> Le mer 31/03/2004 à 02:51, George Bratis a écrit : >> I think that if you want to use a low cost dsp you must rewrite ti in >> assebmly. >> In our tests you will need 4-5 times more MIPS just compiling C code. > > You may need to rewrite 3-4 compute-intensive, but definitely not the > whole thing. > > Jean-Marc >What's your guess on the total MIPS the optimized and unoptimized code will take? --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'speex-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
Le mer 31/03/2004 à 02:51, George Bratis a écrit :> I think that if you want to use a low cost dsp you must rewrite ti in > assebmly. > In our tests you will need 4-5 times more MIPS just compiling C code.You may need to rewrite 3-4 compute-intensive, but definitely not the whole thing. Jean-Marc -- Jean-Marc Valin http://www.xiph.org/~jm/ LABORIUS Université de Sherbrooke, Québec, Canada -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 190 bytes Desc: Ceci est une partie de message numériquement signée. Url : http://lists.xiph.org/pipermail/speex-dev/attachments/20040331/80b7f752/signature-0001.pgp
We are bellow 30MIPS for our 16bit single MAC fixed point DSP. 8khz/8kbps I aggree with Jean-Marc that MIPS is dependent from the target DSP, but it's a very good indication when you compare same architecture DSPs. It is very difficult (impossible) to use effecient a DSP using only c. George ----- Original Message ----- From: "Andre Borrelly" <andre@myavalaunch.com> To: <speex-dev@xiph.org> Sent: Thursday, April 01, 2004 7:40 AM Subject: Re: [speex-dev] speex on a DSP chip? <p>> > Le mer 31/03/2004 à 02:51, George Bratis a écrit :> >> I think that if you want to use a low cost dsp you must rewrite ti in > >> assebmly. > >> In our tests you will need 4-5 times more MIPS just compiling C code. > > > > You may need to rewrite 3-4 compute-intensive, but definitely not the > > whole thing. > > > > Jean-Marc > > > What's your guess on the total MIPS the optimized and unoptimized code > will take? > --- >8 ---- > List archives: http://www.xiph.org/archives/ > Ogg project homepage: http://www.xiph.org/ogg/ > To unsubscribe from this list, send a message to'speex-dev-request@xiph.org'> containing only the word 'unsubscribe' in the body. No subject is needed. > Unsubscribe messages sent to the list will be ignored/filtered.--- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'speex-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
> What's your guess on the total MIPS the optimized and unoptimized code > will take?Please define MIPS (and just not the acronym). Seriously, the number of instruction required will depend a lot on the actual CPU/DSP. Even without that, the complexity of Speex can vary from about 6 mflops to around 50, depending on the sampling rate, the bit-rate and the complexity setting chosen. The best I can say to give you an idea is that the 8 kHz/8 kbps mode requires in the order of 10-12 mflops. Jean-Marc -- Jean-Marc Valin http://www.xiph.org/~jm/ LABORIUS Université de Sherbrooke, Québec, Canada -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 190 bytes Desc: Ceci est une partie de message numériquement signée. Url : http://lists.xiph.org/pipermail/speex-dev/attachments/20040401/27c944c6/signature-0001.pgp