Now scheme min is even faster.. that is ok. But what I don't understand is how that simple C implementation get so low performance.
abs: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) 1 ms real time 0 ms cpu time (0 user, 0 system) no collections no bytes allocated no minor faults no major faults abs-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j)))))) 18 ms real time 17 ms cpu time (17 user, 0 system) no collections no bytes allocated 1 minor fault no major faults min: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) 0 ms real time 0 ms cpu time (0 user, 0 system) no collections no bytes allocated no minor faults no major faults min-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))) 26 ms real time 26 ms cpu time (23 user, 3 system) no collections no bytes allocated no minor faults no major faults
On Sun, Nov 4, 2012 at 7:31 PM, Mikael More mikael.more@gmail.com wrote:
Ah, also throw in the following:
(declare (not interrupts-enabled))
This might be a big deal - when it's off, Gambit doesn't produce any
hooks for stack overflow or thread system interrupts but executes your code according to a typical C model
- (declare (standard-bindings) (extended-bindings)) too?
What is the performance you get from Gambit now, and what kind of performance is it you want to get?
2012/11/4 tcl super dude tcleval@gmail.com
I assumed I could get a bit of performance improvement over min procedure because I don't do any branch. I' thought min-integer to be at least close to gambit-c time. The problem is that those operations and a few more are performed thousands of times. Gambit loops are usually 2x slower than C, so I hoped I could improve some basic operations as I only perform them on integers. I don't Understand how a simple shift to extract the lower bits could make so much difference. Maybe if someone point me to the implementation of min and abs on gambit, I can understand the difference. I ve looked for these procedures on gambit's source but could not point it out.
I'll try to implement them using vectors and see what happens.
now compiled with (block):
abs:
(time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) 2 ms real time
3 ms cpu time (3 user, 0 system) no collections no bytes allocated no minor faults no major faultsabs-integer:
(time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j)))))) 19 ms real time
20 ms cpu time (17 user, 3 system) no collections no bytes allocated 2 minor faults no major faultsmin:
(time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) 2 ms real time
3 ms cpu time (0 user, 3 system) no collections no bytes allocated no minor faults no major faultsmin-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))) 27 ms real time 27 ms cpu time (27 user, 0 system) no collections no bytes allocated no minor faults no major faults
On Sun, Nov 4, 2012 at 5:15 PM, Mikael mikael.rcv@gmail.com wrote:
First, please add (declare (block)) and re-compile, this way Gambit won't need to re-look up the procedures in the global namespace that you call, on each invocation. Please report what kind of improvement you got with this.
Now just to make this round of emails complete:
What would be good/very good performance in your view?
Could you send jobs in some kind of batch format to C?
Note that Gambit-internal fxnum variables have a special encoding (the lowermost two(?) bits are object type identifier, which are shifted out when converting them to C int), this takes a little bit of time.
If for instance you make u32vectors/s32vectors/etc. [and send batch ops made on those to C to be performed etc.] then you can just read out the content of those vectors 'raw' in C, there's no interconversion there.
Brgds
2012/11/4 tcl super dude tcleval@gmail.com
Hi everyone,
I have been working on an image processing library that I want to use on android. I know that performance is a bit of a problem on image processing, so I started to study some bit hacking to improve the performance on my integer operations. My problem now is that I don't know why my c-lambdas take so much time to execute. All operations take constant time and there is no branching. Probably is some sort of conversion time between C types and GAMBIT types. I expected min-integer and abs-integer to be a lot faster than min and abs, respectively. Any help is appreciated.
URL REFERENCE: http://graphics.stanford.edu/~seander/bithacks.html#IntegerAbs http://graphics.stanford.edu/~seander/bithacks.html#IntegerMinOrMax
MY ENVIRONMENT: Linux casa 3.6.4-1-ARCH #1 SMP PREEMPT Mon Oct 29 15:13:56 CET 2012 i686 GNU/Linux
gcc -v Using built-in specs. COLLECT_GCC=gcc COLLECT_LTO_WRAPPER=/usr/lib/gcc/i686-pc-linux-gnu/4.7.2/lto-wrapper Target: i686-pc-linux-gnu Configured with: /build/src/gcc-4.7.2/configure --prefix=/usr --libdir=/usr/lib --libexecdir=/usr/lib --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=https://bugs.archlinux.org/--enable-languages=c,c++,ada,fortran,go,lto,objc,... --enable-shared --enable-threads=posix --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-clocale=gnu --disable-libstdcxx-pch --enable-libstdcxx-time --enable-gnu-unique-object --enable-linker-build-id --with-ppl --enable-cloog-backend=isl --disable-ppl-version-check --disable-cloog-version-check --enable-lto --enable-gold --enable-ld=default --enable-plugin --with-plugin-ld=ld.gold --with-linker-hash-style=gnu --disable-multilib --disable-libssp --disable-build-with-cxx --disable-build-poststage1-with-cxx --enable-checking=release Thread model: posix gcc version 4.7.2 (GCC)
Gambit v4.6.6 (SINGLE HOST and GCC extensions enabled)
SCHEME CODE: compiled with ---> gambitc -exe bithacks.scm
;; bithacks.scm
(declare (not safe) (fixnum) (inline) (inline-primitives))
;; INTEGER ABSOLUTE (define abs-integer (c-lambda (int) int "// we want to find the absolute value of arg1 int const mask = ___arg1 >> sizeof(int) * CHAR_BIT - 1; ___result = (___arg1 + mask) ^ mask;"))
(define min-integer (c-lambda (int int) int "// find the smaller of arg1 arg2 ___result = ___arg2 ^ ((___arg1 ^ ___arg2) & -(___arg1 < ___arg2));"))
(define (test-abs) (let ((result 0)) (begin (display "abs:\n") (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) (display "abs-integer:\n") (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j)))))))))
(define (test-min) (let ((result 0)) (begin (display "min:\n") (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) (display "min-integer:\n") (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))))))
;; run tests
(test-abs) (test-min)
;;;;;;;;;;;;;;;;;;;;;; RESULTS;;;;;;;;;;;;;;;;;;;;;
abs: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) 2 ms real time 3 ms cpu time (3 user, 0 system) no collections no bytes allocated no minor faults no major faults abs-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j)))))) 22 ms real time 23 ms cpu time (23 user, 0 system) no collections no bytes allocated 2 minor faults no major faults min: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) 2 ms real time 3 ms cpu time (0 user, 3 system) no collections no bytes allocated no minor faults no major faults min-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))) 28 ms real time 30 ms cpu time (27 user, 3 system) no collections no bytes allocated no minor faults no major faults
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
Afficher les réponses par date
What about you catch a copy of the C code post- macro expansion, and a copy of the assembly output too, and have a look?
Let the ml know what you found.
2012/11/5 tcl super dude tcleval@gmail.com
Now scheme min is even faster.. that is ok. But what I don't understand is how that simple C implementation get so low performance.
abs: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) 1 ms real time 0 ms cpu time (0 user, 0 system) no collections no bytes allocated no minor faults no major faults abs-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j)))))) 18 ms real time 17 ms cpu time (17 user, 0 system) no collections no bytes allocated 1 minor fault no major faults min: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) 0 ms real time 0 ms cpu time (0 user, 0 system) no collections no bytes allocated no minor faults no major faults min-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))) 26 ms real time 26 ms cpu time (23 user, 3 system) no collections no bytes allocated no minor faults no major faults
On Sun, Nov 4, 2012 at 7:31 PM, Mikael More mikael.more@gmail.com wrote:
Ah, also throw in the following:
(declare (not interrupts-enabled))
This might be a big deal - when it's off, Gambit doesn't produce any
hooks for stack overflow or thread system interrupts but executes your code according to a typical C model
- (declare (standard-bindings) (extended-bindings)) too?
What is the performance you get from Gambit now, and what kind of performance is it you want to get?
2012/11/4 tcl super dude tcleval@gmail.com
I assumed I could get a bit of performance improvement over min procedure because I don't do any branch. I' thought min-integer to be at least close to gambit-c time. The problem is that those operations and a few more are performed thousands of times. Gambit loops are usually 2x slower than C, so I hoped I could improve some basic operations as I only perform them on integers. I don't Understand how a simple shift to extract the lower bits could make so much difference. Maybe if someone point me to the implementation of min and abs on gambit, I can understand the difference. I ve looked for these procedures on gambit's source but could not point it out.
I'll try to implement them using vectors and see what happens.
now compiled with (block):
abs:
(time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) 2 ms real time
3 ms cpu time (3 user, 0 system) no collections no bytes allocated no minor faults no major faultsabs-integer:
(time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j))))))
19 ms real time 20 ms cpu time (17 user, 3 system) no collections no bytes allocated 2 minor faults no major faultsmin:
(time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) 2 ms real time
3 ms cpu time (0 user, 3 system) no collections no bytes allocated no minor faults no major faultsmin-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))) 27 ms real time 27 ms cpu time (27 user, 0 system) no collections no bytes allocated no minor faults no major faults
On Sun, Nov 4, 2012 at 5:15 PM, Mikael mikael.rcv@gmail.com wrote:
First, please add (declare (block)) and re-compile, this way Gambit won't need to re-look up the procedures in the global namespace that you call, on each invocation. Please report what kind of improvement you got with this.
Now just to make this round of emails complete:
What would be good/very good performance in your view?
Could you send jobs in some kind of batch format to C?
Note that Gambit-internal fxnum variables have a special encoding (the lowermost two(?) bits are object type identifier, which are shifted out when converting them to C int), this takes a little bit of time.
If for instance you make u32vectors/s32vectors/etc. [and send batch ops made on those to C to be performed etc.] then you can just read out the content of those vectors 'raw' in C, there's no interconversion there.
Brgds
2012/11/4 tcl super dude tcleval@gmail.com
Hi everyone,
I have been working on an image processing library that I want to use on android. I know that performance is a bit of a problem on image processing, so I started to study some bit hacking to improve the performance on my integer operations. My problem now is that I don't know why my c-lambdas take so much time to execute. All operations take constant time and there is no branching. Probably is some sort of conversion time between C types and GAMBIT types. I expected min-integer and abs-integer to be a lot faster than min and abs, respectively. Any help is appreciated.
URL REFERENCE: http://graphics.stanford.edu/~seander/bithacks.html#IntegerAbs http://graphics.stanford.edu/~seander/bithacks.html#IntegerMinOrMax
MY ENVIRONMENT: Linux casa 3.6.4-1-ARCH #1 SMP PREEMPT Mon Oct 29 15:13:56 CET 2012 i686 GNU/Linux
gcc -v Using built-in specs. COLLECT_GCC=gcc COLLECT_LTO_WRAPPER=/usr/lib/gcc/i686-pc-linux-gnu/4.7.2/lto-wrapper Target: i686-pc-linux-gnu Configured with: /build/src/gcc-4.7.2/configure --prefix=/usr --libdir=/usr/lib --libexecdir=/usr/lib --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=https://bugs.archlinux.org/--enable-languages=c,c++,ada,fortran,go,lto,objc,... --enable-shared --enable-threads=posix --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-clocale=gnu --disable-libstdcxx-pch --enable-libstdcxx-time --enable-gnu-unique-object --enable-linker-build-id --with-ppl --enable-cloog-backend=isl --disable-ppl-version-check --disable-cloog-version-check --enable-lto --enable-gold --enable-ld=default --enable-plugin --with-plugin-ld=ld.gold --with-linker-hash-style=gnu --disable-multilib --disable-libssp --disable-build-with-cxx --disable-build-poststage1-with-cxx --enable-checking=release Thread model: posix gcc version 4.7.2 (GCC)
Gambit v4.6.6 (SINGLE HOST and GCC extensions enabled)
SCHEME CODE: compiled with ---> gambitc -exe bithacks.scm
;; bithacks.scm
(declare (not safe) (fixnum) (inline) (inline-primitives))
;; INTEGER ABSOLUTE (define abs-integer (c-lambda (int) int "// we want to find the absolute value of arg1 int const mask = ___arg1 >> sizeof(int) * CHAR_BIT - 1; ___result = (___arg1 + mask) ^ mask;"))
(define min-integer (c-lambda (int int) int "// find the smaller of arg1 arg2 ___result = ___arg2 ^ ((___arg1 ^ ___arg2) & -(___arg1 < ___arg2));"))
(define (test-abs) (let ((result 0)) (begin (display "abs:\n") (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) (display "abs-integer:\n") (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j)))))))))
(define (test-min) (let ((result 0)) (begin (display "min:\n") (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) (display "min-integer:\n") (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))))))
;; run tests
(test-abs) (test-min)
;;;;;;;;;;;;;;;;;;;;;; RESULTS;;;;;;;;;;;;;;;;;;;;;
abs: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) 2 ms real time 3 ms cpu time (3 user, 0 system) no collections no bytes allocated no minor faults no major faults abs-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j)))))) 22 ms real time 23 ms cpu time (23 user, 0 system) no collections no bytes allocated 2 minor faults no major faults min: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) 2 ms real time 3 ms cpu time (0 user, 3 system) no collections no bytes allocated no minor faults no major faults min-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))) 28 ms real time 30 ms cpu time (27 user, 3 system) no collections no bytes allocated no minor faults no major faults
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
here I have the expansion, assembly, C, and scheme
On Sun, Nov 4, 2012 at 8:01 PM, Mikael mikael.rcv@gmail.com wrote:
What about you catch a copy of the C code post- macro expansion, and a copy of the assembly output too, and have a look?
Let the ml know what you found.
2012/11/5 tcl super dude tcleval@gmail.com
Now scheme min is even faster.. that is ok. But what I don't understand is how that simple C implementation get so low performance.
abs: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) 1 ms real time 0 ms cpu time (0 user, 0 system) no collections no bytes allocated no minor faults no major faults abs-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j)))))) 18 ms real time 17 ms cpu time (17 user, 0 system) no collections no bytes allocated 1 minor fault no major faults min: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) 0 ms real time 0 ms cpu time (0 user, 0 system) no collections no bytes allocated no minor faults no major faults min-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))) 26 ms real time 26 ms cpu time (23 user, 3 system) no collections no bytes allocated no minor faults no major faults
On Sun, Nov 4, 2012 at 7:31 PM, Mikael More mikael.more@gmail.comwrote:
Ah, also throw in the following:
(declare (not interrupts-enabled))
This might be a big deal - when it's off, Gambit doesn't produce any
hooks for stack overflow or thread system interrupts but executes your code according to a typical C model
- (declare (standard-bindings) (extended-bindings)) too?
What is the performance you get from Gambit now, and what kind of performance is it you want to get?
2012/11/4 tcl super dude tcleval@gmail.com
I assumed I could get a bit of performance improvement over min procedure because I don't do any branch. I' thought min-integer to be at least close to gambit-c time. The problem is that those operations and a few more are performed thousands of times. Gambit loops are usually 2x slower than C, so I hoped I could improve some basic operations as I only perform them on integers. I don't Understand how a simple shift to extract the lower bits could make so much difference. Maybe if someone point me to the implementation of min and abs on gambit, I can understand the difference. I ve looked for these procedures on gambit's source but could not point it out.
I'll try to implement them using vectors and see what happens.
now compiled with (block):
abs:
(time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) 2 ms real time
3 ms cpu time (3 user, 0 system) no collections no bytes allocated no minor faults no major faultsabs-integer:
(time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j))))))
19 ms real time 20 ms cpu time (17 user, 3 system) no collections no bytes allocated 2 minor faults no major faultsmin:
(time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) 2 ms real time
3 ms cpu time (0 user, 3 system) no collections no bytes allocated no minor faults no major faultsmin-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))) 27 ms real time 27 ms cpu time (27 user, 0 system) no collections no bytes allocated no minor faults no major faults
On Sun, Nov 4, 2012 at 5:15 PM, Mikael mikael.rcv@gmail.com wrote:
First, please add (declare (block)) and re-compile, this way Gambit won't need to re-look up the procedures in the global namespace that you call, on each invocation. Please report what kind of improvement you got with this.
Now just to make this round of emails complete:
What would be good/very good performance in your view?
Could you send jobs in some kind of batch format to C?
Note that Gambit-internal fxnum variables have a special encoding (the lowermost two(?) bits are object type identifier, which are shifted out when converting them to C int), this takes a little bit of time.
If for instance you make u32vectors/s32vectors/etc. [and send batch ops made on those to C to be performed etc.] then you can just read out the content of those vectors 'raw' in C, there's no interconversion there.
Brgds
2012/11/4 tcl super dude tcleval@gmail.com
Hi everyone,
I have been working on an image processing library that I want to use on android. I know that performance is a bit of a problem on image processing, so I started to study some bit hacking to improve the performance on my integer operations. My problem now is that I don't know why my c-lambdas take so much time to execute. All operations take constant time and there is no branching. Probably is some sort of conversion time between C types and GAMBIT types. I expected min-integer and abs-integer to be a lot faster than min and abs, respectively. Any help is appreciated.
URL REFERENCE: http://graphics.stanford.edu/~seander/bithacks.html#IntegerAbs http://graphics.stanford.edu/~seander/bithacks.html#IntegerMinOrMax
MY ENVIRONMENT: Linux casa 3.6.4-1-ARCH #1 SMP PREEMPT Mon Oct 29 15:13:56 CET 2012 i686 GNU/Linux
gcc -v Using built-in specs. COLLECT_GCC=gcc COLLECT_LTO_WRAPPER=/usr/lib/gcc/i686-pc-linux-gnu/4.7.2/lto-wrapper Target: i686-pc-linux-gnu Configured with: /build/src/gcc-4.7.2/configure --prefix=/usr --libdir=/usr/lib --libexecdir=/usr/lib --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=https://bugs.archlinux.org/--enable-languages=c,c++,ada,fortran,go,lto,objc,... --enable-shared --enable-threads=posix --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-clocale=gnu --disable-libstdcxx-pch --enable-libstdcxx-time --enable-gnu-unique-object --enable-linker-build-id --with-ppl --enable-cloog-backend=isl --disable-ppl-version-check --disable-cloog-version-check --enable-lto --enable-gold --enable-ld=default --enable-plugin --with-plugin-ld=ld.gold --with-linker-hash-style=gnu --disable-multilib --disable-libssp --disable-build-with-cxx --disable-build-poststage1-with-cxx --enable-checking=release Thread model: posix gcc version 4.7.2 (GCC)
Gambit v4.6.6 (SINGLE HOST and GCC extensions enabled)
SCHEME CODE: compiled with ---> gambitc -exe bithacks.scm
;; bithacks.scm
(declare (not safe) (fixnum) (inline) (inline-primitives))
;; INTEGER ABSOLUTE (define abs-integer (c-lambda (int) int "// we want to find the absolute value of arg1 int const mask = ___arg1 >> sizeof(int) * CHAR_BIT - 1; ___result = (___arg1 + mask) ^ mask;"))
(define min-integer (c-lambda (int int) int "// find the smaller of arg1 arg2 ___result = ___arg2 ^ ((___arg1 ^ ___arg2) & -(___arg1 < ___arg2));"))
(define (test-abs) (let ((result 0)) (begin (display "abs:\n") (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) (display "abs-integer:\n") (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j)))))))))
(define (test-min) (let ((result 0)) (begin (display "min:\n") (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) (display "min-integer:\n") (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))))))
;; run tests
(test-abs) (test-min)
;;;;;;;;;;;;;;;;;;;;;; RESULTS;;;;;;;;;;;;;;;;;;;;;
abs: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) 2 ms real time 3 ms cpu time (3 user, 0 system) no collections no bytes allocated no minor faults no major faults abs-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j)))))) 22 ms real time 23 ms cpu time (23 user, 0 system) no collections no bytes allocated 2 minor faults no major faults min: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) 2 ms real time 3 ms cpu time (0 user, 3 system) no collections no bytes allocated no minor faults no major faults min-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))) 28 ms real time 30 ms cpu time (27 user, 3 system) no collections no bytes allocated no minor faults no major faults
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
As on https://github.com/feeley/gambit/blob/master/include/gambit.h.in line 2615 I guess 'min' procedure is just a C macro ....
... /* APPLY-able operations */
#define ___FIX_0 ___FIX(0) #define ___FIXPOS(x)((___WORD)(x)) #define ___FIXMAX(x,y)(((x)<(y))?(y):(x)) #define ___FIXMIN(x,y)(((x)<(y))?(x):(y)) <----------- the line in question #define ___FIXADD(x,y)((___WORD)((x)+(y)))
...
this way there is no call frame. Maybe later Marc can put some light to the matter.
obs: I didn't know if I could call you Marc or Mr. Feely, but given the informal nature of the mailing-list I prefer call people by their first name/nick, I hope you don't mind.
On Sun, Nov 4, 2012 at 9:12 PM, tcl super dude tcleval@gmail.com wrote:
here I have the expansion, assembly, C, and scheme
On Sun, Nov 4, 2012 at 8:01 PM, Mikael mikael.rcv@gmail.com wrote:
What about you catch a copy of the C code post- macro expansion, and a copy of the assembly output too, and have a look?
Let the ml know what you found.
2012/11/5 tcl super dude tcleval@gmail.com
Now scheme min is even faster.. that is ok. But what I don't understand is how that simple C implementation get so low performance.
abs: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) 1 ms real time 0 ms cpu time (0 user, 0 system) no collections no bytes allocated no minor faults no major faults abs-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j)))))) 18 ms real time 17 ms cpu time (17 user, 0 system) no collections no bytes allocated 1 minor fault no major faults min: (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) 0 ms real time 0 ms cpu time (0 user, 0 system) no collections no bytes allocated no minor faults no major faults min-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))) 26 ms real time 26 ms cpu time (23 user, 3 system) no collections no bytes allocated no minor faults no major faults
On Sun, Nov 4, 2012 at 7:31 PM, Mikael More mikael.more@gmail.comwrote:
Ah, also throw in the following:
(declare (not interrupts-enabled))
This might be a big deal - when it's off, Gambit doesn't produce any
hooks for stack overflow or thread system interrupts but executes your code according to a typical C model
- (declare (standard-bindings) (extended-bindings)) too?
What is the performance you get from Gambit now, and what kind of performance is it you want to get?
2012/11/4 tcl super dude tcleval@gmail.com
I assumed I could get a bit of performance improvement over min procedure because I don't do any branch. I' thought min-integer to be at least close to gambit-c time. The problem is that those operations and a few more are performed thousands of times. Gambit loops are usually 2x slower than C, so I hoped I could improve some basic operations as I only perform them on integers. I don't Understand how a simple shift to extract the lower bits could make so much difference. Maybe if someone point me to the implementation of min and abs on gambit, I can understand the difference. I ve looked for these procedures on gambit's source but could not point it out.
I'll try to implement them using vectors and see what happens.
now compiled with (block):
abs:
(time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs i)) (loop-i (fx+ 1 i)))))) 2 ms real time
3 ms cpu time (3 user, 0 system) no collections no bytes allocated no minor faults no major faultsabs-integer:
(time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (abs-integer j)) (loop-j (fx+ 1 j))))))
19 ms real time 20 ms cpu time (17 user, 3 system) no collections no bytes allocated 2 minor faults no major faultsmin:
(time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i result)) (loop-i (fx+ 1 i)))))) 2 ms real time
3 ms cpu time (0 user, 3 system) no collections no bytes allocated no minor faults no major faultsmin-integer: (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result (min-integer j result)) (loop-j (fx+ 1 j)))))) 27 ms real time 27 ms cpu time (27 user, 0 system) no collections no bytes allocated no minor faults no major faults
On Sun, Nov 4, 2012 at 5:15 PM, Mikael mikael.rcv@gmail.com wrote:
First, please add (declare (block)) and re-compile, this way Gambit won't need to re-look up the procedures in the global namespace that you call, on each invocation. Please report what kind of improvement you got with this.
Now just to make this round of emails complete:
What would be good/very good performance in your view?
Could you send jobs in some kind of batch format to C?
Note that Gambit-internal fxnum variables have a special encoding (the lowermost two(?) bits are object type identifier, which are shifted out when converting them to C int), this takes a little bit of time.
If for instance you make u32vectors/s32vectors/etc. [and send batch ops made on those to C to be performed etc.] then you can just read out the content of those vectors 'raw' in C, there's no interconversion there.
Brgds
2012/11/4 tcl super dude tcleval@gmail.com
> Hi everyone, > > I have been working on an image processing library that I want to > use on android. I know that performance is a bit of a problem on image > processing, so I started to study some bit hacking to improve the > performance on my integer operations. My problem now is that I don't know > why my c-lambdas take so much time to execute. All operations take constant > time and there is no branching. Probably is some sort of conversion time > between C types and GAMBIT types. I expected min-integer and abs-integer to > be a lot faster than min and abs, respectively. Any help is appreciated. > > URL REFERENCE: > http://graphics.stanford.edu/~seander/bithacks.html#IntegerAbs > http://graphics.stanford.edu/~seander/bithacks.html#IntegerMinOrMax > > > MY ENVIRONMENT: > Linux casa 3.6.4-1-ARCH #1 SMP PREEMPT Mon Oct 29 15:13:56 CET 2012 > i686 GNU/Linux > > gcc -v > Using built-in specs. > COLLECT_GCC=gcc > COLLECT_LTO_WRAPPER=/usr/lib/gcc/i686-pc-linux-gnu/4.7.2/lto-wrapper > Target: i686-pc-linux-gnu > Configured with: /build/src/gcc-4.7.2/configure --prefix=/usr > --libdir=/usr/lib --libexecdir=/usr/lib --mandir=/usr/share/man > --infodir=/usr/share/info --with-bugurl=https://bugs.archlinux.org/--enable-languages=c,c++,ada,fortran,go,lto,objc,... --enable-shared > --enable-threads=posix --with-system-zlib --enable-__cxa_atexit > --disable-libunwind-exceptions --enable-clocale=gnu --disable-libstdcxx-pch > --enable-libstdcxx-time --enable-gnu-unique-object --enable-linker-build-id > --with-ppl --enable-cloog-backend=isl --disable-ppl-version-check > --disable-cloog-version-check --enable-lto --enable-gold > --enable-ld=default --enable-plugin --with-plugin-ld=ld.gold > --with-linker-hash-style=gnu --disable-multilib --disable-libssp > --disable-build-with-cxx --disable-build-poststage1-with-cxx > --enable-checking=release > Thread model: posix > gcc version 4.7.2 (GCC) > > Gambit v4.6.6 (SINGLE HOST and GCC extensions enabled) > > > SCHEME CODE: compiled with ---> gambitc -exe bithacks.scm > > ;; bithacks.scm > > > (declare > (not safe) > (fixnum) > (inline) > (inline-primitives)) > > ;; INTEGER ABSOLUTE > (define abs-integer > (c-lambda (int) int > "// we want to find the absolute value of arg1 > int const mask = ___arg1 >> sizeof(int) * CHAR_BIT - 1; > ___result = (___arg1 + mask) ^ mask;")) > > > (define min-integer > (c-lambda (int int) int > "// find the smaller of arg1 arg2 > ___result = ___arg2 ^ ((___arg1 ^ ___arg2) & -(___arg1 < > ___arg2));")) > > (define (test-abs) > (let ((result 0)) > (begin > (display "abs:\n") > (time > (let loop-i ((i 0)) > (cond ((fx< i 100000) > (set! result (abs i)) > (loop-i (fx+ 1 i)))))) > (display "abs-integer:\n") > (time > (let loop-j ((j 0)) > (cond ((fx< j 100000) > (set! result (abs-integer j)) > (loop-j (fx+ 1 j))))))))) > > (define (test-min) > (let ((result 0)) > (begin > (display "min:\n") > (time > (let loop-i ((i 0)) > (cond ((fx< i 100000) > (set! result (min i result)) > (loop-i (fx+ 1 i)))))) > (display "min-integer:\n") > (time > (let loop-j ((j 0)) > (cond ((fx< j 100000) > (set! result (min-integer j result)) > (loop-j (fx+ 1 j))))))))) > > ;; run tests > > (test-abs) > (test-min) > > ;;;;;;;;;;;;;;;;;;;;;; RESULTS;;;;;;;;;;;;;;;;;;;;; > > abs: > (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (abs > i)) (loop-i (fx+ 1 i)))))) > 2 ms real time > 3 ms cpu time (3 user, 0 system) > no collections > no bytes allocated > no minor faults > no major faults > abs-integer: > (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result > (abs-integer j)) (loop-j (fx+ 1 j)))))) > 22 ms real time > 23 ms cpu time (23 user, 0 system) > no collections > no bytes allocated > 2 minor faults > no major faults > min: > (time (let loop-i ((i 0)) (cond ((fx< i 100000) (set! result (min i > result)) (loop-i (fx+ 1 i)))))) > 2 ms real time > 3 ms cpu time (0 user, 3 system) > no collections > no bytes allocated > no minor faults > no major faults > min-integer: > (time (let loop-j ((j 0)) (cond ((fx< j 100000) (set! result > (min-integer j result)) (loop-j (fx+ 1 j)))))) > 28 ms real time > 30 ms cpu time (27 user, 3 system) > no collections > no bytes allocated > no minor faults > no major faults > > > _______________________________________________ > Gambit-list mailing list > Gambit-list@iro.umontreal.ca > https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list > >
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
-
Mikael -
tcl super dude