search for: q0

...n". One problem with shl always generating non-poison zeroes in the lower bits is that it then is not equivalent to multiplication. This breaks optimizations like: %t = ... %q = shl i32 %t, 6 =/=> %t = ... %q = mul i32 %t, 64 and (probably more meaningful) %t = ... %q0 = shl i32 %t, 6 %q1 = mul i32 %q0, 3 =/=> %t = ... ;; %q0 = shl i32 %t, 6 %q1 = mul i32 %q0, (3 * 64) and (probably even worse) it also breaks reassociation: %t = ... %q0 = shl i32 %t, %P %q1 = mul i32 %q0, %Q (No other uses of %q0, but there are uses of %q1) =/=>...

...c: received signal 11, cleaning up Not sure what's up with this. This test doesn't appear on the nightly tester although it *should*. > > ===================== MultiSource/Benchmarks/Olden/power > > Only cbe failed: > > 6c6 > < TR=0.81, TI=0.16, P0=6397.86, Q0=1288.50 > --- > > TR=0.81, TI=0.16, P0=6397.87, Q0=1288.50 > 26c26 > < TR=0.80, TI=0.16, P0=7565.44, Q0=1525.86 > --- > > TR=0.80, TI=0.16, P0=7565.44, Q0=1525.85 > 34c34 > < TR=0.79, TI=0.16, P0=7725.99, Q0=1558.54 > --- > > TR=0.79,...

...hese? I see the nightly reports, but they don't say what's XFAIL much less /how/ they're expected to fail. > > ===================== MultiSource/Benchmarks/Olden/power > > > > Only cbe failed: > > > > 6c6 > > < TR=0.81, TI=0.16, P0=6397.86, Q0=1288.50 > > --- > > > TR=0.81, TI=0.16, P0=6397.87, Q0=1288.50 > > 26c26 > > < TR=0.80, TI=0.16, P0=7565.44, Q0=1525.86 > > --- > > > TR=0.80, TI=0.16, P0=7565.44, Q0=1525.85 > > 34c34 > > < TR=0.79, TI=0.16, P0=7725.99, Q0...

From: Jyri Sarha <jsarha at ti.com> I optimized Speex resampler for NEON capable ARM CPUs. The first patch should speed up resampling on any platform that can spare the increased memory usage. It would be nice to have these merged to the master branch. Please let me know if there is anything I can do to help the the merge. The patches have been rebased on top of master branch in

...producing 13,000+ lines of parsing actions, it produces the following two lines: Reading input files ...kc++: segmentation violation kc.llc: received signal 11, cleaning up ===================== MultiSource/Benchmarks/Olden/power Only cbe failed: 6c6 < TR=0.81, TI=0.16, P0=6397.86, Q0=1288.50 --- > TR=0.81, TI=0.16, P0=6397.87, Q0=1288.50 26c26 < TR=0.80, TI=0.16, P0=7565.44, Q0=1525.86 --- > TR=0.80, TI=0.16, P0=7565.44, Q0=1525.85 34c34 < TR=0.79, TI=0.16, P0=7725.99, Q0=1558.54 --- > TR=0.79, TI=0.16, P0=7725.99, Q0=1558.53 36c36 < TR...

...t %[ret], #16, %[a]" + : [ret] "=&r" (ret) + : [a] "r" (a) + : ); + return ret; +} +#else +static inline int32_t saturate_32bit_to_16bit(int32_t a) { + int32_t ret; + asm ("vmov.s32 d0[0], %[a]\n" + "vqmovn.s32 d0, q0\n" + "vmov.s16 %[ret], d0[0]\n" + : [ret] "=&r" (ret) + : [a] "r" (a) + : "q0"); + return ret; +} +#endif +#undef WORD2INT +#define WORD2INT(x) (saturate_32bit_to_16bit(x)) + #define OVERRIDE_INNER_PRODUCT_SINGLE...

...t remainder = len % 16; + len = len - remainder; + + asm volatile (" cmp %[len], #0\n" + " bne 1f\n" + " vld1.16 {d16}, [%[b]]!\n" + " vld1.16 {d20}, [%[a]]!\n" + " subs %[remainder], %[remainder], #4\n" + " vmull.s16 q0, d16, d20\n" + " beq 5f\n" + " b 4f\n" + "1:" + " vld1.16 {d16, d17, d18, d19}, [%[b]]!\n" + " vld1.16 {d20, d21, d22, d23}, [%[a]]!\n" + " subs %[len], %[len], #16\n" + " vmull.s16 q0, d16, d20\n&...

On Thu, 9 Sep 2004 20:16:37 -0700 Jeff Cohen <jeffc at jolt-lang.org> wrote: > On Thu, 9 Sep 2004 08:52:10 -0700 > Jeff Cohen <jeffc at jolt-lang.org> wrote: > > > > I haven't got around to this yet but I will. The odds are good the > > > problem is in a BSD system header file so I need to capture the > > > preprocessed source. > > >

...nt to multiplication. This breaks >> optimizations like: >> >> %t = ... >> %q = shl i32 %t, 6 >> >> =/=> >> >> %t = ... >> %q = mul i32 %t, 64 >> >> and (probably more meaningful) >> >> %t = ... >> %q0 = shl i32 %t, 6 >> %q1 = mul i32 %q0, 3 >> >> =/=> >> >> %t = ... >> ;; %q0 = shl i32 %t, 6 >> %q1 = mul i32 %q0, (3 * 64) >> >> >> and (probably even worse) it also breaks reassociation: >> >> %t = ... >>...

...lane + "vdup.s16 q5, %4;\n" //Duplicate mem in q5 lane + + /* We try to process 16 samples at a time */ + "movs %5, %3, lsr #4;\n" + "beq .celt_fir1_process16_done_%=;\n" + + ".celt_fir1_process16_%=:\n" + /* Load 16 x values in q0, q1 lanes */ + "vld1.16 {q0-q1}, [%0]!;\n" + + /* Init four 32 bits sum in q7, q8, q9, q10 lanes */ + "vshll.s16 q7, d0, %[SIGSHIFT];\n" + "vshll.s16 q8, d1, %[SIGSHIFT];\n" + "vshll.s16 q9, d2, %[SIGSHIFT];\n" + "vshll.s16...

...lane + "vdup.s16 q5, %4;\n" //Duplicate mem in q5 lane + + /* We try to process 16 samples at a time */ + "movs %5, %3, lsr #4;\n" + "beq .celt_fir1_process16_done_%=;\n" + + ".celt_fir1_process16_%=:\n" + /* Load 16 x values in q0, q1 lanes */ + "vld1.16 {q0-q1}, [%0]!;\n" + + /* Init four 32 bits sum in q7, q8, q9, q10 lanes */ + "vshll.s16 q7, d0, %[SIGSHIFT];\n" + "vshll.s16 q8, d1, %[SIGSHIFT];\n" + "vshll.s16 q9, d2, %[SIGSHIFT];\n" + "vshll.s16...

Hi: I have been curious as to Vorbis' performance at -q0 for some time now. When RC3 came out, I used a track I have here and encoded it at q0 at 11.025, 16, 22.05, 32 and 44.1khz, both in stereo and in mono. After having heard the great sound of -q 0 in stereo, I thought I'd see what bitrate I got with -q0 in mono. I was disturbed to discover that...

May I suggest that the up coming RC4 or 1.0 needs some quality selection adjustments. It has been said that if you are happy now with a particular quality setting, later releases will produce the same quality but reduce the average bitrate. This is not currently the case. IMHO -q4 in CVS is producing almost perfect results and features audio frequencies up to 21khz -- almost CD quality.

...nts/Developer/usr/bin/make all-recursive Making all in libspeexdsp CC preprocess.lo CC jitter.lo CC mdf.lo CC fftwrap.lo CC filterbank.lo CC resample.lo In file included from resample.c:104: ./resample_neon.h:134:12: error: unknown register name 'q0' in asm : "q0"); ^ ./resample_neon.h:195:13: error: invalid output constraint '+l' in asm [len] "+l" (len), [remainder] "+l" (remainder) ^ 2 errors generated. make[2]: *** [resample.lo] Error...

..., nrow=n0, ncol=1) #generating the 0/1 exposure Yprob=1/(1+exp(-(10*A-5*(W1)^2+2*W2))) Y=matrix(rbinom(n0, 1, Yprob), nrow=n0, ncol=1) #generating the 0/1 exposure zero=data.frame(rep(0, n0)) Q=glm(cbind(Y, 1-Y) ~ A + W1 + W2, family='binomial') QA=predict(Q, newdata=as.data.frame(A)) Q0=predict(Q,newdata=(A=zero)) I've tried many variations of the last line (Q0) to get the predicted values when A=0 with no luck. With this code, I get errors that my A=zero is a list even though I made it into a data frame. This is the version of the code (after my reading) that *shoul...

Caleb wrote: > i dont understand you people. >-q0 should be poor quality, only that in vorbis, the poor quality is > actually good! :) Another advantage I see in reducing the nominal bitrate for q0 to 48kb/s with a ~13khz lowpass is a smoother transition in average bitrate and frequency resolution from 22khz to 44khz sampling. Currently th...

...quot; argument and rather receives "null" as the string parameter. When calling the web service using an other client (eclipse Web services explorer), I see it generates the following body: <soapenv:Envelope xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/" xmlns:q0="http://web2.web/" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"> <soapenv:Body> <q0:greetMe> <arg0>Olivier-from-eclipse</arg0> </q0:greetMe> </soapenv:Body>...

Does anyone know exactly what ComposerIndices in Target.td is all about? I see just one place where it's used in X86 but it's not clear from the comments in Target.td and it's one usage, exactly what this feature is about. Tia. Reed

I just tried rc4-cvs and I can't believe how good it sounds! I can hear very little difference between the wav and the q0-ogg on my above-average stereo... It really makes MP3 look (sound) *pathetic* in comparision! I did some comparisions on a few songs, from noisy 77-punk to Jean-Michel Jarre to EBM/Industrial, and q0/~64kbit RC4-vorbis sounds as good or better than V9/~128kbit LAME-mp3 on ALL of them! I ripped...

As the subject implies, my question is: is it possible to use the -b (or -M) flag at non-44K sample rates? I'm working with an application that is trying to optimize for very small audio filesize. I found that downsampling to 11K and then using q0 gives high compression, but won't seem to drop below 64kbps or so. It seems like the combination of downsampling, then reducing to 30kps would compress things even further, but I'm not sure if this is allowed. Any ideas? <p>--- >8 ---- List archives: http://www.xiph.org/arch...