apple-style-span" style="font-family: arial, sans-serif; font-size: 13px; border-collapse: collapse; ">delayed code generation/stack-to-register mapping in Cog, specifically in the StackToRegisterMappingCogit. So let me show you the generated code for comparison with the below. </span><span class="Apple-style-span" style="font-family: arial, sans-serif; font-size: 13px; border-collapse: collapse; "> The scheme implemented by </span><span class="Apple-style-span" style="font-family: arial, sans-serif; font-size: 13px; border-collapse: collapse; ">StackToRegisterMappingCogit is to defer creating machine code until compiling a bytecode that consumes operands, i.e. any bytecode that merely produces operands (a stack push) is handled by pushing a descriptor of that operand on a "simulation stack" in the JIT that holds operand descriptors until they're needed and functions as a simple register allocator. When a bytecode such as a send, a store or a return is generated the JIT examines the descriptors on the simulation stack and so has type information, in particular whether an operand is a SmallInteger or not. The VM also implements a register-based calling convention for arg counts 0 and 1 (self and up to 1 arg passed in registers) which has the effect of avoiding the stack for high-dynamic-frequency arithmetic and accessing primitives.</span></div>
<div><font class="Apple-style-span" face="arial, sans-serif"><span class="Apple-style-span" style="border-collapse: collapse;"><br></span></font></div><div><font class="Apple-style-span" face="arial, sans-serif"><span class="Apple-style-span" style="border-collapse: collapse;">So here's the code (see the quote below for code generated by the SimpleStackbasedCogit). <span class="Apple-style-span" style="border-collapse: separate; font-family: arial; "><span class="Apple-style-span" style="font-family: arial, sans-serif; font-size: 13px; border-collapse: collapse; ">The microbenchmark is </span><span class="Apple-style-span" style="font-family: arial, sans-serif; font-size: 13px; border-collapse: collapse; ">[1 to: 100000000 do: [:i| ]] timeToRun. And here's the code the </span></span></span></font><span class="Apple-style-span" style="font-family: arial, sans-serif; font-size: 13px; border-collapse: collapse; ">StackToRegisterMappingCogit </span><span class="Apple-style-span" style="font-family: arial, sans-serif; font-size: 13px; border-collapse: collapse; ">generates for the loop:</span></div>
<div><font class="Apple-style-span" face="arial, sans-serif"><span class="Apple-style-span" style="border-collapse: collapse;"><span class="Apple-style-span" style="border-collapse: separate; font-family: arial; "><span class="Apple-style-span" style="font-family: arial, sans-serif; font-size: 13px; border-collapse: collapse; "><div>
</div><div><br></div><div><div><br></div><div>25 <76> pushConstant: 1</div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86138: movl $0x00000003, %eax : B8 03 00 00 00 </div><div>26 <68> popIntoTemp: 0</div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>8613d: movl %eax, -16(%ebp) : 89 45 F0</div><div>27 <10> pushTemp: 0</div><div>28 <20> pushConstant: 100000000</div><div>29 <B4> send: <=</div>
<div>30 <9D> jumpFalse: 37</div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86140: movl -16(%ebp), %edx : 8B 55 F0 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86143: movl %edx, %eax : 89 D0 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>86145: andl $0x00000001, %eax : 83 E0 01 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86148: jz .+0x0000000a (0x00086154=16r86090@C4) : 74 0A </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>8614a: cmpl $0x0bebc201, %edx : 81 FA 01 C2 EB 0B </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86150: jnle .+0x00000058 (0x000861aa=16r86090@11A) : 7F 58 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>86152: jmp .+0x00000022 (0x00086176=16r86090@E6) : EB 22 </div><div>(29 <B4> send: <= if i is not a SmallInteger; args passed in edx and esi, result in edx)</div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>86154: movl $0x0bebc201, %esi : BE 01 C2 EB 0B </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86159: movl $0x00448c9c=#<=, %ecx : B9 9C 8C 44 00 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>8615e: call .+0xfff7a30d (0x00000470=ceSend1Args) : E8 0D A3 F7 FF </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>IsSendCall:</div>
<div>(30 <9D> jumpFalse: 37 if #<= is actually sent)</div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86163: movl %edx, %eax : 89 D0 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86165: subl $0x00162270=false, %eax : 2D 70 22 16 00 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>8616a: jz .+0x0000003e (0x000861aa=16r86090@11A) : 74 3E </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>8616c: cmpl $0x00000008, %eax : 83 F8 08 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>8616f: jz .+0x00000005 (0x00086176=16r86090@E6) : 74 05 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86171: call .+0xfff7a7e2 (0x00000958=ceSendMustBeBooleanAddFalseTrampoline) : E8 E2 A7 F7 FF </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>HasBytecodePC:</div><div>31 <10> pushTemp: 0</div><div>32 <76> pushConstant: 1</div><div>33 <B0> send: +</div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86176: movl -16(%ebp), %edx : 8B 55 F0 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>86179: movl %edx, %eax : 89 D0 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>8617b: andl $0x00000001, %eax : 83 E0 01 </div><div>
<span class="Apple-tab-span" style="white-space:pre"> </span>8617e: jz .+0x00000008 (0x00086188=16r86090@F8) : 74 08 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86180: addl $0x00000002, %edx : 83 C2 02 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>86183: jno .+0x00000012 (0x00086197=16r86090@107) : 71 12</div><div>(33 <B0> send: + if either i is not a SmallInteger or if the addition overflows; args passed in edx and esi)</div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>86185: subl $0x00000002, %edx : 83 EA 02 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86188: movl $0x00000003, %esi : BE 03 00 00 00 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>8618d: movl $0x004496a8=#+, %ecx : B9 A8 96 44 00 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86192: call .+0xfff7a2d9 (0x00000470=ceSend1Args) : E8 D9 A2 F7 FF </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>IsSendCall:</div><div>34 <68> popIntoTemp: 0</div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86197: movl %edx, -16(%ebp) : 89 55 F0 </div>
<div>35 <A3 F6> jumpTo: 27</div><div><span class="Apple-tab-span" style="white-space:pre"> </span>8619a: movl %ds:0x0013aafc=&stackLimit, %eax : A1 FC AA 13 00 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>8619f: cmpl %eax, %esp : 39 C4 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>861a1: jnb .+0xffffff9d (0x00086140=16r86090@B0) : 73 9D </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>861a3: call .+0xfff7a920 (0x00000ac8=ceCheckForInterruptsTrampoline) : E8 20 A9 F7 FF </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>HasBytecodePC:</div><div><span class="Apple-tab-span" style="white-space:pre"> </span>861a8: jmp .+0xffffff96 (0x00086140=16r86090@B0) : EB 96 </div></div>
<div><br></div><div>So now Cog is evaluating the following instructions round the loop:</div></span></span></span></font></div><div><font class="Apple-style-span" face="arial, sans-serif"><span class="Apple-style-span" style="border-collapse: collapse; "><br>
</span></font></div><div><font class="Apple-style-span" face="arial, sans-serif"><span class="Apple-style-span" style="border-collapse: collapse;"><span class="Apple-style-span" style="border-collapse: separate; font-family: arial; "><span class="Apple-style-span" style="font-family: arial, sans-serif; border-collapse: collapse; "></span></span><div>
<span class="Apple-tab-span" style="white-space:pre"> </span>86140: movl -16(%ebp), %edx : 8B 55 F0 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86143: movl %edx, %eax : 89 D0 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86145: andl $0x00000001, %eax : 83 E0 01 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>86148: jz .+0x0000000a (0x00086154=16r86090@C4) : 74 0A </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>8614a: cmpl $0x0bebc201, %edx : 81 FA 01 C2 EB 0B </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>86150: jnle .+0x00000058 (0x000861aa=16r86090@11A) : 7F 58 </div><div>...</div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86176: movl -16(%ebp), %edx : 8B 55 F0 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>86179: movl %edx, %eax : 89 D0 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>8617b: andl $0x00000001, %eax : 83 E0 01 </div><div>
<span class="Apple-tab-span" style="white-space:pre"> </span>8617e: jz .+0x00000008 (0x00086188=16r86090@F8) : 74 08 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86180: addl $0x00000002, %edx : 83 C2 02 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>86183: jno .+0x00000012 (0x00086197=16r86090@107) : 71 12</div><div>...</div><div><span class="Apple-tab-span" style="white-space:pre"> </span>86197: movl %edx, -16(%ebp) : 89 55 F0 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>8619a: movl %ds:0x0013aafc=&stackLimit, %eax : A1 FC AA 13 00 </div><div><span class="Apple-tab-span" style="white-space:pre"> </span>8619f: cmpl %eax, %esp : 39 C4 </div>
<div><span class="Apple-tab-span" style="white-space:pre"> </span>861a1: jnb .+0xffffff9d (0x00086140=16r86090@B0) : 73 9D </div><div><br></div><div>which compares well with those executed by VW:</div><div><br></div></span></font><span class="Apple-style-span" style="font-family: arial, sans-serif; font-size: 13px; border-collapse: collapse; "><div>
<span style="white-space: pre-wrap; "> </span>nm@040:<span style="white-space: pre-wrap; "> </span>mov -0x10(bFrame),rReceiver<span style="white-space: pre-wrap; "> </span>! 8b 5d f0</div><div><span style="white-space: pre-wrap; "> </span>nm@043:<span style="white-space: pre-wrap; "> </span>mov $0x17d78403,rArg1<span style="white-space: pre-wrap; "> </span>! be 03 84 d7 17</div>
<div><span style="white-space: pre-wrap; "> </span>nm@048:<span style="white-space: pre-wrap; "> </span>testb $2,%bl<span style="white-space: pre-wrap; "> </span>! f6 c3 02 a.k.a. testb $2,rReceiver</div><div><span style="white-space: pre-wrap; "> </span>nm@04b:<span style="white-space: pre-wrap; "> </span>jz nm@056<span style="white-space: pre-wrap; "> </span>! 74 09</div>
<div><span style="white-space: pre-wrap; "> </span>nm@04d:<span style="white-space: pre-wrap; "> </span>cmp rArg1,rReceiver<span style="white-space: pre-wrap; "> </span>! 3b de</div><div><span style="white-space: pre-wrap; "> </span>nm@04f:<span style="white-space: pre-wrap; "> </span>jle nm@075<span style="white-space: pre-wrap; "> </span>! 7e 24</div>
<div><span style="white-space: pre-wrap; "> </span>nm@051:<span style="white-space: pre-wrap; "> </span>jmp nm@0ab<span style="white-space: pre-wrap; "> </span>! e9 55 00 00 00</div><div>...</div><div><span style="white-space: pre-wrap; "><span class="Apple-style-span" style="white-space: normal; "><div>
<span style="white-space: pre-wrap; "> </span>nm@075:<span style="white-space: pre-wrap; "> </span>mov -0x10(bFrame),rReceiver<span style="white-space: pre-wrap; "> </span>! 8b 5d f0</div><div><span style="white-space: pre-wrap; "> </span>nm@078:<span style="white-space: pre-wrap; "> </span>testb $2,%bl<span style="white-space: pre-wrap; "> </span>! f6 c3 02 a.k.a. testb $2,rReceiver</div>
<div><span style="white-space: pre-wrap; "> </span>nm@07b:<span style="white-space: pre-wrap; "> </span>jz nm@085<span style="white-space: pre-wrap; "> </span>! 74 08</div></span> </span>nm@07d:<span style="white-space: pre-wrap; "> </span>add $4,rReceiver<span style="white-space: pre-wrap; "> </span>! 83 c3 04</div>
<div><span style="white-space: pre-wrap; "> </span>nm@080:<span style="white-space: pre-wrap; "> </span>jno nm@092<span style="white-space: pre-wrap; "> </span>! 71 10</div><div>...</div><div><span style="white-space: pre-wrap; "> </span>nm@092:<span style="white-space: pre-wrap; "> </span>mov rReceiver,-0x10(bFrame)<span style="white-space: pre-wrap; "> </span>! 89 5d f0</div>
<div><span style="white-space: pre-wrap; "> </span>nm@095:<span style="white-space: pre-wrap; "> </span>cmp 0xfb190,bSP<span style="white-space: pre-wrap; "> </span>! 3b 25 90 b1 0f 00</div><div><span style="white-space: pre-wrap; "> </span>nm@09b:<span style="white-space: pre-wrap; "> </span>jae nm@040<span style="white-space: pre-wrap; "> </span>! 0f 83 9f ff ff ff</div>
<div><br></div><div><br></div><div>Using the non-destructive testb avoids a register copy, but there's only one more instruction. I'm doing slightly better than VW in that I don't assign the limit 100000000 (<span class="Apple-style-span" style="font-size: small; ">0x0bebc201 in Squeak, </span>0x17d78403 in VW) to the arg register unless i is not a SmallInteger, and I'm able to generate short jumps (2 bytes vs 5 bytes). But these are angels on the heads of pins. The code generators are essentially equivalent.</div>
<div><br></div><div>best</div><div>Eliot</div><div><br></div></span><div class="gmail_quote">On Fri, Nov 19, 2010 at 11:48 AM, Eliot Miranda <span dir="ltr"><<a href="mailto:eliot.miranda@gmail.com">eliot.miranda@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Hi All,<div><br></div><div> so the performance benefit of the special selectors is interesting. One thing the interpreter does on all relational special selectors #= #~= #< #<= #> #>= is statically predict SmallIntegers and/or Floats and look ahead for a following conditional branch, evaluating the branch immediately, avoiding reifying the result of the relational as either true or false and later testing for it, hence jumping directly on the condition codes.</div>
<div><br></div><div>e.g.</div><div><div>bytecodePrimLessThan</div><div><span style="white-space:pre-wrap"> </span>| rcvr arg aBool |</div><div><span style="white-space:pre-wrap"> </span>rcvr := self internalStackValue: 1.</div>
<div><span style="white-space:pre-wrap"> </span>arg := self internalStackValue: 0.</div><div><span style="white-space:pre-wrap"> </span>(self areIntegers: rcvr and: arg) ifTrue:</div><div>
<span style="white-space:pre-wrap"> </span>["The C code can avoid detagging since tagged integers are still signed.</div><div><span style="white-space:pre-wrap"> </span> But this means the simulator must override to do detagging."</div>
<div><span style="white-space:pre-wrap"> </span>^self cCode: [self booleanCheat: rcvr < arg]</div><div><span style="white-space:pre-wrap"> </span>inSmalltalk: [self booleanCheat: (objectMemory integerValueOf: rcvr) < (objectMemory integerValueOf: arg)]].</div>
<div><br></div><div><span style="white-space:pre-wrap"> </span>self initPrimCall.</div><div><span style="white-space:pre-wrap"> </span>aBool := self primitiveFloatLess: rcvr thanArg: arg.</div>
<div><span style="white-space:pre-wrap"> </span>self successful ifTrue: [^ self booleanCheat: aBool].</div><div><br></div><div><span style="white-space:pre-wrap"> </span>messageSelector := self specialSelector: 2.</div>
<div><span style="white-space:pre-wrap"> </span>argumentCount := 1.</div><div><span style="white-space:pre-wrap"> </span>self normalSend</div><div><br></div><div><div>booleanCheat: cond</div>
<div><span style="white-space:pre-wrap"> </span>"cheat the interpreter out of the pleasure of handling the next bytecode IFF it is a jump-on-boolean. Which it is, often enough when the current bytecode is something like bytecodePrimEqual"</div>
<div><span style="white-space:pre-wrap"> </span><inline: true></div><div><br></div><div><span style="white-space:pre-wrap"> </span>cond</div><div><span style="white-space:pre-wrap"> </span>ifTrue: [self booleanCheatTrue]</div>
<div><span style="white-space:pre-wrap"> </span>ifFalse: [self booleanCheatFalse]</div></div><div><br></div><div><div>booleanCheatFalse</div><div><span style="white-space:pre-wrap"> </span>"cheat the interpreter out of the pleasure of handling the next bytecode IFF it is a jump-on-boolean. Which it is, often enough when the current bytecode is something like bytecodePrimEqual"</div>
<div><span style="white-space:pre-wrap"> </span>| bytecode offset |</div><div><span style="white-space:pre-wrap"> </span><sharedCodeNamed: 'booleanCheatFalse' inCase: 179></div>
<div><br></div><div><span style="white-space:pre-wrap"> </span>bytecode := self fetchByte. "assume next bytecode is jumpIfFalse (99%)"</div><div><span style="white-space:pre-wrap"> </span>self internalPop: 2.</div>
<div><span style="white-space:pre-wrap"> </span>(bytecode < 160 and: [bytecode > 151]) ifTrue: "short jumpIfFalse"</div><div><span style="white-space:pre-wrap"> </span>[^self jump: bytecode - 151].</div>
<div><br></div><div><span style="white-space:pre-wrap"> </span>bytecode = 172 ifTrue: "long jumpIfFalse"</div><div><span style="white-space:pre-wrap"> </span>[offset := self fetchByte.</div>
<div><span style="white-space:pre-wrap"> </span>^self jump: offset].</div><div><br></div><div><span style="white-space:pre-wrap"> </span>"not followed by a jumpIfFalse; undo instruction fetch and push boolean result"</div>
<div><span style="white-space:pre-wrap"> </span>localIP := localIP - 1.</div><div><span style="white-space:pre-wrap"> </span>self fetchNextBytecode.</div><div><span style="white-space:pre-wrap"> </span>self internalPush: objectMemory falseObject</div>
</div><div><br></div><div><div>booleanCheatTrue</div><div><span style="white-space:pre-wrap"> </span>"cheat the interpreter out of the pleasure of handling the next bytecode IFF it is a jump-on-boolean. Which it is, often enough when the current bytecode is something like bytecodePrimEqual"</div>
<div><span style="white-space:pre-wrap"> </span>| bytecode |</div><div><span style="white-space:pre-wrap"> </span><sharedCodeNamed: 'booleanCheatTrue' inCase: 178></div><div>
<br></div><div><span style="white-space:pre-wrap"> </span>bytecode := self fetchByte. "assume next bytecode is jumpIfFalse (99%)"</div><div><span style="white-space:pre-wrap"> </span>self internalPop: 2.</div>
<div><span style="white-space:pre-wrap"> </span>(bytecode < 160 and: [bytecode > 151]) ifTrue: "short jumpIfFalse"</div><div><span style="white-space:pre-wrap"> </span>[^self fetchNextBytecode].</div>
<div><br></div><div><span style="white-space:pre-wrap"> </span>bytecode = 172 ifTrue: "long jumpIfFalse"</div><div><span style="white-space:pre-wrap"> </span>[self fetchByte.</div>
<div><span style="white-space:pre-wrap"> </span>^self fetchNextBytecode].</div><div><br></div><div><span style="white-space:pre-wrap"> </span>"not followed by a jumpIfFalse; undo instruction fetch and push boolean result"</div>
<div><span style="white-space:pre-wrap"> </span>localIP := localIP - 1.</div><div><span style="white-space:pre-wrap"> </span>self fetchNextBytecode.</div><div><span style="white-space:pre-wrap"> </span>self internalPush: objectMemory trueObject</div>
</div><div><br></div><div><br></div>Until yesterday Cog didn't do this for jitted code. The conversation about using #= for testing got me motivated to implement it. Note that VisualWorks' HPS VM does something even but more aggressive than booleanCheat:. So what are we talking about? Here's my micro-benchmark:</div>
<div><br></div><div>Time millisecondsToRun: [1 to: 100000000 do: [:i| ]]</div><div><br></div><div>The bytecode compiler optimizes the inner loop to code equivalent to</div><div><br></div><div>| i |</div><div>i := 1.</div>
<div>[i <= 100000000] whileTrue: [i := i + 1]</div><div><br></div><div>Here's the Squeak bytecode:</div><div><br></div><div><div>25 <76> pushConstant: 1</div><div>26 <68> popIntoTemp: 0</div><div>27 <10> pushTemp: 0</div>
<div>28 <20> pushConstant: 100000000</div><div>29 <B4> send: <=</div><div>30 <9D> jumpFalse: 37</div><div>31 <10> pushTemp: 0</div><div>32 <76> pushConstant: 1</div><div>33 <B0> send: +</div>
<div>34 <68> popIntoTemp: 0</div><div>35 <A3 F6> jumpTo: 27</div></div><div>37:</div><div><br></div><div>And here's the VisualWorks bytecode:</div><div><div> 6 <4A> push 1</div><div> 7 <4C> store local 0; pop</div>
<div> 8 <67> loop head</div><div> 9 <10> push local 0</div><div>10 <1C> push 100000000</div><div>11 <A4> send <=</div><div>12 <C4> jump false 18</div><div>13 <10> push local 0</div>
<div>
14 <C8> push 1; send +</div><div>15 <4C> store local 0; pop</div><div>16 <E3 F6> jump 8</div></div><div>18:</div><div><br></div><div>Note that loopHead is a no-op that allows the HPS JIT to be one pass; it tells the JIT that there is a backward branch to this instruction and so knows that loopHead is a control-flow join, which means that the registers that cache the receiver and its base pointer (VW has an indirection from the object header to an object body) must be reloaded if needed.</div>
<div><br></div><div>So the current version of Cog generates the following code for the loop:</div><div><br></div><div><div>25 <76> pushConstant: 1</div><div><span style="white-space:pre-wrap"> </span>a1ea4: movl $0x00000003, %eax : B8 03 00 00 00 </div>
<div><span style="white-space:pre-wrap"> </span>a1ea9: pushl %eax : 50 </div><div>26 <68> popIntoTemp: 0</div><div><span style="white-space:pre-wrap"> </span>a1eaa: popl %eax : 58 </div>
<div><span style="white-space:pre-wrap"> </span>a1eab: movl %eax, %ss:0xfffffff0(%ebp) : 89 45 F0 </div><div>27 <10> pushTemp: 0</div><div><span style="white-space:pre-wrap"> </span>a1eae: movl %ss:0xfffffff0(%ebp), %eax : 8B 45 F0 </div>
<div><span style="white-space:pre-wrap"> </span>a1eb1: pushl %eax : 50 </div><div>28 <20> pushConstant: 100000000</div><div><span style="white-space:pre-wrap"> </span>a1eb2: pushl $0x0bebc201 : 68 01 C2 EB 0B </div>
<div>29 <B4> send: <=</div><div><span style="white-space:pre-wrap"> </span>a1eb7: movl %ss:0x4(%esp), %edx : 8B 54 24 04 </div><div><span style="white-space:pre-wrap"> </span>a1ebb: movl $0x0045a458=#<=, %ecx : B9 58 A4 45 00 </div>
<div><span style="white-space:pre-wrap"> </span>a1ec0: call .+0xfff5e5cb (0x00000490=ceSend1Args) : E8 CB E5 F5 FF </div><div><span style="white-space:pre-wrap"> </span>IsSendCall:</div>
<div><span style="white-space:pre-wrap"> </span>a1ec5: pushl %edx : 52 </div><div>30 <9D> jumpFalse: 37</div><div><span style="white-space:pre-wrap"> </span>a1ec6: popl %eax : 58 </div>
<div><span style="white-space:pre-wrap"> </span>a1ec7: subl $0x00172270=false, %eax : 2D 70 22 17 00 </div><div><span style="white-space:pre-wrap"> </span>IsObjectReference:</div><div>
<span style="white-space:pre-wrap"> </span>a1ecc: jz .+0x0000003d (0x000a1f0b=16rA1E50@BB) : 74 3D </div><div><span style="white-space:pre-wrap"> </span>a1ece: cmpl $0x00000008, %eax : 83 F8 08 </div>
<div><span style="white-space:pre-wrap"> </span>a1ed1: jz .+0x0000000b (0x000a1ede=16rA1E50@8E) : 74 0B </div><div><span style="white-space:pre-wrap"> </span>a1ed3: addl $0x00172270=false, %eax : 05 70 22 17 00 </div>
<div><span style="white-space:pre-wrap"> </span>IsObjectReference:</div><div><span style="white-space:pre-wrap"> </span>a1ed8: pushl %eax : 50 </div><div><span style="white-space:pre-wrap"> </span>a1ed9: call .+0xfff5e902 (0x000007e0=ceSendMustBeBooleanTrampoline) : E8 02 E9 F5 FF </div>
<div><span style="white-space:pre-wrap"> </span>HasBytecodePC:</div><div>31 <10> pushTemp: 0</div><div><span style="white-space:pre-wrap"> </span>a1ede: movl %ss:0xfffffff0(%ebp), %eax : 8B 45 F0 </div>
<div><span style="white-space:pre-wrap"> </span>a1ee1: pushl %eax : 50 </div><div>32 <76> pushConstant: 1</div><div><span style="white-space:pre-wrap"> </span>a1ee2: movl $0x00000003, %eax : B8 03 00 00 00 </div>
<div><span style="white-space:pre-wrap"> </span>a1ee7: pushl %eax : 50 </div><div>33 <B0> send: +</div><div><span style="white-space:pre-wrap"> </span>a1ee8: movl %ss:0x4(%esp), %edx : 8B 54 24 04 </div>
<div><span style="white-space:pre-wrap"> </span>a1eec: movl $0x0045ae64=#+, %ecx : B9 64 AE 45 00 </div><div><span style="white-space:pre-wrap"> </span>a1ef1: call .+0xfff5e59a (0x00000490=ceSend1Args) : E8 9A E5 F5 FF </div>
<div><span style="white-space:pre-wrap"> </span>IsSendCall:</div><div><span style="white-space:pre-wrap"> </span>a1ef6: pushl %edx : 52 </div><div>34 <68> popIntoTemp: 0</div><div>
<span style="white-space:pre-wrap"> </span>a1ef7: popl %eax : 58 </div><div><span style="white-space:pre-wrap"> </span>a1ef8: movl %eax, %ss:0xfffffff0(%ebp) : 89 45 F0 </div><div>35 <A3 F6> jumpTo: 27</div>
<div><span style="white-space:pre-wrap"> </span>a1efb: movl %ds:0x0013aafc=&stackLimit, %eax : A1 FC AA 13 00 </div><div><span style="white-space:pre-wrap"> </span>a1f00: cmpl %eax, %esp : 39 C4 </div>
<div><span style="white-space:pre-wrap"> </span>a1f02: jnb .+0xffffffaa (0x000a1eae=16rA1E50@5E) : 73 AA </div><div><span style="white-space:pre-wrap"> </span>a1f04: call .+0xfff5e907 (0x00000810=ceCheckForInterruptsTrampoline) : E8 07 E9 F5 FF </div>
<div><span style="white-space:pre-wrap"> </span>HasBytecodePC:</div><div><span style="white-space:pre-wrap"> </span>a1f09: jmp .+0xffffffa3 (0x000a1eae=16rA1E50@5E) : EB A3 </div><div>
37:<span style="white-space:pre-wrap"> </span>0x000a1f0b/16rA1E50@BB:</div></div><div><br></div><div>And here's the primitive part of the SmallInteger>#<= method; the SmallInteger<#+ method is similar. I'll omit it for brevity.</div>
<div><br></div><div><div><span style="white-space:pre-wrap"> </span>entry: (check that the receiver matches the inline cache)</div><div><span style="white-space:pre-wrap"> </span>00002498: movl %edx, %eax : 89 D0 </div>
<div><span style="white-space:pre-wrap"> </span>0000249a: andl $0x00000001, %eax : 83 E0 01 </div><div><span style="white-space:pre-wrap"> </span>0000249d: jnz .+0x00000010 (0x000024af=<=@37) : 75 10 (jump if the receiver is a SmallInteger, which in our case it is)</div>
<div><span style="white-space:pre-wrap"> </span>0000249f: movl %ds:(%edx), %eax : 8B 02 </div><div><span style="white-space:pre-wrap"> </span>000024a1: shrl $0x0a, %eax : C1 E8 0A </div>
<div><span style="white-space:pre-wrap"> </span>000024a4: andl $0x0000007c, %eax : 83 E0 7C </div><div><span style="white-space:pre-wrap"> </span>000024a7: jnz .+0x00000006 (0x000024af=<=@37) : 75 06 </div>
<div><span style="white-space:pre-wrap"> </span>000024a9: movl %ds:0xfffffffc(%edx), %eax : 8B 42 FC </div><div><span style="white-space:pre-wrap"> </span>000024ac: andl $0xfffffffc, %eax : 83 E0 FC </div>
<div><span style="white-space:pre-wrap"> </span>000024af: cmpl %ecx, %eax : 39 C8 (compare the receiver's tags against the inline cache ecx, which in our case will match)</div>
<div><span style="white-space:pre-wrap"> </span>000024b1: jnz .+0xffffffdf (0x00002492=<=@1A) : 75 DF </div><div><span style="white-space:pre-wrap"> </span>noCheckEntry: (SmallInteger primitive #<=)</div>
<div><span style="white-space:pre-wrap"> </span>000024b3: movl %ss:0x4(%esp), %eax : 8B 44 24 04 </div><div><span style="white-space:pre-wrap"> </span>000024b7: movl %eax, %ecx : 89 C1 </div>
<div><span style="white-space:pre-wrap"> </span>000024b9: andl $0x00000001, %eax : 83 E0 01 </div><div><span style="white-space:pre-wrap"> </span>000024bc: jz .+0x00000014 (0x000024d2=<=@5A) : 74 14 (is the argument a SmallInteger, which in our case it will be)</div>
<div><span style="white-space:pre-wrap"> </span>000024be: cmpl %ecx, %edx : 39 CA </div><div><span style="white-space:pre-wrap"> </span>000024c0: jle .+0x00000008 (0x000024ca=<=@52) : 7E 08 </div>
<div><span style="white-space:pre-wrap"> </span>000024c2: movl $0x00172270=false, %edx : BA 70 22 17 00 (answer the false object)</div><div><span style="white-space:pre-wrap"> </span>IsObjectReference:</div>
<div><span style="white-space:pre-wrap"> </span>000024c7: ret $0x0008 : C2 08 00 </div><div><span style="white-space:pre-wrap"> </span>000024ca: movl $0x00172278=true, %edx : BA 78 22 17 00 (answer the true object)</div>
<div><span style="white-space:pre-wrap"> </span>IsObjectReference:</div><div><span style="white-space:pre-wrap"> </span>000024cf: ret $0x0008 : C2 08 00 </div><div><span style="white-space:pre-wrap"> </span>000024d2:</div>
</div><div><br></div><div>You can see that the Cog code generator is extremely naive; it's generating RISC code. There are no compares against memory; only against registers, etc.</div><div><br></div><div>The question I want you to ask yourself is how much faster the loop will go if we add code to short-circuit the send of #<= and the comparison against true and false with a tag test and a direct comparison. Try answering it as we go along. I'll answer it below but try and answer it yourself. Basically we should speed up about a third of the loop substantially where the loop has three main parts, a) the compare and branch [i <= 100000000] whileTrue:, b) the increment i := i + 1, and c) the backward branch at the end of the while loop which checks the stackLimit to break out if there is an event (e.g. ctrl-.).</div>
<div><br></div><div>OK, so here's the code with the booleanCheat implemented in the JIT:</div><div><div><br></div><div><div>25 <76> pushConstant: 1</div><div><span style="white-space:pre-wrap"> </span>a52cc: movl $0x00000003, %eax : B8 03 00 00 00 </div>
<div><span style="white-space:pre-wrap"> </span>a52d1: pushl %eax : 50 </div><div>26 <68> popIntoTemp: 0</div><div><span style="white-space:pre-wrap"> </span>a52d2: popl %eax : 58 </div>
<div><span style="white-space:pre-wrap"> </span>a52d3: movl %eax, %ss:0xfffffff0(%ebp) : 89 45 F0 </div><div>27 <10> pushTemp: 0</div><div><span style="white-space:pre-wrap"> </span>a52d6: movl %ss:0xfffffff0(%ebp), %eax : 8B 45 F0 </div>
<div><span style="white-space:pre-wrap"> </span>a52d9: pushl %eax : 50 </div><div>28 <20> pushConstant: 100000000</div><div><span style="white-space:pre-wrap"> </span>a52da: pushl $0x0bebc201 : 68 01 C2 EB 0B </div>
<div>29 <B4> send: <=</div><div><span style="white-space:pre-wrap"> </span>a52df: movl %ss:(%esp), %esi : 8B 34 24 </div><div><span style="white-space:pre-wrap"> </span>a52e2: movl %esi, %eax : 89 F0 </div>
<div><span style="white-space:pre-wrap"> </span>a52e4: movl %ss:0x4(%esp), %edx : 8B 54 24 04 </div><div><span style="white-space:pre-wrap"> </span>a52e8: andl %edx, %eax : 23 C2 </div>
<div><span style="white-space:pre-wrap"> </span>a52ea: andl $0x00000001, %eax : 83 E0 01 (are receiver and arg SmallIntegers, which they are)</div><div><span style="white-space:pre-wrap"> </span>a52ed: jz .+0x00000009 (0x000a52f8=16rA5278@80) : 74 09 </div>
<div><span style="white-space:pre-wrap"> </span>a52ef: addl $0x00000008, %esp : 83 C4 08 </div><div><span style="white-space:pre-wrap"> </span>a52f2: cmpl %esi, %edx : 39 F2 (compare directly)</div>
<div><span style="white-space:pre-wrap"> </span>a52f4: jnle .+0x00000056 (0x000a534c=16rA5278@D4) : 7F 56 (jump on condition codes)</div><div><span style="white-space:pre-wrap"> </span>a52f6: jmp .+0x00000027 (0x000a531f=16rA5278@A7) : EB 27 (jump past the send and the jumpFalse:)</div>
<div><span style="white-space:pre-wrap"> </span>a52f8: movl %ss:0x4(%esp), %edx : 8B 54 24 04 </div><div><span style="white-space:pre-wrap"> </span>a52fc: movl $0x0045a458, %ecx : B9 58 A4 45 00 </div>
<div><span style="white-space:pre-wrap"> </span>a5301: call .+0xfff5b18a (0x00000490=ceSend1Args) : E8 8A B1 F5 FF </div><div><span style="white-space:pre-wrap"> </span>IsSendCall:</div>
<div><span style="white-space:pre-wrap"> </span>a5306: pushl %edx : 52 </div><div>30 <9D> jumpFalse: 37</div><div><span style="white-space:pre-wrap"> </span>a5307: popl %eax : 58 </div>
<div><span style="white-space:pre-wrap"> </span>a5308: subl $0x00172270=false, %eax : 2D 70 22 17 00 </div><div><span style="white-space:pre-wrap"> </span>IsObjectReference:</div><div>
<span style="white-space:pre-wrap"> </span>a530d: jz .+0x0000003d (0x000a534c=16rA5278@D4) : 74 3D </div><div><span style="white-space:pre-wrap"> </span>a530f: cmpl $0x00000008, %eax : 83 F8 08 </div>
<div><span style="white-space:pre-wrap"> </span>a5312: jz .+0x0000000b (0x000a531f=16rA5278@A7) : 74 0B </div><div><span style="white-space:pre-wrap"> </span>a5314: addl $0x00172270=false, %eax : 05 70 22 17 00 </div>
<div><span style="white-space:pre-wrap"> </span>IsObjectReference:</div><div><span style="white-space:pre-wrap"> </span>a5319: pushl %eax : 50 </div><div><span style="white-space:pre-wrap"> </span>a531a: call .+0xfff5b4c1 (0x000007e0=ceSendMustBeBooleanTrampoline) : E8 C1 B4 F5 FF </div>
<div><span style="white-space:pre-wrap"> </span>HasBytecodePC:</div><div>31 <10> pushTemp: 0</div><div><span style="white-space:pre-wrap"> </span>a531f: movl %ss:0xfffffff0(%ebp), %eax : 8B 45 F0 </div>
<div><span style="white-space:pre-wrap"> </span>a5322: pushl %eax : 50 </div><div>32 <76> pushConstant: 1</div><div><span style="white-space:pre-wrap"> </span>a5323: movl $0x00000003, %eax : B8 03 00 00 00 </div>
<div><span style="white-space:pre-wrap"> </span>a5328: pushl %eax : 50 </div><div>33 <B0> send: +</div><div><span style="white-space:pre-wrap"> </span>a5329: movl %ss:0x4(%esp), %edx : 8B 54 24 04 </div>
<div><span style="white-space:pre-wrap"> </span>a532d: movl $0x0045ae64=#+, %ecx : B9 64 AE 45 00 </div><div><span style="white-space:pre-wrap"> </span>a5332: call .+0xfff5b159 (0x00000490=ceSend1Args) : E8 59 B1 F5 FF </div>
<div><span style="white-space:pre-wrap"> </span>IsSendCall:</div><div><span style="white-space:pre-wrap"> </span>a5337: pushl %edx : 52 </div><div>34 <68> popIntoTemp: 0</div><div>
<span style="white-space:pre-wrap"> </span>a5338: popl %eax : 58 </div><div><span style="white-space:pre-wrap"> </span>a5339: movl %eax, %ss:0xfffffff0(%ebp) : 89 45 F0 </div><div>35 <A3 F6> jumpTo: 27</div>
<div><span style="white-space:pre-wrap"> </span>a533c: movl %ds:0x0013aafc=&stackLimit, %eax : A1 FC AA 13 00 </div><div><span style="white-space:pre-wrap"> </span>a5341: cmpl %eax, %esp : 39 C4 </div>
<div><span style="white-space:pre-wrap"> </span>a5343: jnb .+0xffffff91 (0x000a52d6=16rA5278@5E) : 73 91 </div><div><span style="white-space:pre-wrap"> </span>a5345: call .+0xfff5b4c6 (0x00000810=ceCheckForInterruptsTrampoline) : E8 C6 B4 F5 FF </div>
<div><span style="white-space:pre-wrap"> </span>HasBytecodePC:</div><div><span style="white-space:pre-wrap"> </span>a534a: jmp .+0xffffff8a (0x000a52d6=16rA5278@5E) : EB 8A </div></div>
<div><br></div><div>So roughly 1/3 of the loop has been sped up substantially. How much speedup?</div><div><br></div><div>2.7%. 2.7 measly percent. A good 2 - 3 hours work and 2.7 measly % ?!?!? i.e. 691 milliseconds fell to 672 milliseconds (and the measurements are nicely repeatable).</div>
<div><br></div><div>Well, one reason for the limited performance increase could be that the event check at the backward branch is being taken a lot and dominating costs. So what happens if I change the heartbeat from 2KHz to 20KHz? 672 falls to 664, so only 9 milliseconds or so is due to the stack limit event check.</div>
<div><br></div><div>Perhaps the x86 is so good at optimizing the code that it can't be sped up. Well, let's have a look at HPS's performance. It is *4* times faster, 168 ms vs 672 (that's *exactly* 4 times faster :) ). So what code does HPS generate? First its back-end is less naive; it doesn't move intermediate results through registers all the time. But most significantly it is able to know that both the 100000000 and the + 1 are SmallIntegers because it delays generating code until it gets to a send. So its code is a /lot/ better. It short-circuits bth the compare and branch and the increment. Here it is (remember that Squeak has only one tagged type, SmallInteger with tag 1, and 32-bit VW has two tagged types SmallInteger and Character with tags 3 and 1 respectively, so in VW 1 == 0x7):</div>
<div><br></div><div><div> 6 <4A> push 1</div><div> 7 <4C> store local 0; pop</div><div><span style="white-space:pre-wrap"> </span>nm@039:<span style="white-space:pre-wrap"> </span>xor rTemp,rTemp<span style="white-space:pre-wrap"> </span>! 33 c0</div>
<div><span style="white-space:pre-wrap"> </span>nm@03b:<span style="white-space:pre-wrap"> </span>movb $7,%al<span style="white-space:pre-wrap"> </span>! b0 07</div><div>
<span style="white-space:pre-wrap"> </span>nm@03d:<span style="white-space:pre-wrap"> </span>mov rTemp,-0x10(bFrame)<span style="white-space:pre-wrap"> </span>! 89 45 f0</div>
<div> 8 <67> loop head</div><div> 9 <10> push local 0</div><div><span style="white-space:pre-wrap"> </span>nm@040:<span style="white-space:pre-wrap"> </span>mov -0x10(bFrame),rReceiver<span style="white-space:pre-wrap"> </span>! 8b 5d f0</div>
<div>10 <1C> push 100000000</div><div><span style="white-space:pre-wrap"> </span>nm@043:<span style="white-space:pre-wrap"> </span>mov $0x17d78403,rArg1<span style="white-space:pre-wrap"> </span>! be 03 84 d7 17</div>
<div>11 <A4> send <=</div><div><span style="white-space:pre-wrap"> </span>nm@048:<span style="white-space:pre-wrap"> </span>testb $2,%bl<span style="white-space:pre-wrap"> </span>! f6 c3 02 a.k.a. testb $2,rReceiver</div>
<div><span style="white-space:pre-wrap"> </span>nm@04b:<span style="white-space:pre-wrap"> </span>jz nm@056<span style="white-space:pre-wrap"> </span>! 74 09</div><div>
<span style="white-space:pre-wrap"> </span>nm@04d:<span style="white-space:pre-wrap"> </span>cmp rArg1,rReceiver<span style="white-space:pre-wrap"> </span>! 3b de</div>
<div><span style="white-space:pre-wrap"> </span>nm@04f:<span style="white-space:pre-wrap"> </span>jle nm@075<span style="white-space:pre-wrap"> </span>! 7e 24</div><div>
<span style="white-space:pre-wrap"> </span>nm@051:<span style="white-space:pre-wrap"> </span>jmp nm@0ab<span style="white-space:pre-wrap"> </span>! e9 55 00 00 00</div>
<div><span style="white-space:pre-wrap"> </span>nm@056:<span style="white-space:pre-wrap"> </span>mov $0x16a8ec6c,rClass<span style="white-space:pre-wrap"> </span>! ba 6c ec a8 16</div>
<div><span style="white-space:pre-wrap"> </span>nm@05b:<span style="white-space:pre-wrap"> </span>call _81080=send1Args<span style="white-space:pre-wrap"> </span>! e8 28 e7 6b ea</div>
<div><span style="white-space:pre-wrap"> </span>map: 0x3a: send1(26) #<=</div><div>vpc 12:</div><div>12 <C4> jump false 18</div><div><span style="white-space:pre-wrap"> </span>nm@060:<span style="white-space:pre-wrap"> </span>cmp $0x16b33da4,rReceiver<span style="white-space:pre-wrap"> </span>! 81 fb a4 3d b3 16</div>
<div><span style="white-space:pre-wrap"> </span>nm@066:<span style="white-space:pre-wrap"> </span>jz nm@0ab<span style="white-space:pre-wrap"> </span>! 74 43</div><div>
<span style="white-space:pre-wrap"> </span>nm@068:<span style="white-space:pre-wrap"> </span>cmp $0x16da90e4,rReceiver<span style="white-space:pre-wrap"> </span>! 81 fb e4 90 da 16</div>
<div><span style="white-space:pre-wrap"> </span>nm@06e:<span style="white-space:pre-wrap"> </span>jz nm@075<span style="white-space:pre-wrap"> </span>! 74 05</div><div>
<span style="white-space:pre-wrap"> </span>nm@070:<span style="white-space:pre-wrap"> </span>call _159cadd8<span style="white-space:pre-wrap"> </span>! e8 6b 84 00 00</div>
<div><span style="white-space:pre-wrap"> </span>nm@075:<span style="white-space:pre-wrap"> </span>mov -0x10(bFrame),rReceiver<span style="white-space:pre-wrap"> </span>! 8b 5d f0</div>
<div><span style="white-space:pre-wrap"> </span>nm@078:<span style="white-space:pre-wrap"> </span>testb $2,%bl<span style="white-space:pre-wrap"> </span>! f6 c3 02 a.k.a. testb $2,rReceiver</div>
<div><span style="white-space:pre-wrap"> </span>nm@07b:<span style="white-space:pre-wrap"> </span>jz nm@085<span style="white-space:pre-wrap"> </span>! 74 08</div><div>
13 <10> push local 0</div><div>14 <C8> push 1; send +</div><div><span style="white-space:pre-wrap"> </span>nm@07d:<span style="white-space:pre-wrap"> </span>add $4,rReceiver<span style="white-space:pre-wrap"> </span>! 83 c3 04</div>
<div><span style="white-space:pre-wrap"> </span>nm@080:<span style="white-space:pre-wrap"> </span>jno nm@092<span style="white-space:pre-wrap"> </span>! 71 10</div><div>
<span style="white-space:pre-wrap"> </span>nm@082:<span style="white-space:pre-wrap"> </span>sub $4,rReceiver<span style="white-space:pre-wrap"> </span>! 83 eb 04</div>
<div><span style="white-space:pre-wrap"> </span>nm@085:<span style="white-space:pre-wrap"> </span>push $7<span style="white-space:pre-wrap"> </span>! 6a 07</div><div><span style="white-space:pre-wrap"> </span>nm@087:<span style="white-space:pre-wrap"> </span>pop rArg1<span style="white-space:pre-wrap"> </span>! 5e</div>
<div><span style="white-space:pre-wrap"> </span>nm@088:<span style="white-space:pre-wrap"> </span>mov $0x16bbab34,rClass<span style="white-space:pre-wrap"> </span>! ba 34 ab bb 16</div>
<div><span style="white-space:pre-wrap"> </span>nm@08d:<span style="white-space:pre-wrap"> </span>call _81080=send1Args<span style="white-space:pre-wrap"> </span>! e8 f6 e6 6b ea</div>
<div><span style="white-space:pre-wrap"> </span>map: 0x22: send1(2) #+</div><div>15 <4C> store local 0; pop</div><div><span style="white-space:pre-wrap"> </span>nm@092:<span style="white-space:pre-wrap"> </span>mov rReceiver,-0x10(bFrame)<span style="white-space:pre-wrap"> </span>! 89 5d f0</div>
<div>16 <E3 F6> jump 8</div><div><span style="white-space:pre-wrap"> </span>nm@095:<span style="white-space:pre-wrap"> </span>cmp 0xfb190,bSP<span style="white-space:pre-wrap"> </span>! 3b 25 90 b1 0f 00</div>
<div><span style="white-space:pre-wrap"> </span>nm@09b:<span style="white-space:pre-wrap"> </span>jae nm@040<span style="white-space:pre-wrap"> </span>! 0f 83 9f ff ff ff</div>
<div><span style="white-space:pre-wrap"> </span>nm@0a1:<span style="white-space:pre-wrap"> </span>call _80d30<span style="white-space:pre-wrap"> </span>! e8 92 e3 6b ea</div>
<div>vpc 18:</div><div><span style="white-space:pre-wrap"> </span>nm@0a6:<span style="white-space:pre-wrap"> </span>jmp nm@040<span style="white-space:pre-wrap"> </span>! e9 95 ff ff ff</div>
</div><div><br></div><div><br></div><div>Interesting. VW executes the following instructions around the loop:</div><div><br></div><div><div><span style="white-space:pre-wrap"> </span>nm@040:<span style="white-space:pre-wrap"> </span>mov -0x10(bFrame),rReceiver<span style="white-space:pre-wrap"> </span>! 8b 5d f0</div>
<div><span style="white-space:pre-wrap"> </span>nm@043:<span style="white-space:pre-wrap"> </span>mov $0x17d78403,rArg1<span style="white-space:pre-wrap"> </span>! be 03 84 d7 17</div>
<div><span style="white-space:pre-wrap"> </span>nm@048:<span style="white-space:pre-wrap"> </span>testb $2,%bl<span style="white-space:pre-wrap"> </span>! f6 c3 02 a.k.a. testb $2,rReceiver</div>
<div><span style="white-space:pre-wrap"> </span>nm@04b:<span style="white-space:pre-wrap"> </span>jz nm@056<span style="white-space:pre-wrap"> </span>! 74 09</div><div>
<span style="white-space:pre-wrap"> </span>nm@04d:<span style="white-space:pre-wrap"> </span>cmp rArg1,rReceiver<span style="white-space:pre-wrap"> </span>! 3b de</div>
<div><span style="white-space:pre-wrap"> </span>nm@04f:<span style="white-space:pre-wrap"> </span>jle nm@075<span style="white-space:pre-wrap"> </span>! 7e 24</div><div>
<span style="white-space:pre-wrap"> </span>nm@051:<span style="white-space:pre-wrap"> </span>jmp nm@0ab<span style="white-space:pre-wrap"> </span>! e9 55 00 00 00</div>
<div>...</div><div><span style="white-space:pre-wrap"> </span>nm@07d:<span style="white-space:pre-wrap"> </span>add $4,rReceiver<span style="white-space:pre-wrap"> </span>! 83 c3 04</div>
<div><span style="white-space:pre-wrap"> </span>nm@080:<span style="white-space:pre-wrap"> </span>jno nm@092<span style="white-space:pre-wrap"> </span>! 71 10</div><div>
...</div><div><span style="white-space:pre-wrap"> </span>nm@092:<span style="white-space:pre-wrap"> </span>mov rReceiver,-0x10(bFrame)<span style="white-space:pre-wrap"> </span>! 89 5d f0</div>
<div><span style="white-space:pre-wrap"> </span>nm@095:<span style="white-space:pre-wrap"> </span>cmp 0xfb190,bSP<span style="white-space:pre-wrap"> </span>! 3b 25 90 b1 0f 00</div>
<div><span style="white-space:pre-wrap"> </span>nm@09b:<span style="white-space:pre-wrap"> </span>jae nm@040<span style="white-space:pre-wrap"> </span>! 0f 83 9f ff ff ff</div>
</div><div><br></div><div>12 instructions, one read and one write. Cog executes many more, quite a few of them reads and writes. How I itch to find the time to do delayed code generation/stack-to-register mapping in Cog...</div>
<div><br></div><div>best,</div><div>Eliot</div><div><br></div><div class="gmail_quote">On Tue, Nov 16, 2010 at 3:32 PM, Levente Uzonyi <span dir="ltr"><<a href="mailto:leves@elte.hu" target="_blank">leves@elte.hu</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">On Tue, 16 Nov 2010, Juan Vuletich wrote:<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi Levente,<br>
<br>
Levente Uzonyi wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi,<br>
<br>
I'm mostly ready with this cleanup, only a few methods left in the image that use #== for integer comparison. Some of them must use #==, others may be changed, but I wasn't sure if the code will work if it's changed or not. If you'd like to check these methods, then evaluate the following in a workspace:<br>
<br>
... code here...<br>
<br>
Cheers,<br>
Levente<br>
</blockquote>
<br>
Thanks for the snippet! It is now a method in Cuis. I also checked a bit on trunk. In all senders of #nextObject you can apply the pattern in #allObjectsDo:. Instead of assuming == 0 for last #nextObject, it assumes that Object new will go at the end. If that ever changed, a few places would need fixing...<br>
</blockquote>
<br>
I'm not sure if it's ok to use that pattern in ImageSegment. Even if it's ok, the code is so messy, that it seems to be hard to apply the pattern without breaking the code.<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
As an experiment, in #allObjectsDo: I tried to remove the Object new stuff and replace with '[ 0 = object and: [ object isMemberOf: SmallInteger ]] whileFalse: ', but my test became extremely slow. I suspect #= is sending the message even if it has its own bytecode... So, the only method that remains with the '0 ==' pattern in Cuis is #critical:ifLocked: , as I'm not sure if we can clean it.<br>
</blockquote>
<br>
#= is a special selector, so I'm sure it won't send a message if it doesn't have to. It's possible that #= is a bit slower than #== because it has to do some extra checks, but the difference is probably neglible. Here's a benchmark:<br>
<br>
| offset equality identity |<br>
Smalltalk garbageCollect.<br>
offset := [ 1 to: 1000000 do: [ :i | ] ] timeToRun.<br>
equality := [ :x | [ 1 to: 1000000 do: [ :i | 0 = x ] ] timeToRun ].<br>
identity := [ :x | [ 1 to: 1000000 do: [ :i | 0 == x ] ] timeToRun ].<br>
{ 0. 1. SmallInteger maxVal. SmallInteger maxVal + 1. nil. Object new. 0.0. Array new. 1/2 } collect: [ :each |<br>
each -> ({<br>
equality value: each.<br>
identity value: each.<br>
equality value: each.<br>
identity value: each.<br>
equality value: each.<br>
identity value: each } - offset) ].<br>
<br>
And my results on CogVM:<br>
<br>
{<br>
0->#(3 2 3 1 3 2).<br>
1->#(2 4 3 0 3 4).<br>
1073741823->#(3 3 2 1 3 5).<br>
1073741824->#(221 4 223 1 223 4).<br>
nil->#(14 4 14 0 15 4).<br>
an Object->#(14 5 13 1 14 4).<br>
0.0->#(622 5 623 2 624 4).<br>
#()->#(16 5 14 1 16 4).<br>
(1/2)->#(260 4 259 1 258 5)<br>
}<br>
<br>
So the cause of the slowdown in your #allObjectsDo: implementation is that #= is only fast if both the receiver and the argument are SmallIntegers, otherwise there will be message sends.<br>
<br>
Another way to implement #allObjectsDo: without the marker object is to replace the loop condition with: object class == SmallInteger. It's actually the same as your implementation without the #= check. #isMemberOf: is optimized to non-real sends. This is as fast as the current implementation on my pc.<br>
<br>
<br>
Levente<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
Cheers,<br>
Juan Vuletich<br>
<br>
<br>
</blockquote>
<br>
</blockquote></div><br></div>
</blockquote></div><br></div>