a all would have to know my name is the one who's on i'm phones or nation and universe you know signal processing and williams the a team and uh much computation you yeah i especially indeed you these are the a polite detail signal processing i detection of processing that it would in terms of two or high sampling applications yeah so you you gaze yeah actually so we used to use uh we we we as i is the circus about you know what you had a architecture based the sheep use that you know it was also to and place up to the of a stopwatch but here you have this to some problems top top four is the problem is so that she the other sam she can share everybody so that you can how do you have a lot word or i guess was always signal processing algorithms are high high hell of it so one cup to what your voice you can so i would assume so uh that that based on feedback so you could have an adaptive filter is uh one audible discontinuity and the another one is uh architecture space people were not or don't have outermost in so i E twenty four we a pit i had already would you jones there are some well what have the but we do all word uh sometimes at the dollar alignment or word awhile soon corning journal where the white in top artists communication or what little the algorithm that i'm going to project here at least uh all the two but that is a filter yeah so you know what are the two that is a filter is uh use the for channel equalization noise cancelling beamforming and many applications and all the team of filter so uh components available mood you up with that so why not forties we use uh this kind of able more generality so that is awful wants to have the but but the video over panel panel of them that you can get we do this kind of will so show that it had on that you can be so high say this days ten so the filter then don't the video will have that you is just ten so for a parallel or yeah we needed to see are the kind of a of that so there's of your point for the the thought of it of it a not that question is so why they all that to elect is if it does all that table tenth to filter is um wall frequent to used well the to of the or the T what had the delay line filters uh that you've got the to had a light that had a that that can be paralysed but the what where there is a bit is do so we each to pet lights all that to that use a it and the use of those two gonna a system performance such as uh what that patients with the yeah but with a kind of battle for that you will that is so well we try to implement the disagreed the sin the and the mote acquired this dog so first uh i'm going to i'm going to pretty pretty discuss a sin then white choir architectures and a i'm like the and the pad a computational of all that two would like to the filters here ah i'm with the discuss a or of the and take out of a parallel port and that top home with this the work so have a a have a red id H a noble the in you can all see creations the you you can see K john so i use the for all the two that two filters then i'm going to sure experiment to reject to and but vol oh or in space space if you seven that had other processing feature are you one i don't use those who but also pa skylark take your we use a a what that you call a compile a couple a C does second parallel architectures is seemed that picture that is a single are used stuck what but data for the case of which and architecture we don't one instruction you can come down to say in in this a four addition at one time so this is a very efficient or processing approach well this so architecture or shows the some or word of then date i an a line this say you put be one to at the age get a to be you on a you want to P tool then and uh i there should be well well i want to step and then a novel so and always so and architecture i use a problem too or then what do but to to and the doors on not the architect had our a processor architecture that are used the shape you is a a to clock detection you know in in the you this to be here well do you are off for the claw up a lot in in soul do on global "'cause" side in you we're most there we here because a on more and more out so is uh implemented on the chip so we here form what quark architecture we need a to and play by to pursue of the so this is all point for the set of that the have of the region and of force so that that have that have a really do we need to consider either the synchronization model where the peak or do do you do pause or not working in no so uh so so all for ten so speed that can be the me to the bias in coney need additional way the that that this of the whole so in shall you can and priori they turn that there a and so that that that of the to and uh for the seemed architecture that is a date other of it and the V you to can see that date all alignment or what and of for the case of but entire as as the you for O D video of a are you miss of the media the for and of for the case on work well but T deal will have other things a proportional out to do a mobile courses so the or number bubble pause so you know that yeah you well okay this is the gradient all that to will that is to filter and that this a fit closest of will yeah a uh several steps but don't operation and each is that is a i didn't colour and the one important thing is that do fit a coefficient to a a or or that thilo by low cut and a signal or all we can all that to dull phil top quite efficient to using the global that our signal but in my to process the implementation it is a very important to use a local feed the X so oh one all the T job a parallel and i guess some for this a filter use of coming in from the fact to their uh you can use a look card field of F for for all that station yeah so so as a of filter shows but filter coefficient the K and the filter coefficient to a each she our or or that to the by a local fusion no as a i mentioned a or you'll one are be also a is so we pat a parallel i do each is that okay so this so the this is open code the crunch sort of a parallel region here are to do so we pull type dining that used is at each is stage you close that so i menu compared this to figure your and this to the guillaume E we put pipelining registers at each class it so oh but yeah i didn't type or but you turn you couldn't do computational what this so these the D just that independent to i but computational ball or it this just that have and this is step can not be yeah does same it there because of type dining they "'cause" so each is that oh one step delay does so then you we compute at time and for this is that then we need to to compute a and plus eight and plus a one for this is that so uh yeah this a well if you're the yeah this uh oh it is it shows times that and that is so it shows others yeah stage a number so it the we need to top i some skew the process but and you a what we do know is sort the pipelining used to this yeah the that in use these of high more because the but normal pipelining registers is just them but well this so the band to your boat this top low is the total number of able don't do really of a parallel a jim is a the emitted the bike to two and uh so on a approach is to use a T turn have that have a the computation take that a back a our computation means that we we try to compute to board with and press that the time so you want for of four is the uh what that to for the you can a like take out of that that the um and the you C O A you can use the ten thousand process so for eight are that it had a yeah since the V yeah fast fast racing many many samples of but they are in signal processing don't think a pair of a jim is not be to the by filter told us so use some well i didn't so is to more idea but i use a more complex here L for and so the state G it we norm value of the key and minus a lot of F and minus than and do we need the to compute to D and and a computation computation couldn't be transformed or to store the the you know the currency creation so that you shown use of high complex about do you can be a compute either we that depend noise do not depend on it means the a a three the back but i are he and my so one depend you and yeah and minus one and minus one so yeah so well this can be translated to to yeah and minus so one and you participate bit i E N minus one M mine some X the N so this it becomes up for told a constant coefficient the in you know the can creation yeah is so there are several you creations here and that or not the case a used computational able to yeah and in this case disk this can always still be come but to do to time be quite efficient to the you know the colours see cajun so okay K M and comes at eight K end times K N and minus one plus X K N for the case of what time they are the in you you can seek shown oh that's all the to be more complex so then though parallel order like duration of will tossed the coefficient case so now i and like to put a forty well you spain how be compared a like you the in you know the common seek kitchen you know the in you the currency cajun and use a the back so then that that makes the dependence problem because also to compute to Y N we need the why and my so on to Y M minus want and so oh so you since a very difficult to each is seems not a large got to compute the why and a Y and minus and Y and plus the simultaneous it but it is possible so say yeah why i though we so say you at the times a why my some plus sixty at all then oh why why mine a one is so the initial condition then since that this is a the in you know yeah different C creation you couldn't you could do calm or just leading to two part one is not time you to a solution and the yeah is a particular solution so you for be you go lot the initial condition you can compute to particular solution and that you if we can see though one you'd a particular solution a particular solution can be computed that you had other people's so yeah i i even lot thing in each can be shy and that that breaks the dependence problem then to problems are how we can well compensate for that depend problem you you these days so the in your the currency K john we can a polite blue i'd the mess that to compensate for the well effect all good initial condition so by computing had you can also lose and you a solution oh a to D no become paralysed a computational able you know the can seek age this so at the forces that it do we compute the particular solution and that does step two we prop a okay to the initial condition and that the propagation is kind to the by i in particular solution and the initial condition and the finally we compute the companies solution so do you do a of computation you the some or where the pattern or idea a symbol for the would for the case of O for although though the in you know the because uh though or why there is uh not more than twice though number of will she of computation so that instead you full we use of four processors in C D you can guess say a a on the tools that the pose and top was speed the and the for the constant the coefficient the case of tar speed or can be a how you okay this is the payment L reader a a did X meant to be the well you inter oh for the quasi C pure and the this so has a for a she in the instructions and that this this shows of though or reject to this is the fiddle the and it's of for a a was sixty in and the a a this is it you can sell which are to then processing L samples and that this is thus in the parallel the uh a processing and sample so by a plan seemed that oh be you get the speed P T in a about two hundred all fifty percent the tool two seven used to represent and that this is so these are two for data that the parallel region we sin the so this to mean that using in the instructions so you i no data level parallel to a and a be the that we get or or or are they are higher speed up well don't well but the or over a to lead them is of for whistle four a for the case to an to quell one one show that had L later we can hard but you get any game because some to quite cheap you well need the some C duration mobile but yeah a a a i'm so we do we do some sort of a parallel region with the multiple acquired B can that a a up or sort is it talk case or but data that that have a and we the not to quell and the you couldn't get a bar one hundred the pose in both speed up and the by combining portal data to better parallel do using by to core and C and you get a ball for hundred the nist yeah so to pose and top was speed up this so all experiment experiment a are two shows that oh i for for yeah for pat computation to read the took why cpus though effect of well in a mean you cajun and that's and corny duration will where the P comes of thirty problem so we need a on this or that or tech and or a but this kind of a single you a channel with that we do i'm sure about had out of a job it it a very difficult to to up so will this kind of a sink one you H job would that be do how by applying data never parallel and you can oh well or would this kind of a or where the and the still get some speed okay this is the company no but this do you have we implement T the to but tight this fit on but to in the for type the void no bought that will filter is and uh we get oh fairly close speed uh with a small right toys C then and police in the in for data that have are processing and we couldn't guess speed uh for some more to Y and dark picture and uh we believe but that this top to can be a provide the to are don't or the T that two filters and those was a sure D yeah or that to beamforming are you would thank or questions from you uh this is a i i okay there are no other questions uh