Comments on: Common Lisp vs. C/ASM: Battle FFT http://nklein.com/2009/10/common-lisp-vs-casm-battle-fft/ software development and consulting Tue, 10 Aug 2010 13:38:03 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: Paul Khuong http://nklein.com/2009/10/common-lisp-vs-casm-battle-fft/comment-page-1/#comment-430 Paul Khuong Sun, 25 Oct 2009 14:28:49 +0000 http://nklein.com/?p=987#comment-430 Yes. Unfortunately, the version currently available on the website lacks some floating point improvement that were merged in for 1.0.30. You'll have to build it yourself from the latest source (a 5-10 minute point and shoot process, except on 10.6 where it seems to be point and pray for some). Yes. Unfortunately, the version currently available on the website lacks some floating point improvement that were merged in for 1.0.30. You’ll have to build it yourself from the latest source (a 5-10 minute point and shoot process, except on 10.6 where it seems to be point and pray for some).

]]> By: pat http://nklein.com/2009/10/common-lisp-vs-casm-battle-fft/comment-page-1/#comment-429 pat Sun, 25 Oct 2009 06:09:07 +0000 http://nklein.com/?p=987#comment-429 CCL doesn't notice when I <b>(declaim (ftype (function ... ...) foo))</b>. So, I've had similar problems finding good ways to optimize CCL code in the past. I really haven't put any effort yet into optimizing this better for CCL. Under CCL, this code conses a ton where it doesn't cons at all under SBCL. So, certainly, there is work to be done to get the declarations happy with CCL. CCL doesn’t notice when I (declaim (ftype (function … …) foo)). So, I’ve had similar problems finding good ways to optimize CCL code in the past. I really haven’t put any effort yet into optimizing this better for CCL. Under CCL, this code conses a ton where it doesn’t cons at all under SBCL. So, certainly, there is work to be done to get the declarations happy with CCL.

]]> By: pat http://nklein.com/2009/10/common-lisp-vs-casm-battle-fft/comment-page-1/#comment-428 pat Sun, 25 Oct 2009 06:06:01 +0000 http://nklein.com/?p=987#comment-428 Are there x86-64 builds of SBCL for Mac? My Linux box is very slow. Are there x86-64 builds of SBCL for Mac? My Linux box is very slow.

]]> By: Paul Khuong http://nklein.com/2009/10/common-lisp-vs-casm-battle-fft/comment-page-1/#comment-427 Paul Khuong Sat, 24 Oct 2009 20:47:18 +0000 http://nklein.com/?p=987#comment-427 The timings will probably be even more interesting for x86-64 builds, at least in the case of SBCL (definitely, for Bordeaux FFT) and I would expect for FFTW. I don't know how CCL handles FP, but I wouldn't be surprised if you also saw a speed-up there. The timings will probably be even more interesting for x86-64 builds, at least in the case of SBCL (definitely, for Bordeaux FFT) and I would expect for FFTW. I don’t know how CCL handles FP, but I wouldn’t be surprised if you also saw a speed-up there.

]]> By: Anton Vodonosov http://nklein.com/2009/10/common-lisp-vs-casm-battle-fft/comment-page-1/#comment-415 Anton Vodonosov Sat, 17 Oct 2009 10:08:14 +0000 http://nklein.com/?p=987#comment-415 Is it because your optimizations were more sbcl-specific, or ccl is just that slow? Is it because your optimizations were more sbcl-specific, or ccl is just that slow?

]]> By: pat http://nklein.com/2009/10/common-lisp-vs-casm-battle-fft/comment-page-1/#comment-413 pat Fri, 16 Oct 2009 23:47:17 +0000 http://nklein.com/?p=987#comment-413 That CCL number is 512x512 (2D). That SBCL number is 512x512x16 (3D). These numbers are v1.3 on a 1M (1D). Apple, oranges, and bananas. v1.3 SBCL is 0.13 on a 512x512 while CCL is 2.39. That CCL number is 512×512 (2D).

That SBCL number is 512×512x16 (3D).

These numbers are v1.3 on a 1M (1D).

Apple, oranges, and bananas. v1.3 SBCL is 0.13 on a 512×512 while CCL is 2.39.

]]> By: Anton Vodonosov http://nklein.com/2009/10/common-lisp-vs-casm-battle-fft/comment-page-1/#comment-412 Anton Vodonosov Fri, 16 Oct 2009 23:13:31 +0000 http://nklein.com/?p=987#comment-412 CCL 30 secs? For version 1.2 CCL vs SBCL was 9.77 secs vs 2.72 secs. Now it's 0.72 vs 30? CCL 30 secs?

For version 1.2 CCL vs SBCL was 9.77 secs vs 2.72 secs. Now it’s 0.72 vs 30?

]]>