1 / 12

TELL40 VELO time ordering

TELL40 VELO time ordering. Pablo Vázquez , Jan Buytaert , Karol Hennessy , Marco Gersabeck, Pablo Rodríguez. Time ordering. Packet (34b)=ASIC(4b)+ SuperpixelAddress (13b )+BCID(9b)+Hitpattern(8b) Data packets arrive to the TELL40 disordered in time

shima
Télécharger la présentation

TELL40 VELO time ordering

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. TELL40 VELO time ordering Pablo Vázquez, JanBuytaert, Karol Hennessy, Marco Gersabeck, Pablo Rodríguez P. Vazquez (U. Santiago)

  2. Time ordering • Packet(34b)=ASIC(4b)+SuperpixelAddress(13b)+BCID(9b)+Hitpattern(8b) • Data packets arrive to the TELL40 disordered in time • Time ordering = grouping of packets by 9bit BCID value • Grouping is done in 2 steps: • Router  based on 4-bits of BCID (MSB or LSB, see later) • Memory storage  remaining 5-bits BCID  Goalofthistalk “VELO” “Standard LHCb” memory routing Decoding Dataprocessing P. Vazquez (U. Santiago)

  3. 4-bit BCID VELO packetrouter • Non blocking schemeusing internal memory (FIFO’s) • Different number of input links considered: • 16 @160 MHz (320) for a half (full) module as described in theTDR • 10 @320 MHz for a full module, when reducing opticallinks • 2 architectures under study: • “Cascade” of 1-bit router elements. Each element routes packets based on 1-bit • “crossbar” : singlestage N x 2M portsroutespacketsbased on all bits • Use of altera library megafunctions: lpm_scfifo, lpm_compare... P. Vazquez (U. Santiago)

  4. 1-bit router elements r0 w0 FIFO_00 If (BID[i]=0) Write FIFO_00 else Write FIFO_01 More Full? MUX 2-to-1 FLIP FLOP Data 0 Out 0 FIFO_01 r0 FlipFlop0 r1 More Full? w1 Out 0 If (BID[i]=0) Write FIFO_10 else Write FIFO_11 FIFO_10 MUX 2-to-1 FLIP FLOP w0 Out 1 Data 1 r1 FIFO_11 Data 0 FlipFlop1 Out 1 If (BID[i]=0) Write FlipFlop0 else Write FlipFlop1 2x2 routerelementuses 4 fifoswhile 1x1 routerelementdoesn’t use fifos P. Vazquez (U. Santiago)

  5. 10x16 router with “1-bit router elements” 0000 2x2 port 1-bit router 1x2 port 1-bit router xxx0 xx01 xxxx xxx1 xx11 xxxx 1111 P. Vazquez (U. Santiago)

  6. FPGA resources and max speed • Several configurations compiled for the actual device: alterastratixV 5SGXEA7N3F45C2 • Router can run @ 320 MHz with 1-5% ofresourceswith a cross-sectional bandwidth 1.6-2.6 higher than required (2G pakets/s peak for full module) P. Vazquez (U. Santiago)

  7. Simulation • Generate simulated input data using the known distribution of the event size and latency (see next slides) • Aim: the dependency of packet loss vs fifo depth and routing strategy (MSB vs LSB bits) • Aim: data packet loss << 0.1%? • Every 34-bit FIFO is allocated on a M20K memory block => optimal is 512 x 40bit. Could be extended • We will start simulating the 10x16 @320 MHz case P. Vazquez (U. Santiago)

  8. Event sizedistribution • The 20 inputsof a full module are below 40% ofpeakpacketratecapacity At theinputofrouter @320 MHz Links 1-8 mean 0.37 pkt/clk Links 9-12 mean 0.31 pkt/clk Link 13,14 mean 0.25 pkt/clk Link15,16 mean 0.30 pkt/clk Link17,18 mean 0.26 pkt/clk Link19,20 mean 0.16 pkt/clk Data fromMartin van Beuzekom P. Vazquez (U. Santiago)

  9. Simulationlatency • Packets arrive at the tell40 with variable latency: • Routing with the MSB bits generates data rate peaks 10 times higher than routing with LSB bits • increases packet loss probability (but maybe still acceptable), • but is nicer for MEP: consecutive events are stored in the samememory • Maybe better suited for data-processing (idle time between peaks) probability P. Vazquez (U. Santiago)

  10. Summary • TELL40 VELO router block has been investigated • Compilation of designs based on 1-bit router elements shows that with 1-5% of FPGA resources, the router can handle easily the requiredbandwidth • Simulations with self-generated data will show the packet loss (with respect to fifodepth and MSB/LSB routing strategy) P. Vazquez (U. Santiago)

  11. backup P. Vazquez (U. Santiago)

  12. 10x16 router 1 stage w0 r0 FIFO_0 Data 0 BCID= 0000 MUX 10-to-1 FLIP FLOP Out 0 r15 BCID= 1111 MUX 10-to-1 FLIP FLOP w9 Out 15 FIFO_9 Data 9 P. Vazquez (U. Santiago)

More Related