I'd really like to run a bench mark before and after the commit in question. I spent a short time trying to run the benchmarks a while back but was unsuccessful getting it to run. Can you help me with this? Thanks. Blake On Wed, Oct 6, 2010 at 4:25 PM, Blake McBride <blake@mcbride.name> wrote:

That's helpful.  Thanks.  So now we need to do benchmark / runtime tests.

Thanks.

Blake

On Wed, Oct 6, 2010 at 4:17 PM, Erik Huelsmann <ehuels@gmail.com> wrote:

...

Blake,

Before going to bed, I did a quick test - as discussed over GMail chat - to see how much they differ in compilation times on other software.

timing new code, Maxima compilation: 223.416s, loading: 25.8 timing old code, Maxima compilation: 204.063s, loading: 29.174

I have no idea of the variation of the Maxima compilation times; it looks like the new code is 10% slower at compiling, but 20% more efficient at loading. However, these were single runs, so my conclusions may be way off, depending on the variations.

Bye,

Erik.

On Tue, Oct 5, 2010 at 5:52 PM, Blake McBride <blake@mcbride.name> wrote:

...

Greetings,

I hadn't built ABCL in a little while so I checked out the latest version today and built it.  It seemed to be significantly slower than before so I decided to investigate.  This is what I found.

In the past I could build ABCL in 2:43.  It now takes me 4:40.  That (IMO) represents a pretty significant change in build time.  I did a binary search and discovered that all of the change occurred at revision 12918 - Generic Class File Branch Merge.

In general, I could't care less about the build time unless it is indicative of a problem that could rear its head in my application. Where is that time being spent?  Is there a change in runtime? Loading? Compiling?

I'd be real interested in this.

Thanks.

Blake McBride

_______________________________________________ armedbear-devel mailing list armedbear-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/armedbear-devel

On Nov 11, 2010, at 11:22 , Erik Huelsmann wrote:

Hi Blake,

On Sat, Nov 6, 2010 at 7:11 PM, Blake McBride <blake@mcbride.name> wrote:

...

I'd really like to run a bench mark before and after the commit in question. I spent a short time trying to run the benchmarks a while back but was unsuccessful getting it to run. Can you help me with this?

Sure. What's your platform? If it's Linux/unix, getting the tests to run is relatively easy: make sure you have "make" and follow the instructions in the README file.

There is support in the ASDF system definition 'abcl.asd' for running the cl-bench tests as well. Simply install the cl-bench suite in a sibling directory to the ABCL source, then either 1) From the REPL CL-USER> (asdf:load-system :abcl) […] Cl-USER> (asdf:test-system :cl-bench) or 2) Use the ant target 'test.cl-bench' cmd$ ant test.cl-bench -- "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now."

Hi Blake, What's more, tonight, I hacked around a bit more. What I did was to disable the peephole optimizer for constructors (that's the "old" as in 0.22) behaviour. It reduced my build-time from 5m50s to roughly 5m20s. So, at least part of the added build time seems to come from "deteriorated" performance; the remainder would be coming from extra sources to be compiled or performance degradation. Since we compile structures pretty slowly and jvm-class-file.lisp contains quite a number of structure definitions, I tend to blame the former. BTW: Alessio is working on defining static class initializer methods in ABCL and moving those to trunk. I propose that we always peephole optimize all methods *except* the static initializer (it's being run only once; a few extra byte codes can't matter that much). The net effect of that is that I can throw away my current hack, because most of our constructors will be trivial super() calls. The real initialization code will move to the static initializer - which won't be optimized. Bye, Erik. On Sun, Nov 14, 2010 at 12:22 PM, Erik Huelsmann <ehuels@gmail.com> wrote:

Hi Blake, others,

Running our cl-bench tests on Windows finally succeeded (but they did need some tweaking).

Below the results. The reference is the revision before the merge, the comparison (0.24) is a trunk revision of the last days; it's not exactly the comparison you were talking about, but it does seem like an indication.

Here are the results:

Armed Bear Common Lisp 0.23.0-dev Java 1.6.0_20 Sun Microsystems Inc. Java HotSpot(TM) Client VM Low-level initialization completed in 0.289 seconds. Startup completed in 3.552 seconds. Benchmark                 Reference  Armed ------------------------------------------------------------------------------------- SUM-PERMUTATIONS         [      2.31]   1.02 BOYER                    [      3.66]   0.95 BROWSE                   [      3.19]   0.19 DDERIV                   [      0.99]   1.01 DERIV                    [      0.71]   1.02 DESTRUCTIVE              [      1.03]   0.99 DIV2-TEST-1              [      0.51]   1.03 DIV2-TEST-2              [      2.02]   0.99 FFT                      [      0.42]   1.12 FRPOLY/FIXNUM            [      1.56]   1.03 FRPOLY/BIGNUM            [      0.73]   1.02 FRPOLY/FLOAT             [      1.53]   1.02 PUZZLE                   [      4.67]   1.02 TAK                      [      7.70]   0.98 CTAK                     [     20.68]   1.01 TRTAK                    [      7.70]    1.0 TAKL                     [      5.94]   0.97 STAK                     [     10.92]   0.98 FPRINT/UGLY              [      2.69]   1.18 FPRINT/PRETTY            [     33.24]   1.01 TRAVERSE                 [     20.80]   1.05 TRIANGLE                 [     12.32]   1.06 RICHARDS                 [     23.68]   1.04 FACTORIAL                [      0.42]    1.0 FIB                      [      0.89]    1.0 FIB-RATIO                [      0.23]   1.12 ACKERMANN                [     31.31]   0.93 MANDELBROT/COMPLEX       [      0.57]   1.48 MANDELBROT/DFLOAT        [      0.07]   1.41 MRG32K3A                 [      1.90]   1.11 CRC40                    [     19.55]   1.05 BIGNUM/ELEM-100-1000     [      2.78]   1.14 BIGNUM/ELEM-1000-100     [      0.77]   1.13 BIGNUM/ELEM-10000-1      [      1.05]   1.21 BIGNUM/PARI-100-10       [      0.05]   0.98 BIGNUM/PARI-200-5        [      0.15]   0.99 PI-DECIMAL/SMALL         [     40.27]   1.07 PI-DECIMAL/BIG           [     87.37]   1.06 PI-ATAN                  [      1.58]   1.04 PI-RATIOS                [      4.41]   1.04 HASH-STRINGS             [      1.72]   1.08 HASH-INTEGERS            [      1.38]   1.26 SLURP-LINES              [      0.00]   0.67 BOEHM-GC                 [      9.29]   1.16 DEFLATE-FILE             [     12.08]   1.02 1D-ARRAYS                [      2.66]   1.17 2D-ARRAYS                [     13.33]   1.07 3D-ARRAYS                [     36.69]   1.06 BITVECTORS               [      3.49]   1.01 BENCH-STRINGS            [     21.73]   0.97 fill-strings/adjustable  [     11.19]   1.07 STRING-CONCAT            [    143.53]   0.95 SEARCH-SEQUENCE          [      2.28]    1.0 CLOS/defclass            [       1.0]   1.31 CLOS/defmethod           [      0.87]   1.37 CLOS/instantiate         [     42.83]   1.02 CLOS/simple-instantiate  [    143.36]   1.06 CLOS/methodcalls         [      8.87]   1.01 CLOS/method+after        [      5.36]   1.05 CLOS/complex-methods     [      3.55]   1.19 EQL-SPECIALIZED-FIB      [      1.40]   0.90 Reference time in first column is in seconds; other columns are relative Reference implementation: Armed Bear Common Lisp 0.23.0-dev Impl Armed: Armed Bear Common Lisp 0.24.0-dev === Test machine ===   Machine-type: X86   Machine-version: NIL

Bye,

Erik.

On Thu, Nov 11, 2010 at 11:22 AM, Erik Huelsmann <ehuels@gmail.com> wrote:

...

Hi Blake,

On Sat, Nov 6, 2010 at 7:11 PM, Blake McBride <blake@mcbride.name> wrote:

...

I'd really like to run a bench mark before and after the commit in question.  I spent a short time trying to run the benchmarks a while back but was unsuccessful getting it to run.  Can you help me with this?

Sure. What's your platform? If it's Linux/unix, getting the tests to run is relatively easy: make sure  you have "make" and follow the instructions in the README file.

When you have Windows, it's a bit harder. This is what I do to run the tests on my windows machine:

<in the cl-bench root directory> <open support.lisp> <search #+win32, replace with #+(or win32 windows)> <search #-win32, replace with #-(or win32 windows)> <save, close> cd files copy *.lisp *.olisp cd .. abcl :ld do-compilation-script :exit abcl :ld do-execute-script

I hope the above works for you!

Bye,

Erik.

...

Thanks.

Blake

On Wed, Oct 6, 2010 at 4:25 PM, Blake McBride <blake@mcbride.name> wrote:

...

That's helpful.  Thanks.  So now we need to do benchmark / runtime tests.

Thanks.

Blake

On Wed, Oct 6, 2010 at 4:17 PM, Erik Huelsmann <ehuels@gmail.com> wrote:

...

Blake,

Before going to bed, I did a quick test - as discussed over GMail chat - to see how much they differ in compilation times on other software.

timing new code, Maxima compilation: 223.416s, loading: 25.8 timing old code, Maxima compilation: 204.063s, loading: 29.174

I have no idea of the variation of the Maxima compilation times; it looks like the new code is 10% slower at compiling, but 20% more efficient at loading. However, these were single runs, so my conclusions may be way off, depending on the variations.

Bye,

Erik.

On Tue, Oct 5, 2010 at 5:52 PM, Blake McBride <blake@mcbride.name> wrote:

...

Greetings,

I hadn't built ABCL in a little while so I checked out the latest version today and built it.  It seemed to be significantly slower than before so I decided to investigate.  This is what I found.

In the past I could build ABCL in 2:43.  It now takes me 4:40.  That (IMO) represents a pretty significant change in build time.  I did a binary search and discovered that all of the change occurred at revision 12918 - Generic Class File Branch Merge.

In general, I could't care less about the build time unless it is indicative of a problem that could rear its head in my application. Where is that time being spent?  Is there a change in runtime? Loading? Compiling?

I'd be real interested in this.

Thanks.

Blake McBride

_______________________________________________ armedbear-devel mailing list armedbear-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/armedbear-devel

Dear Erik, Thanks a lot. If I'm reading it correctly it definitely answers my question. The way I am reading it is as follows: The numbers in the brackets represent actual times for the revision prior to revision 12918. The numbers outside the brackets represent numbers relative to the numbers in the brackets for the latest trunk branch. Therefor 1.0 outside the brackets means it took the same amount of time is the number indicated in the brackets. 2.0 would mean it took twice as long, 0.5 would meant it took have as much time, and so on. Is that correct? If m interpretation is correct then there clearly isn't a runtime issue. This puts my questions to rest. Thank you very much! Blake McBride On Sun, Nov 14, 2010 at 5:22 AM, Erik Huelsmann <ehuels@gmail.com> wrote:

Hi Blake, others,

Running our cl-bench tests on Windows finally succeeded (but they did need some tweaking).

Below the results. The reference is the revision before the merge, the comparison (0.24) is a trunk revision of the last days; it's not exactly the comparison you were talking about, but it does seem like an indication.

Here are the results:

Armed Bear Common Lisp 0.23.0-dev Java 1.6.0_20 Sun Microsystems Inc. Java HotSpot(TM) Client VM Low-level initialization completed in 0.289 seconds. Startup completed in 3.552 seconds. Benchmark                 Reference  Armed ------------------------------------------------------------------------------------- SUM-PERMUTATIONS         [      2.31]   1.02 BOYER                    [      3.66]   0.95 BROWSE                   [      3.19]   0.19 DDERIV                   [      0.99]   1.01 DERIV                    [      0.71]   1.02 DESTRUCTIVE              [      1.03]   0.99 DIV2-TEST-1              [      0.51]   1.03 DIV2-TEST-2              [      2.02]   0.99 FFT                      [      0.42]   1.12 FRPOLY/FIXNUM            [      1.56]   1.03 FRPOLY/BIGNUM            [      0.73]   1.02 FRPOLY/FLOAT             [      1.53]   1.02 PUZZLE                   [      4.67]   1.02 TAK                      [      7.70]   0.98 CTAK                     [     20.68]   1.01 TRTAK                    [      7.70]    1.0 TAKL                     [      5.94]   0.97 STAK                     [     10.92]   0.98 FPRINT/UGLY              [      2.69]   1.18 FPRINT/PRETTY            [     33.24]   1.01 TRAVERSE                 [     20.80]   1.05 TRIANGLE                 [     12.32]   1.06 RICHARDS                 [     23.68]   1.04 FACTORIAL                [      0.42]    1.0 FIB                      [      0.89]    1.0 FIB-RATIO                [      0.23]   1.12 ACKERMANN                [     31.31]   0.93 MANDELBROT/COMPLEX       [      0.57]   1.48 MANDELBROT/DFLOAT        [      0.07]   1.41 MRG32K3A                 [      1.90]   1.11 CRC40                    [     19.55]   1.05 BIGNUM/ELEM-100-1000     [      2.78]   1.14 BIGNUM/ELEM-1000-100     [      0.77]   1.13 BIGNUM/ELEM-10000-1      [      1.05]   1.21 BIGNUM/PARI-100-10       [      0.05]   0.98 BIGNUM/PARI-200-5        [      0.15]   0.99 PI-DECIMAL/SMALL         [     40.27]   1.07 PI-DECIMAL/BIG           [     87.37]   1.06 PI-ATAN                  [      1.58]   1.04 PI-RATIOS                [      4.41]   1.04 HASH-STRINGS             [      1.72]   1.08 HASH-INTEGERS            [      1.38]   1.26 SLURP-LINES              [      0.00]   0.67 BOEHM-GC                 [      9.29]   1.16 DEFLATE-FILE             [     12.08]   1.02 1D-ARRAYS                [      2.66]   1.17 2D-ARRAYS                [     13.33]   1.07 3D-ARRAYS                [     36.69]   1.06 BITVECTORS               [      3.49]   1.01 BENCH-STRINGS            [     21.73]   0.97 fill-strings/adjustable  [     11.19]   1.07 STRING-CONCAT            [    143.53]   0.95 SEARCH-SEQUENCE          [      2.28]    1.0 CLOS/defclass            [       1.0]   1.31 CLOS/defmethod           [      0.87]   1.37 CLOS/instantiate         [     42.83]   1.02 CLOS/simple-instantiate  [    143.36]   1.06 CLOS/methodcalls         [      8.87]   1.01 CLOS/method+after        [      5.36]   1.05 CLOS/complex-methods     [      3.55]   1.19 EQL-SPECIALIZED-FIB      [      1.40]   0.90 Reference time in first column is in seconds; other columns are relative Reference implementation: Armed Bear Common Lisp 0.23.0-dev Impl Armed: Armed Bear Common Lisp 0.24.0-dev === Test machine ===   Machine-type: X86   Machine-version: NIL

Bye,

Erik.

On Thu, Nov 11, 2010 at 11:22 AM, Erik Huelsmann <ehuels@gmail.com> wrote:

...

Hi Blake,

On Sat, Nov 6, 2010 at 7:11 PM, Blake McBride <blake@mcbride.name> wrote:

...

I'd really like to run a bench mark before and after the commit in question.  I spent a short time trying to run the benchmarks a while back but was unsuccessful getting it to run.  Can you help me with this?

Sure. What's your platform? If it's Linux/unix, getting the tests to run is relatively easy: make sure  you have "make" and follow the instructions in the README file.

When you have Windows, it's a bit harder. This is what I do to run the tests on my windows machine:

<in the cl-bench root directory> <open support.lisp> <search #+win32, replace with #+(or win32 windows)> <search #-win32, replace with #-(or win32 windows)> <save, close> cd files copy *.lisp *.olisp cd .. abcl :ld do-compilation-script :exit abcl :ld do-execute-script

I hope the above works for you!

Bye,

Erik.

...

Thanks.

Blake

On Wed, Oct 6, 2010 at 4:25 PM, Blake McBride <blake@mcbride.name> wrote:

...

That's helpful.  Thanks.  So now we need to do benchmark / runtime tests.

Thanks.

Blake

On Wed, Oct 6, 2010 at 4:17 PM, Erik Huelsmann <ehuels@gmail.com> wrote:

...

Blake,

Before going to bed, I did a quick test - as discussed over GMail chat - to see how much they differ in compilation times on other software.

timing new code, Maxima compilation: 223.416s, loading: 25.8 timing old code, Maxima compilation: 204.063s, loading: 29.174

I have no idea of the variation of the Maxima compilation times; it looks like the new code is 10% slower at compiling, but 20% more efficient at loading. However, these were single runs, so my conclusions may be way off, depending on the variations.

Bye,

Erik.

On Tue, Oct 5, 2010 at 5:52 PM, Blake McBride <blake@mcbride.name> wrote:

...

Greetings,

I hadn't built ABCL in a little while so I checked out the latest version today and built it.  It seemed to be significantly slower than before so I decided to investigate.  This is what I found.

In the past I could build ABCL in 2:43.  It now takes me 4:40.  That (IMO) represents a pretty significant change in build time.  I did a binary search and discovered that all of the change occurred at revision 12918 - Generic Class File Branch Merge.

In general, I could't care less about the build time unless it is indicative of a problem that could rear its head in my application. Where is that time being spent?  Is there a change in runtime? Loading? Compiling?

I'd be real interested in this.

Thanks.

Blake McBride

_______________________________________________ armedbear-devel mailing list armedbear-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/armedbear-devel